We have known about media bias effect for decades: the belief that the media is biased against your side of a debate. New research finds that the same belief applies to misinformation. While the research was focused on political issues, the underlying cause applies equally to misinformation about brands, companies, and business issues. In this short midweek episode, Neville and Shel find that the PR industry has not yet acknowledged the phenomenon, which requires strategies to address it.

Links from this episode:

• Think the Media’s Biased Against You? You Probably Think Misinformation Is, Too
• The Hostile Media Effect
• The Influence of Hostile Media Perceptions on Misinformation Beliefs and Sharing
• Hostile Media Effects on Twitter, Social Identity, and Media Bias Perceptions
• Fake News Has Real Effects on Consumer Demand
• The Impact of Fake News on Consumer Behavior and Market Outcomes
• Political Identity, Media Trust, and Susceptibility to Misinformation

The next monthly, long-form episode of FIR will drop on Monday, June 29.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville Hobson:
Hi everyone, and welcome to For Immediate Release. This is episode 519. I’m Neville Hobson.

Shel Holtz:
And I’m Shel Holtz. When you think about all the misinformation out there—fake news, bad-faith spin—do you think it’s mostly aimed at your side of an argument or the other side? Most of us, if we’re honest, feel like it’s aimed at us. And there’s now research saying that feeling is nearly universal.

Even though the research was based on political discourse, it has a direct connection to organizational communication. We’ll explain right after this.

All right, let’s start by backing up for a second. There’s a concept called the hostile media effect. It’s been around since the 1980s. The original study showed pro-Israeli and pro-Arab students the exact same news coverage of the exact same event. Both groups walked away convinced it was biased against their side.

Everyone saw exactly the same footage, but they reached opposite conclusions. And the more committed you were, the more certain you were that the media was out to get you.

That finding has held up for 40 years, and it’s a big reason trust in news has collapsed as politics has gotten more tribal.

Now let’s add the new wrinkle. A team at the University of Amsterdam asked whether that same instinct applies to misinformation—to fake news. They surveyed 4,000 people across Germany, the Netherlands, and Poland around the 2024 European elections.

Nearly half said their preferred party was particularly targeted by misinformation. Ask about the party they liked least, and that number got cut in half. They’re calling it the hostile misinformation effect, and it got stronger the more politically engaged people were. The more plugged in people felt, the more victimized they felt.

Now, Neville, you might think that’s a political science finding. But the mechanism underneath isn’t about politics; it’s about identity and motivated reasoning. Every brand, every company, every department is an identity group.

Your most loyal customers are partisans. Your most engaged employees are partisans. The research says the people most attached to your organization are exactly the ones primed to believe any criticism out there is unfairly targeting them.

Now think about a crisis. Your defenders don’t need convincing that your critics are unfair. They already assume it. The minds still open are the uncommitted people in the middle.

Among neutrals, knowing more made them see less bias. It’s only partisans who dig in.

So if someone criticizes a brand that some people love, the brand’s biggest fans may see that as an attack rather than just an honest review—and respond in kind. There was no crisis, but now maybe there is.

There’s an internal angle here, too. Picture a layoff memo or a return-to-office announcement. Leadership reads it as fair. But every faction inside the company—by department, by level, by tenure—is wired to read the same message as unfair to them.

“We said it neutrally” is no defense because neutrality is in the eye of the beholder.

This notion reveals a trap for communicators. When bad coverage hits, it’s tempting to wave it away as misinformation. But “fake news” self-destructed as a term the moment it got weaponized to mean “any story I don’t like.”

Cry misinformation every time you’re criticized, and you train your audience to tune out the label. You also look evasive to the exact neutrals you need to reach.

So this is where I want to bring you in, Neville. We’ve spent years on this show talking about declining trust and the misinformation environment. This research says the problem isn’t just that there’s more bad information out there; it’s that people are wired to feel personally besieged by it.

And I’m not sure our profession has reckoned with what that means.

Neville Hobson:
Yeah, it doesn’t sound like it, Shel. I don’t think so.

It’s actually quite fascinating looking at the Nieman Lab article you shared with me in our Slack channel and seeing the depth of the research on a topic that I had no idea was even a thing to look into.

I found it interesting in a number of areas.

For instance, the study you quoted from the 2024 European Parliament elections got me thinking. The tendency to see misinformation as directed at you seems more pronounced the farther right politically someone is.

That caught my attention because isn’t that precisely what we’re seeing in the United States with the Trump MAGA movement?

Here in the UK, we’ve got Reform and an even newer party that’s emerged further to the right. Those groups often function as an echo chamber for the kinds of messages Trump promotes. They’re constantly criticizing anything anyone else says as an attack and talking about issues in ways that rile people up and stimulate hostile reactions in return.

We see a lot of that in this country right now.

It’s interesting that this study has been done, and I think the way you’re connecting it to organizational communication is a good call. It certainly gives us a lot to think about.

One question it prompted in my mind concerns the point about engagement and partisanship. If the more engaged and partisan someone is, the stronger this effect becomes, does that mean an organization’s most loyal stakeholders are actually its most vulnerable to this kind of perception?

What do you think?

Shel Holtz:
Absolutely. I think that’s exactly the connection we can draw between this study and organizational communication.

If somebody criticizes the company based on an experience they had—and let’s say that criticism goes viral—and it was sincere and well-intentioned, then the partisan defenders of that organization are going to feel attacked. They’re likely to respond in kind and escalate a situation that probably would have faded into the background if left alone.

I think that’s one of the fallouts organizations can experience from this phenomenon. The more partisan you are, the more besieged you’re going to feel when you perceive something being said about the brand or organization as unfair—even if it was perfectly fair.

Neville Hobson:
So how do you address that within the organization?

Shel Holtz:
That’s an interesting question, and it’s hard to fight because you really can’t argue people out of it.

One related concept is the third-person effect—the idea that other people are more susceptible to media influence than we are ourselves.

In other words: I can see what the media is trying to do, but other people are going to be fooled by it.

When you stack that together with the idea that your group is being unfairly targeted, you get a complete worldview: I’m clear-eyed, my group is the victim, and everyone else is gullible.

There was a fascinating study where researchers took 661 Coca-Cola drinkers and showed them a real fake-news story—a 2016 hoax claiming that Dasani water was being recalled because parasites had been found in it.

The finding was that the people most confident in their own ability to spot fake news were the most convinced that other people would be fooled by it. They were also the ones most loudly demanding that Coca-Cola take corrective action.

Sometimes the stakeholders who are screaming “Do something about misinformation!” aren’t reacting to the actual threat. They’re reacting to a belief that other, less discerning people are being duped.

That makes the challenge even more complicated for communicators.

Neville Hobson:
Yeah, it’s weird, isn’t it?

The next question that comes to mind is this: If both sides feel targeted regardless of what’s actually out there, what should communicators do? Is there an approach that works when perception is this detached from reality?

Shel Holtz:
From an organizational standpoint—and I’m less interested in the political implications for purposes of this podcast—I think there are a couple of things.

First, the more prebunking you can do, the better.

When one of these situations comes up—a bad review, criticism from the media, negative reporting—you can immediately point people to information you’ve already published that addresses the issue. Having a bank of credible material you can reference may keep people from getting unnecessarily riled up.

The other thing is to respond quickly, but not emotionally.

If you can maintain a sense of calm—or even a sense of humor—you increase the likelihood that others will follow suit.

If people see that the company isn’t feeling besieged and isn’t acting attacked, that may help tamp down some of the reaction.

Neville Hobson:
So this becomes a major issue for trust, doesn’t it?

And I imagine the role of artificial intelligence in all of this only exacerbates the problem. Is that how you see it?

Shel Holtz:
Yeah, I do.

The flood of AI-generated slop out there—content targeting your organization, your brand, or your leaders—is only going to increase exponentially.

If somebody has an axe to grind and wants to flood search engines or AI summaries with negative content, AI makes that dramatically easier.

Now, to be fair, the research we’re discussing focused on information published through media outlets. That may be an important distinction.

People might be more skeptical of something posted on a blog or LinkedIn than something published by a mainstream news outlet.

That’s where this research suggests people feel especially attacked.

Neville Hobson:
That was my thought as well.

Going back to the study, we’re looking at this from a political perspective. We all consume information online, and most of us have preferred sources.

Meanwhile, mainstream media is going through what seems like a growing crisis of trust.

You see constant battles online between people citing one newspaper versus another. It’s distracting, and it wastes an enormous amount of time and energy.

It also got me thinking about how I react to some of this. The suggestion that people on the political right are more susceptible to this phenomenon doesn’t really describe me. I’m more in the middle.

I don’t react the way I see some people reacting—especially on that delightful conversation platform known as X.

People vent their spleens there. Maybe it makes them feel better, but I don’t think it advances understanding in any meaningful way.

Then again, perhaps that’s not the point.

The point seems to be: I win, you lose.

And that’s very much the Trump approach. It feels like that’s where much of this is headed.

Shel Holtz:
Yeah.

One of the stranger findings from the research was that when researchers examined whether people felt more victimized after their party lost an election, the results weren’t what you’d expect.

You’d think the losers would feel most targeted.

Instead, the people whose party won were more likely to believe their side was being targeted by misinformation.

Feeling besieged isn’t necessarily about being under threat. It’s about identity.

And since you mentioned X, there’s another interesting strand of research.

People may dismiss something because they saw it on X. But researchers found that the source of a message can trigger hostile media perceptions independently of the content itself.

Your company’s name on a statement can be enough to act as a bias cue for people who already have feelings about you.

The exact same words can land very differently depending on whose logo appears at the top.

That’s worth thinking about because it means message discipline alone can’t solve a credibility problem.

Neville Hobson:
This is a bigger dilemma than it might seem at first. A real conundrum for communicators.

The Nieman Lab article is lengthy, but it’s definitely worth reading.

Shel Holtz:
Yeah. And there will be links in the show notes to some of the original research on identity bias as well.

The reason I chose this topic is that I’ve never heard it discussed in PR circles. I’ve never seen it covered in PR textbooks or books about public relations.

This was completely new to me.

That’s one reason I’m still struggling with an answer about how to deal with it.

But it certainly starts with recognizing that it’s happening.

Neville Hobson:
I agree. It was new to me as well.

The more I think about it, though, the more it seems to describe the environment we’re operating in today.

Now we just have to figure out what to do about it.

Shel Holtz:
Yes, we do.

And that will be a 30 for this episode of For Immediate Release.

The post FIR #519: Is Misinformation Biased Against You? appeared first on FIR Podcast Network.

The history of public relations over the last 30 years is a litany of one failure after another — failures to recognize and embrace technologies that represented seismic shifts in how people and organizations communicate. The internet. The web. Social media. Smartphones. The video shift. And now, with AI, the industry seems poised to do it again. As many organizations explore how AI will reshape them, PR agencies still seem unable to figure out billing models to replace the now-useless hourly rate. In this short midweek episode, Neville looks at a post from Stephen Waddington that laments the industry’s intransigence, and Shel and Neville discuss what PR should be doing.

Links from this episode:

• The future of jobs in PR: will we get the third technology shift wrong too? (by Stephen Waddington)
• It looks like PR has its head in the sand about AI (by Neville Hobson)
• Senior practitioner neglect of digital/social skills a huge threat to PR’s future (2015 post by Shel Holtz)
• Once Again, This Time with AI, the Communications Profession Will Be Late to Embrace a Valuable Technology (2023 post by Shel Holtz)

The next monthly, long-form episode of FIR will drop on Monday, June 22.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Shel Holtz: Hi everybody and welcome to episode number five hundred and eighteen of For Immediate Release. I’m Shel Holtz.

Neville Hobson: And I’m Neville Hobson. So here’s a question I want to put to you right at the start, and I’d like you to sit with it as Shel and I work through this topic today. Public relations as a profession has faced two seismic technology shifts in the last 30 years. In fact, more than two, but I’m just going to mention these two. The internet arrived in 1995. Social media arrived around 2007. And in both cases, PR largely got it wrong. Not wrong in the sense of ignoring the technology. Wrong in the sense of fundamentally misreading what it meant. In 1995, we thought the internet was a publishing problem. In 2007, we thought social media was just another broadcast channel. And the disciplines that grew out of both—search, content marketing, influencer marketing—were largely built by people who weren’t us, people outside the profession who saw what we missed. So the question is: are we about to do it a third time? We’ll address that question in just a minute.

That’s the challenge Stephen Waddington lays down in a piece he’s just written for Influence, the member magazine of the CIPR, the Chartered Institute of Public Relations. Stephen is someone whose thinking I respect considerably. He’s been one of the sharper and more honest voices in UK PR for years. And this article comes off the back of a book he’s just co-edited, AI and Public Relations: A How-To Guide for Implementation and Management, published in May. And what he’s arguing in this piece is that this is no longer a theoretical debate, as job reductions are happening now. He gives specific examples. Three account executives doing media monitoring—that’s now one tool. A two-person intranet team—that’s now a fraction of the effort. The UK government has listed public relations professionals among the twenty occupations most exposed to large language models. We’re on the list. Early career employment in those sectors is also in relative decline.

Now, Waddington is not a pure pessimist. He sees a plausible optimistic path. The career pyramid becomes a diamond. Firms building roles around insight and risk management rather than billable hours. A rough near-term reduction of perhaps fifteen to twenty percent in entry-level positions, followed by net growth as scope expands and new roles emerge, the way digital did after 2000. He thinks in-house teams especially have an opportunity here. When AI absorbs the routine, it frees space for the work that corporate communication teams have always needed but rarely had capacity for. But he gives serious, genuine weight to the pessimistic case too. And this is where I think the article gets interesting. He references Martin Ford, author of The Rise of the Robots in 2015, and Ford’s argument that previous technology waves hit one tier of the workforce and the tier above absorbed the displaced.

This time Ford says there’s no tier above. The advisory work that absorbed previous shifts is itself the target. Waddington doesn’t fully accept that in his article, but he doesn’t dismiss it either. And then there’s the argument that I think should be keeping every agency head and comms director awake at night—the pipeline. He’s hearing a common response from firms right now: freeze your apprenticeship schemes, freeze your graduate intake, let AI cover the production work. And he calls that, bluntly, organizational self-harm. Because in five years, those organizations will have nobody who understands how the systems actually work, why they fail, and crucially when to override them. You cannot run an advisory profession without a pipeline. And you cannot build a pipeline if you spent five years dismantling the entry points. So that’s where I think we should start today’s conversation. Not with the technology, with the choices.

Because Waddington’s closing argument, and it’s what I find compelling, is that human agency still exists here. The technology isn’t making decisions. We are. The question is whether we’re making them wisely, or whether for the third time in thirty years, we’re about to hand the future of our profession to people who aren’t us. Shel, what’s your instinct on this?

Shel Holtz: Very much what yours and Stephen’s is. I have been saying for decades that the public relations industry is always, always, always late to the game when there is a new technology that is going to shape the way communicators do their jobs. We were late to the internet, for sure. We were late to the World Wide Web. My first book on communicating online—well, actually, my first book was on intranets, but the first one that got any attention was Public Relations on the Net—came out before the World Wide Web, before there was a graphical user interface. So there were plenty of opportunities for PR before the web, based on the capabilities of the internet. Then we missed the web, then we missed social media. In between we missed some other seismic shifts—mobile, being able to communicate with people based on the fact that they now had this computer in their pocket. We missed the pivot to visual communication, we missed the pivot to video communication. And now, yeah, we are poised to miss the pivot to AI. And that’s not to suggest that PR people aren’t using it. I think they are, but I think they’re using it at a very superficial level and are succumbing to a lot of the hype out there about things like job loss and “get rid of your entry-level people.” That’s all mundane drudge work that the partners and senior people don’t want to do—the account execs—so hand that all off to the AI and you don’t need to pay those people anymore.

And you’re exactly right. I was listening to a podcast over the weekend where they were talking about the same issue, but they were talking about it in the context of law firms. And they were making the point that the associates that are brought in out of law school do the drudge work that the partners don’t want to do. They write contracts, right? They do things like that. And now that the AI can do that, who needs them? Well, the question becomes: where do the future partners come from when the ones who are already at the partner level retire? There’ll be nobody to take those jobs. We are not rethinking the industry, and we’re not rethinking it from two perspectives. One of those perspectives is the agency. The other perspective is the in-house side of communications. They’re two sides of the same coin. But I think we need to split them apart and look at them in terms of how we need to reinvent the profession. You and I have talked about reinventing how we bill, how we price, because the hourly model makes no sense anymore. But what does an entry-level person do if the AI can handle a lot of that drudge work? And it can.

I mean, we’ve talked about on this show that I’ve set up a Hermes instance and it is out there. In fact, I haven’t checked my Telegram account yet, but there should be 10 links to recent news stories that are prime for me to news-check because I set up an agent to do that. I have an agent set up, a skill set up, that I can deploy anytime I want to. It is set up to analyze the websites of twenty-two of our competitors. And all I have to do is tell it what I want it to analyze. Do I want it to look at how they handle their project portfolios? Do I want it to look at how they handle their thought leadership? I can ask it any of those questions and it’ll come back and give me a very nice report. I could absolutely set it up to do media monitoring. I’m starting to question the need for my media monitoring service at work, although the agent that I have set up to do some of this certainly can’t get behind the paywall the way that the media monitoring service can, because they pay the licensing fee for all of those. So if the AI can assume all of this work, it’s not a question of saying we don’t need entry-level people. It’s a question of reimagining what entry-level people should be doing.

In terms of AI: What should they be doing with AI, and what new things can we be having them do that we haven’t thought of before, or that we always wished they could do if they didn’t have all of this drudge work that they had to spend their time on? It’s time for a reinvention, and I don’t see anybody talking about that. I haven’t seen a whole lot of ideas about where all this should go.

Neville Hobson: Yeah, I’m with you on that a hundred percent. Exactly my sentiment as well, that you don’t see people talking about this in a truly serious way. I see on LinkedIn—if that’s any barometer, I don’t know if it is or not—but I see people mentioning this now and again and “we ought to do something about that.” But there’s no webinars, no seminars, no get-togethers on the topic of reinventing the agency, let’s say. It’s a topic I’ve written about myself, and value-based pricing versus time-based pricing. And it’s interesting how Stephen Waddington addresses that topic in his article. It’s quite a pointed observation he makes that’s worth pushing on. If you’re still selling time rather than value, he says, AI will break your model. That’s a direct challenge to the billable-hour structure that much of agency PR still runs on. So the firms getting this right are building around insight, outcome, and risk management instead. It’s worth asking how many firms are actually making that structural shift versus just talking about it. Not enough. Doesn’t mean to say they’re ignoring it. Far from it. I think it’s largely because they don’t know what to do. How do they address this? So there’s an opportunity for someone with some insights and answers to help educate firms like that. There’s a consulting opportunity, if you like.

Shel Holtz: I was thinking exactly the same thing. If somebody’s looking for a pivot in their career, that sounds like one to me.

Neville Hobson: Yeah, yeah. So we are at that place. Again, go back just three years, 2023, when we wrote our pieces about that CIPR survey, and twenty-five percent of the respondents said they’d never ever use AI. It was pretty absolute, the answers. Here we are, three years later, and I bet you that number’s down to five percent, if not less. I can’t imagine anyone—and it causes a very broad question, “would you use AI, yes or no?” It’s a bit like “should we stay in the EU, yes or no?” I mean the Brexit referendum—well, people, what a dumb question. But so that’s where we’re at. But I believe a lot of the landscape is now so polluted with everyone’s opinion that it’s very confusing to zero in on what are the issues I need to be thinking about in an organization. Plus, I see so many people—I saw one just this morning—someone’s got a PDF book on how to move your business to selling value, basically, not time. And it’s not how many hours you did, it’s what did you deliver to the client.

So it’s great, but it needs to be more authority than that, I think. And this is where the profession comes in—professional bodies like the CIPR, the PRSA in the US. The CIPR has done a good job in raising awareness about AI in the right way, in context related to public relations. They’ve had this AI panel for some time now with senior practitioners leading it. This book’s come out and it’s got a lot of support from practitioners in the UK and beyond. So maybe now is the time that this is going to get taken a bit more seriously than people do. I think though what Stephen worries about—and I think it’s not a misplaced worry—is the point that people are being laid off. Layoffs are happening all the time and most people believe it’s because AI is going to be more efficient and all that kind of stuff. And there must be some truth in some of that. But he also mentions something quite interesting in his article, because he says that most of the conversation about AI and jobs focuses on redundancy risks from above—leadership cutting roles. We’ve talked about that quite a bit. But Waddington notes a quieter pressure running in the opposite direction. Junior and mid-career practitioners are walking out of organizations they consider too far behind the curve.

So firms that move too slowly aren’t just at risk of getting the technology wrong, they’re at risk of losing the people who could help them get it right. The talent drain is bi-directional. Now that’s an interesting element to bring into this discussion, I think—that it’s those folks who are walking away. He doesn’t say, and I hadn’t found anything before we started recording, as to where they’re all going. Are they leaving the profession entirely, or are they just looking for a place that—in a sense they feel it’s worth going to this company because they’ve got it switched on, that they’re clued into this? So maybe that’s the state we’re in. Doesn’t answer the questions, mind you, and they’re coming thick and fast now, I think. I see, again, LinkedIn is a kind of barometer of sentiment, if you will—not in the analytics way, but the feeling you see expressed in some posts from some people who are worth reading about it. And that includes many of the people that I follow and that you would follow as well. So you’re seeing this, but it’s all very random. That’s the thing. And it requires something more than that. And voices like Stephen’s, yours when you were talking about this—we’ve missed about three, four, five times, that sort of thing. What’s going to make people really pay attention to this?

Shel Holtz: I hate to say it, but it’s the same thing that has always made the industry pay attention, and that’s when they suffer financial pain. The reason we have not embraced as an industry these technological changes is our billings were fine. We were doing just fine as an industry financially. So why should we make this risky change to something that we don’t quite understand and we’re not convinced is going to have all that much impact or will necessarily stick around all that long? That leaves an opening for other industries—advertising and marketing—to sneak in. It also leaves an opportunity for boutiques that specialize in this to start up and take money off the table that was there for the PR agencies that were already in business. And this seems to be a recurring pattern: if we’re not feeling the financial pain, we’re not gonna make any change. As soon as we start to feel that pain, as soon as we see our clients going to the boutiques and going to the marketing agencies, then we go, “we better change.” And then we’re behind the curve. So I think that’s the big issue and the big challenge—to be proactive rather than reactive when these technologies create these opportunities, or create the requirement, if we wait, that we must change because we’ve already seen these revenues go to somebody else.

One thing to keep in mind: absolutely there have been layoffs within the industry and they have been attributed to AI. It is important to keep in mind though—and this was reinforced in that very same podcast I was listening to that I mentioned earlier—that if you look at economic data, there’s no evidence of mass layoffs as a result of AI. The unemployment rate is pretty much where it was before all of this. The number of new jobs that are being reported, at least in the US, has actually been pretty strong. The jobs report the last month was quite encouraging. So we keep hearing about the mass layoffs and they may be coming. They may not.

Because frankly, what I see—and I don’t know if this is unique to the construction industry, I doubt it; I think it may be a bigger issue in the construction industry, but I think this is probably true of most jobs—is it’s not the job that gets replaced by AI, it’s tasks within the job. And then there are other tasks that the AI can’t do. The other thing is that there are things that we have wished that we could do, but haven’t had the time to do, from an internal comms standpoint and even, I suspect, a PR standpoint from inside the organization, the client side. I mean, I remember when I was in my first corporate job. This was with Arco. I was there from ’77 to ’83 with some brilliant communicators, but the company believed in it. So they funded the internal comms department. We had 25 employees in internal comms in five cities.

And each of us had beats, just like you were a newspaper reporter with a beat. I had two beats. I had Arco Petroleum Products, which was the gas stations and the merchandising of cans of motor oil and things like that, and Arco Marine, which was the oil tankers that transported oil mainly from Alaska down to the refineries along the West Coast. And I spent time—I mean, that was my job, was to go hang out, to spend time, to shadow somebody, to do a ride-along, to ride on a tanker, to spend a day at one of the gas stations and really get a sense, and to be able to report on this a little more intimately than just calling somebody and doing an interview over the phone. And in public relations, I think it’s important to remember that “relations” part of the public relations label. How do you build relations? Well, if AI really does take away a lot of that drudge work that we spend the time doing while we’re sitting at our desk, then we have time to get up from our desks and go out and hang out with the publics that we are dealing with and build those relations. And why wouldn’t we want to do that? AI can’t do that. AI can’t get up, get their car and go to where the public is. Maybe it’s a community relations organization, maybe it’s a division of your business. Maybe it’s a customer base that is gathering—well, let’s say it’s Ford Motor Company and there’s a car club that’s meeting. Whatever it may be, we have the opportunity now to become much more entwined with those publics.

And do a much better job of understanding them. Yeah, we still want to do the data, we still want to do the analytics, but there’s nothing like sitting with them and looking them in the eye and talking with them to build an understanding that’s going to help you communicate with them and help you build trust among them. That’s just one idea of what we can do with this freed-up time. And this is an important point—and I saw this in one of the reports that came out just last week, I think it was—the value that we get from saving an hour because AI can do it just leaks out of the bottom of the organization if we don’t know what we’re going to replace that hour with that has value. And we hear about all the savings of time that AI is going to give us. I haven’t heard a whole lot about how organizations are figuring out how to reallocate that time among those employees.

Neville Hobson: Yeah, yeah, neither me. No, I agree. And you do hear a lot of talk about the concept of that. I mean there’s lots in this topic, Shel, really, and you’ve thrown some bright light on some of the things we should be doing. I like the idea of going out to meet your publics, as it were. It’s winding the clock back, actually, to how we used to do all this back in the day, before all this tech was there.

Shel Holtz: Yeah, it really is.

Neville Hobson: We had to go out and find the sources and interview them face to face and, you know, meet down the pub or whatever. So maybe we need to examine what worked in the past and bring it to the fore again.

Shel Holtz: When I was a newspaper reporter, before I made the switch to corporate communications, I was with a local community daily newspaper, and I used to go hang out at the bar after work where all of the government workers hung out after work. Got to know them, got to listen in, got some pretty good stories out of that. But also I could pick up the phone and call some of these people because they knew me. I wasn’t just the reporter who called when I needed a quote or needed some information. I was the guy they just had a drink with.

Neville Hobson: Yeah, exactly. Lessons to learn there, I think. So yeah, lots of good ideas here. I think Stephen Waddington did a good job in literally describing the landscape and expressing some of his concerns. That’s prompted this conversation. So let’s hope this adds to the topics that people need to be talking about. So listeners, hope this is helpful.

Shel Holtz: And listeners, if your organization is actually making some changes and doing some pivots, we’d love to hear about it. And that’ll be a 30 for this episode of For Immediate Release.

 

The post FIR #518: Is the PR Industry Blowing It Again? appeared first on FIR Podcast Network.

First, they were told to use AI. Experiment! Add it to your workflows! Go wild! Then the bills started piling up, and companies realized the cost was not tenable. Now the walk-backs are happening. Usage caps! Caution! Slow down! Among the issues communicators need to address is employees questioning leadership’s judgment. In this short midweek episode, Shel and Neville explore approaches communicators can take to help employees understand the pivot while maintaining the perception of leader competence.

Links from this episode:

• AI can cost more than human workers now
• Microsoft reports are exposing AI’s real cost problem: Using the tech is more expensive than paying human employees
• When AI Costs More Than the Worker It Replaced
• AI isn’t paying off in the way companies think. Layoffs driven by automation are failing to generate returns, study finds
• AI layoffs may be backfiring on companies
• Uber, Microsoft, and Others Burning Through AI Budgets. Now What?
• Uber burned through its entire 2026 AI budget in four months. Now its COO is questioning whether it’s worth it
• Uber Burns Its 2026 AI Budget In Four Months On Claude Code
• Sam Altman says OpenAI’s top token spender uses 100 billion tokens a month — and they’re not even the world leader
• OpenAI CEO Sam Altman admits AI token costs are becoming ‘a huge issue’ — company seeks improved value as overspending becomes a meme
• Token Billing Exposes AI’s Missing ROI And Puts Billion-Dollar Bets At Risk
• AI savings misses should make executives uncomfortable
• AI saves workers a day a week, but they don’t know what to do with it

The next monthly, long-form episode of FIR will drop on Monday, June 22.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville Hobson: Hi everyone and welcome to For Immediate Release. This is episode 517. I’m Neville Hobson.

Shel Holtz: I’m Shel Holtz. In some companies right now, that AI that was supposed to replace expensive humans is costing more than the humans it replaced. The numbers are kind of breathtaking. Uber burned through its entire 2026 AI budget in four months. In fact, I just heard today that they’re introducing a monthly AI spending cap of $1,500 per employee. One unnamed company, a real one though, spent half a billion dollars on AI in a single month because nobody had bothered to set a spending limit. An NVIDIA executive flat out admitted that for his team, compute now costs more than the engineers using it. A lot of these companies didn’t just overspend, they made decisions on the strength of what they thought AI could do. And in plenty of cases, those decisions cost people their jobs. The pitch was that AI can do this work for a fraction of the cost.

Then Bain and Company studied a thousand companies and finds that most aren’t getting those savings. Gartner found the layoffs delivered no better returns than not laying anyone off at all. In its study, Bain looked at the books and saw money leaking out of the top, companies spending the budget without the savings showing up. And Boston Consulting Group went and asked employees and found that the leak runs from the bottom too. Over 40% of regular AI users say they’re saving a full workday every week. But Boston Consulting Group’s point is that saved time doesn’t automatically become value. If nobody tells an employee where to redirect those reclaimed hours, that value just evaporates. So two consultancies looking at two completely different ends of the organization landed on the same diagnosis. This is a management failure. It’s not a technology failure. So put yourself in the shoes of employees who are still there.

They watched colleagues walked out the door because they were told the machine could do it cheaper. And now they’re watching leadership start to walk it back. In most cases, walk it back really quietly. What does that do to employees who see their leaders’ judgment, their competence? Because that is where this becomes a communication story. It’s about trust, credibility, and what we as communicators are supposed to do when leaders make a big public painful bet that doesn’t pay off. We’ll share our thoughts about that right after this.

Now there’s a lot we can talk about with this story, like communicating a suddenly altered governance model. But let’s start here. The bet companies made was about people, that AI could replace human labor at a fraction of the cost. When that turns out to be wrong, employees don’t just see a line item on a P&L. They see leaders who either didn’t understand the technology they bet the company on, or who used the AI story as cover for cuts they were going to make anyway, what we’ve come to call AI washing. Both of these readings are poison. The second one travels fastest. A communicator’s first job is to make sure that the accurate story is the one that gets out there. And to add a little context to that idea, Sam Altman, the CEO of OpenAI, just recently said that this cost discussion is new. It started early this year. Before that, nobody was talking about it. And that’s probably because before early this year, most employees were prompting AI chatbots, and that didn’t blow up budgets. What changed early this year? Agents. Now employees have agents running complex tasks in an endless loop, and that burns tokens like nobody’s business. I heard about one employee who burned through a billion tokens in a month. That means costs are exploding. It’s not something anybody really anticipated, and a lot of CEOs were caught unaware. You know, they paid for the subscription cost to say Anthropic, now they’re paying the subscription, but they’re also paying for tokens. So Neville, if you were leading a comms department in a company that laid off a bunch of people because AI could do their jobs, and now it’s either costing more for the AI to do those jobs, or the AI isn’t doing it as well as the people did, how do you communicate that without making leadership look like fools?

Neville Hobson: Yeah, it’s a good question, isn’t it, Shel? I mean, you’ve painted a picture that’s pretty dire, it seems to me. And I like to think that this is a kind of outlier territory we’re in. This is not the mainstream. But I’m willing to be proven wrong. You know, I don’t recognize this in the UK, so it could not yet be a big deal over here. But I’m thinking you mentioned that the original AI narrative, if I can describe it that way, was sold to people as we’re replacing you with robots or we’re replacing you with tech. I wonder, is that the case everywhere? Because I would have thought it was, many would have sold it as additive. It’s AI plus you, not AI instead of you. And that’s a wholly different kind of message if that were the case. Either way, they’re walking it back. And I think there’s a handful of things the communication leader should do. And that person would also be the counselor and the advisor to the leadership of the organization. So I think one of the first things, if not the first thing, is that you mustn’t let the leadership hide behind euphemisms. You know, restructuring, right sizing, optimizing for the future, that kind of stuff.

Employees see through all of that, particularly in this example. And it makes things even worse, I think. The communicator’s job is to push for plain language, even when that’s uncomfortable. So that’s one thing. The second, and again, this may well be the first. I’m just putting these out there as bullet points effectively without saying which is the most significant. Acknowledge the whiplash directly. Acknowledge it. Don’t pretend that the earlier message just didn’t happen. All those lovely messages about this rosy future’s coming when we’re introducing AI. You could say, we told you AI would be a productivity multiplier for everyone. We believe that. The reality has turned out differently, and you deserve an honest explanation of why. Of course, your next bit is so what are you going to say? But that’s the framing you need to do. Separate the human cost from the technology story. The people losing jobs are not a line item.

Communication that treats them as a budget adjustment will destroy trust with everyone who remains behind. And the people who remain are watching very carefully. Give the survivors something real, not platitudes about the future, concrete information about what the new operating model looks like, what’s expected of them, and what support they’re getting. And there’s a harder truth for communicators, it seems to me, and this kind of popped out of the woodwork sideways when I was looking into this. If you’re the comms person in a situation like this, and leadership won’t let you be honest, won’t acknowledge the contradiction, they won’t speak plainly to people losing their jobs, then you have a professional and ethical problem, not a communication problem. And that’s a different ball game. And if you’re in that situation, I would say, easy to say this naturally, is get your resume brushed up and start looking around someplace else. You don’t want to be in that kind of environment. So the best communication advice in the world can’t fix a leadership team that wants to paper over a broken promise. And that’s worth saying out loud. And I think these are the things that should be on the communicator’s to-do list in a situation like this that could, I suppose, go some way towards restoring or maintaining some level of trust in the leadership, where we do, we hate what you did, we’re angry at you, but we believe that you are willing to fix it in some form. And even saying like, yeah, I’m with you on that, you thought this would work and it didn’t. So there’s other questions that will arise too. But either way it’s going to be an uncomfortable journey to get to an outcome that you like.

Shel Holtz: Yeah, to your first point, no, I don’t think every organization is going through this. I think it’s the ones that are actually providing employees with a budget to use tokens. I mean, we heard the stories about the leaderboards that were encouraging people to rise to the top of the leaderboard for the number of tokens they were burning, because that would indicate, it would signal that they’re using AI, which the companies wanted them to do. But you know, the models, the frontier labs were not really forthright with their pricing structures on the tokens. This was something new. They introduced it. It’s not very transparent. Not all of them even have the ability to show you how many tokens you’ve burned through. A couple of them do, but it’s not clear what that means in terms of your costs or what you have left available. But you know, from the communication standpoint, yeah, the trap I see here is that the original decisions were announced really loudly.

Shel Holtz: They issued press releases. They had all hands meetings. It was discussed on earnings calls. The correction here is happening more like a whisper. Spend caps are being added quietly to governance language. I know Uber is out there talking about it, but not everybody. The leaderboards for those token maxing exercises are just vanishing without explanation. Contractors are appearing in company offices to do the work that the AI was supposed to do. Some of them are the employees who were laid off in the first place. So, you know, when you boast loudly and walk back quietly, employees are going to fill that silence, right? Information abhors a vacuum. And they’re going to fill it with the worst possible interpretation. But, you know, leaders do have this instinct to either double down, you know, this is what we said we were going to do, and by God, we’re going to do it. Or they go quiet, and both those approaches are wrong. I think what actually rebuilds credibility isn’t a groveling apology. It’s a clear here’s what we expected, here’s what the data actually showed, and here’s what we’re changing in response. I think that reframes what looked like a wrong bet into a process that we employed to arrive at the best outcome.

And it’s a continuum, and we’re here at the continuum. We’re not at the end game. We’re still learning and adapting and adjusting, and we still believe in the promise of AI, which by the way, I do. Look, employees don’t lose respect for leaders who update their thinking. They lose respect for leaders who pretend they were right all along when they weren’t. So we should counsel leaders to stop measuring AI adoption and hours saved, and start measuring whether that saved time is actually being reinvested into something that matters. We should never have been celebrating who can burn through the most tokens the fastest. In communication terms, that’s like focusing on which communicator can produce the most articles for the intranet and ignoring whether people are modifying their behaviors or reinforcing their support for company goals or whatever other outcomes you had determined that those articles were supposed to produce. Deciding what an organization celebrates, what it puts on the scoreboard, that’s partly our job.

Neville Hobson: Yeah, I’d say you’re right. It’s interesting. I think there’s a kind of an underlying question behind all this, it seems to me, that I was mulling over when I was looking at this. The real question is, can communications actually repair the damage that’s been caused by this? Can it? I think the honest answer is partially. So good communication can reduce the damage, preserve some trust with the people who stay and give the organization a chance to rebuild its credibility over time. But it can’t undo the promise. It can only help people understand what happened and why. Now that may sound like, well, in that case, this is doomed. No, not at all. Because if you get to that point, help people understand what happened and why, you’ve then got a foundation where you can build from, it seems to me. And that’s connected directly to your point you just made, that you have to, you know, grasp the nettle as it were, take the bull by the horns, etc., think of the metaphors, and be honest and truthful, fess up. It’s not like, yeah, we screwed up, not at all. We made the bet, we had all the research, everyone was convinced this would work, and indeed our vendors who persuaded us to sign up for all those tokens were saying the same thing.

There’s also another element. I’ve been reading about this in, I’ve forgotten, one of the US papers that I subscribe to, about this supposed huge AI backlash in the US that isn’t happening here. Yeah, it’s not happening here, although it’s a whole different landscape here in that context. Not on the scale that you have in the States. But I’ve been reading a bit about that. Maybe that needs to be factored into this as well, because that would fuel people’s anger, I think, at the outcome that they’re experiencing, particularly if there’s kind of iffy communication somewhere in there that doesn’t resonate with employees. So it’s a complex picture. But I think the real question is, can communications actually repair the damage? And I think the answer is precisely that, partially.

Shel Holtz: Yeah, you raise a really interesting point. If 70% of the population is feeling negative about AI, you have to figure that that is reflected in your employee population. And now you’ve got this piled on top of that. That’s just fueling those views. So you have to factor that into the approach that you’re taking to communication and maybe even into the employee profiles that you’re using to craft those communications on that whiplash that you were talking about. Important to remember that for, you know, like a year, employees were pushed. Yeah, there were mandates, those leaderboards I mentioned. They were pushed to use it as much as humanly possible, and now they’re being told, oops, use it carefully. Yeah, and I think communications does set the narrative that connects those two poles. You know, without it, the reversal just looks like leadership didn’t know what they’re doing.

Shel Holtz: And again, that reads as incompetent. So I think we do have to build the bridge. We learned something on the road. Here’s what’s changed and why. It doesn’t mean that the promise of AI isn’t still there. It’s just the road that we are taking to get there is a little more winding than we expected. And by the way, speaking of the whiplash, according to that Boston Consulting Group survey, nearly half of workers, 47%, say they spend more time managing and directing AI than doing the actual work they were originally hired to do. Think about that against what leaders promised, because the pitch was that this was going to be a labor-reducing thing, a productivity improver. And what employees are actually living is the labor didn’t get easier, it just got transformed from doing the work to supervising the AI that does the work. Boston Consulting Group found that four in ten reported an increased cognitive load, which is, I have to say, exactly what I have been experiencing. I’m not spending less time, I’m spending more time, and AI is part of the reason for that. So when the lived experience doesn’t match what was said in the announcement, the gap is where leaders’ credibility can go to die, right? It’s the communicator’s job to close that gap. Make sure that the story leadership is telling matches the work that people are actually doing and the experience they’re actually having.

Neville Hobson: Yeah, I would subscribe to that view. I’m also thinking, Shel, that, you know, the way this has been presented, let’s say, is as if, you know, lots of communication when this started, big promises from leadership, and then silence. And now there’s this, we’ve got to rethink all this and you’re not going to be able to do that. So there was no effective communication in the meantime, there was no updates on what’s going on and this is what people are doing and these are the benefits they’re having. So I feel there was a lack of continuity in the communication. Otherwise this wouldn’t have landed as a big surprise, or as big a surprise as it has to many. There would have been signs that might have prompted some to say, whoa, hang on a second. I’m getting this little note here saying I’ve used so many tokens. I mean it could even be, what are these tokens it keeps talking about? How was that communicated? That every time you do this, this is going to happen. And I’m thinking of my own experience as an individual and the experiments I’m running with Claude, for instance, that when I was experimenting with Claude Cowork, it would, or it might have been the project, I can’t remember which one it was now, but there was a little script I could put into the model that told me how many tokens I used. And it didn’t tell me that and then some kind of tech gobbledygook that I wouldn’t have a clue what the hell does that mean. It actually told me this project you’ve done used so many tokens, which is nought point one two percent of your total allowance until the next reset. Now that makes me think, that’s okay.

So, did they have anything like that in the organization that enabled people to just keep a running total on what they’re doing? You know, it’s kind of like the average MPG in your car kind of thing. Everyone knows that kind of thing. That’s helpful, even though it’s not very scientific or detailed, but it would be enough. And maybe I haven’t encountered it myself. I see people talking about this on LinkedIn quite a bit, where they are getting fed up with this tool telling me I’ve reached my token limit, and I’m thinking, okay, they’re obviously doing the kind of work that not everyone is doing that burns through tokens like this. And it’s quite tech oriented, those kind of comments. So you’ve got to take that into account. But I think it does come back to, from a comms point of view, that you told people this is what was going to happen, this is why we were doing this, and there’s this lovely rosy future for all of us. Did you explain the detail about how things were going to actually happen? Probably not, I would say. So otherwise there wouldn’t have been the backlash at the end. The backlash would have started earlier and you might have been able to head it off. You might have been able to make a difference. So these are all, you know, what-ifs perhaps, but nevertheless, it comes back to that underlying question, that can communicators actually repair this? The answer is partially, it seems to me.

Shel Holtz: Yeah, and I could definitely see myself burning through a lot of tokens without a tech rationale for it. If I set up ten agents to scan the media environment 24 hours a day, seven days a week, each with a different area of focus, and they’re just constantly running, that’s tokens being burned. That’s like the car never being turned off and going through the gas. So, by the way, imagine the message that companies can send to their employees if the company was more deliberate and slower about the implementation of AI. They know it’s coming. They’ve given Copilot to employees, but in terms of everybody going out and building agents and things like that, well, we’re still remediating our internal data to make it useful for an AI model. And we haven’t launched our training yet. No, they can rub their hands and say, you know, being slow and deliberate paid off. Look what’s happening to all these other organizations. We’re pretty smart leaders here in this company. By the way, one other communication opportunity I want to address before we wrap up, and that’s around governance, because there’s, you know, a remarkable detail in this story. The companies that overspent mostly already had the tools to prevent it. You know, spend limits like you set up in the agent that you had that told you how many tokens you’d gone through. Routing simpler tasks to cheaper models. I just heard this morning on a podcast that there’s something you can sign up for if you use their tools, it’s automatically going to route you to the right model for the task. So you’re not always using the high-end model, the reasoning model, the thinking model, that costs a lot. Budget caps, all of these rules have always existed as capabilities inside the platforms, the companies just never switched them on. So for a communicator, that’s a gift because it lets leadership say we’re putting real discipline in place without it sounding like punishing employees for using too many tokens, like they were initially told to do. The framing here is everything. The governance change needs to be framed as here’s how we make sure this actually pays off for us. And it lets employees know the company’s approach to AI is maturing. If you frame it as AI use is now restricted, well, that signals panic and confirms the exact incompetence story you’re trying to tamp down in the first place.

Neville Hobson: Yeah. Again, what was communicated, how was it communicated, to what depth was it communicated, and how did you know that people absorbed it well? So what kind of feedback mechanism did you have to see if people understood the message? But I think there’s also another element, this is an organizational one, which is, you know, you’ve got a hundred and fifty employees, let’s say you’ve got five hundred employees in your company, did every single one of those five hundred need unlimited token access to do anything they wanted? No, they shouldn’t have done. So the example you mentioned, a tech person or enabling someone to be monitoring something twenty-four seven, not everyone would need to have that. And in fact, you’d need to structure it so that only these guys who need to do that. And I would imagine you’ve done a comparison to say this is definitely going to be cheaper than doing it that way, which you’ve always been doing it. So that would make sense. I wonder how many didn’t even do those kind of, you know.

Shel Holtz: Many, if not most. Yeah.

Neville Hobson: Right, I wouldn’t be surprised. So perhaps we shouldn’t be surprised at this outcome that we’re discussing here. And again, that is now the fifth time I think I’ve mentioned this, that communications can only repair this damage partially. And I think the bits it can repair are going to have some good outcomes, I would say.

Shel Holtz: One other thing that communications can do here is to set the stage. It’s, this is not the end game. We have not reached the place where, now we get it, everything’s right. The message is anticipate more change from this. I mean, the frontier model companies, Anthropic, OpenAI, you know, they’re going to change their pricing models because they know that people are dissatisfied. They know that it’s opaque. They know that it’s costing more than people and companies can afford. So you can look for this to get changed. I just recently heard somebody suggesting an outcome-based pricing model as opposed to a token-based model. I don’t know how that would work, but it’ll be interesting to see what they come up with. But you need to prepare employees for more change. You can’t let them be surprised again. Look at what agents did to the way this has pivoted. There’s going to be more of this on the horizon. You know, prepare yourself, but in the meantime, we are still on the right path of figuring out how to apply this to the organization’s best advantage. And that’ll be a thirty for this episode of For Immediate Release.

The post FIR #517: How to Communicate AI Whiplash to Employees appeared first on FIR Podcast Network.

The Economist has gone public with an experiment: it has created a shadow website featuring an AI-friendly version of its front-of-paywall content. The idea is to improve the odds of this content surfacing in AI answers and responses to AI queries. It’s based on a new standard, llms.txt, which has been described as the robot.txt of AI. What does this mean for communicators? Neville and Shel break it down in this short midweek episode.

Links from this episode:

• The Economist tests AI-ready web pages
• The Economist prepares for a two‑track internet: one for humans and one for AI agents
• The Economist is testing content read by AI agents
• How The Economist is using AI to extend its global reach
• The next version of the web will be built for machines, not humans

The next monthly, long-form episode of FIR will drop on Monday, June 22.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Shel Holtz: Hi, everybody, and welcome to For Immediate Release. This is episode number 516. I’m Shel Holtz.

Neville Hobson: And I’m Neville Hobson.

Something quiet is happening to the web, and The Economist is one of the first major publishers to talk openly about how it’s responding. A piece published by Digiday last week describes how The Economist is building what its VP of generative AI, Josh Munker, calls two versions of the web.

One version is the one we’re all familiar with: richly designed pages, feature photography, navigation, everything optimized for a human reader browsing with intent. The other version is quite different: stripped back, structured around questions and answers, designed not for you, but for an AI agent acting on your behalf.

Now, if that framing sounds familiar, it should. In episode 515 last week, we spent some time on what Google announced at its developer conference in May: that searching the web will increasingly be done by AI agents rather than by humans, and that people will focus on acting on the information those agents surface rather than clicking links themselves. I made the point then that the question for communicators was shifting from, “How do we get found?” to, “How do we become part of the information environment that AI systems draw from?”

What The Economist is doing is a direct practical answer to exactly that question. And here’s what makes this particularly interesting. The Economist itself published a piece last December describing this shift in precise terms: a move from a pull internet, where people initiate actions, to a push model, where agents act unprompted, setting up meetings, flagging research, handling tasks, often without a human ever typing a query. They wrote about it then as an emerging phenomenon. Now, six months later, their own team is operationally responding to it. They’re not just observers of this trend; they’re participants in it.

The logic behind their approach is straightforward. A growing share of people, particularly in B2B contexts, no longer start their discovery process with a search engine or a home page. They start with ChatGPT or Gemini or Claude. They ask a question, get a synthesized answer, and may never visit the original source at all.

For a publisher like The Economist, that creates an obvious problem. If your content isn’t structured in a way that an AI agent can parse and surface clearly, you effectively become invisible. Not because your content is poor, but because the intermediary can’t read it properly.

So The Economist is experimenting. Right now, the focus is on content that already sits outside the paywall: marketing copy, B2B sales material, the kinds of pages where you want a potential subscriber or corporate client to find you. They’re building parallel versions: the polished human-facing page alongside a clean, agent-readable equivalent. The aim is to show up accurately and usefully in AI-generated answers.

Now, why does this matter to communicators beyond the publishing world? Because what The Economist is describing isn’t a publishing problem. It’s a communication problem.

And it connects to something that one researcher quoted in The Economist’s December piece put plainly: Marketers and communicators may need to pitch not to people, but to agent attention. The audience increasingly will be algorithms, and the humans will act on what these algorithms surface.

Think about your own organization’s public-facing content: press releases, executive bios, policy statements, corporate FAQs, product and service descriptions. All of that content is increasingly being read and summarized by AI agents before it ever reaches a human. If that content isn’t structured to be understood accurately by an agent, you lose control of how your organization is represented in AI-generated answers. And unlike a Google snippet, you may not even know it’s happening.

Alessandro DeSantis, a media consultant quoted in the Digiday piece, puts it bluntly. He calls agent optimization a defensive baseline, not a competitive advantage, but the minimum requirement to remain visible at all.

There’s a deeper question sitting underneath all of this, which we’ll get into: Who do you trust in the AI-intermediated world? What does it mean for the communicator’s job when the first reader of your content isn’t a person at all?

Shel, you and I discussed the Google side of this in FIR 515. Here’s a publisher responding in real time. What’s your take?

Shel Holtz: I have lots of takes on this.

This is, I think, a big issue. The first thing I want to point out is that, as I read the commentary of people who are talking about this, there’s an expectation that in the not-too-distant future, the AI version is all that we’re going to need to publish because we’re going to be publishing for AI as people rely on AI to get their information. I find this a troubling idea.

I think people are ignoring the fact that right now, 25 to 60 percent, depending on the nature of the site, of visits to a website are direct. They are not coming from a search engine. It’s somebody who already knows the URL. As I mentioned in a post I published to LinkedIn last week, nobody going to Amazon starts at Google and says “online retail site” and waits for the URL to come up. They just type Amazon.com.

There are a lot of people who know the URLs. There are URLs published in magazine articles, in advertising, in TV commercials, for example. And then there is the dark web: I send you a link by email or in our Slack channel, and you click it. There’s no search involved at all, so there is no opportunity to see that AI overview.

So I think we have to keep in mind that there are still a lot of people who are coming to our websites, not through Google or some other search mechanism, or starting with Claude or ChatGPT or Gemini or what have you. They’re coming directly to your website, either because they know the URL or it has been shared with them by somebody else. So I think we do need to keep that in mind.

The other reason I think we need to maintain our own websites is because we own them, and we don’t own that intermediary. You publish that Markdown version of a web page and you provide the proper router to it. Was it called LLM text, I think? They’re calling this the robots.txt of the AI era. And it’s going to share with the person who’s making the query what it shares. It may not be exactly what is on your page. So now you’re down to using a third party.

Neville Hobson: Something like that.

Shel Holtz: So, yeah, it’s good to have at least as a statement of record what your original content was. I have some other thoughts about this, but I’ll let you react to that first.

Neville Hobson: Yeah, no, I get it totally. Yet the trend seems to be quite clear. This is the way it’s moving. And I would say that, from what I’ve been reading, not just this, but The Economist is actually a probably good signal for what other media properties may or may not be doing or might want to do, depending on the outcome of The Economist’s experimentation. Being discoverable in this new way is critical; otherwise, you are invisible.

And I think the behaviors of people are gradually changing, not to the extent that it’s a groundswell. I don’t see it like that myself. Maybe it is generational, potentially, I suppose. But I think applying that kind of “it’s because of age differences in the different generations” is a bit of a tired argument nowadays, I think, as a kind of answer to everything.

We’re seeing a shift in how people not only find information, but how they think about going about finding information. So the idea of thinking, “Okay, let me just type into this search box: How do I do so-and-so? Where do I go to?” or whatever words you might use. I kind of talk in literally conversational language when I’m searching for something, and I’ve had more good experiences, way more than not good, with Google Overviews as a result.

And again, I sometimes stop and think, “What did I just do then?” I took Google Overview and I didn’t go any further than that because it gave me exactly what I wanted. And not only that, I’ve got a list of all the sources it looked at as well. So why am I going to spend any more time on this? I got what I need. So this is evolving fast, even though it’s still not totally clear exactly what’s happening.

And I think the idea — and this is my take on this, Shel — the idea of, as a matter of course, sharing URLs in an email or in a WhatsApp message with colleagues in the workplace, that ain’t going to be the way it’s going to be as a matter of course. There are going to be other, better ways of doing it than that, generally speaking, of course.

So this isn’t suddenly going to happen next week. This is going to take some time to evolve. But it seems from what I’ve read, Digiday being one good example, and a handful of others I’ve also looked at, this is the clear trend that’s evolving. And The Economist is clearly betting on their experiment that AI-agent content prepared for AI agents is an essential move for them to take as a publisher so that they do not become invisible. And I get that.

Shel Holtz: I agree that it’s an essential move. I’m not suggesting that organizations not do this. I’m just suggesting that they not abandon their websites. For now, we’ve got to keep that 25 to 60 percent that are coming directly to your website in mind.

You also have to keep in mind that your website may have things on it that can’t be duplicated in an AI Overview. In fact, the advice that’s circulating out there is that if you want to drive traffic still to your website, you need to have that kind of content. It may be calculators, for example, interactive elements that just can’t be summarized. You have to get to the website and interact with that tool.

I know that the ability to purchase from an AI Overview is coming. It’s not here yet. I haven’t heard a timeline for when it’s going to be here. It’s coming, but for now, if you want to make a purchase, you need to go to the website to do that. So I’m suggesting that there are still reasons to provide a website. You need to do both.

Now, when I say you need to do both, I think it’s important to understand that you don’t need to have a one-for-one of all of the content on your website. You need to make a judgment about what on your website you want to have summarized by AI and what is unnecessary or superfluous, or what you otherwise just don’t need to spend the time on. It doesn’t matter if the AI gets this particular content summarized in Markdown so that it can give you the AI summary.

So I think you have to do some strategizing. But undoubtedly, what’s coming from the web development platforms that are out there is the automation of this. I think it’s inevitable that WordPress and Wix and Squarespace and Webflow and all of the other services out there for creating websites are going to add the ability to automatically convert this page to Markdown and put it in the right place. I think that’s probably going to ease this. I haven’t heard any conversations from these website creation companies about this, but I do think it’s inevitable.

Neville Hobson: It is, I agree. One hundred percent, it’s inevitable. And like I said earlier, this is not sudden. So your point about keeping your website with links and all that stuff — for some years, that’s still going to be the case for many people, but it’s going to decline.

I think it’s interesting what you said about some stuff not lending itself well. I totally agree. And in fact, one of the tips for communicators that I came up with from all the stuff I’ve been reading is to go through your key external pages in your organization — the About page, FAQ, leadership bios, product or service descriptions — and ask, “If an AI agent were summarizing this, what would it say?”

And you talked about the examples you made: content buried in carousels, PDFs, JavaScript-heavy pages, or dense narrative prose is harder for agents to parse accurately. So restructuring even a handful of high-priority pages into clear, direct, quick Q&A formats is a practical first step. It could matter a great deal.

And the other thing, which I picked up from Digiday, but also from some others, is worth noting for everyone, I think. The technical landscape is still forming. Nobody has the full picture yet. But as The Economist example tells us, they’re not waiting, and neither should we as communicators. You start with what you can control: your own content, your direct audience relationships, your messaging clarity. And as Digiday concludes, the organizations that experiment now, even tentatively, will be better positioned than those waiting for a definitive playbook that may never arrive.

And I think if I were doing this — and again, it’s easy to say, “Well, if I were doing this, I’d do X.” It sounds easy, doesn’t it? — but the idea of literally kind of duplicating your website for this purpose is quite a bit of work. But it could stand you in good stead where your competitors haven’t done that and they’ve still got these beautifully designed, laid out, immaculate photos and stuff that’s animated, designed for human readers, without the version designed for the algorithms.

And I think that, to me, would be a key thing to truly investigate thoroughly in your organization: What am I committing to if I say, “Okay, our gorgeous website wins awards for its design, and here are all the metrics showing all the leads it delivers and all that kind of stuff. What do we need to do to get on the radar of the AI agents version?”

So you build a version. What’s entailed with that? One thing I didn’t see in Digiday’s piece, and I haven’t found it anywhere else yet, is what exactly is involved in creating a duplicate of your website designed around this? Where, for instance, you’re treating everything as structured data, not just copywritten text, which doesn’t translate well to an agent-intermediated world, to quote Digiday.

So it talks about, you know, starting to think about your key messages as discrete, unambiguous statements that can be lifted and represented accurately without context. That doesn’t mean dumbing things down. And this bit I did like: It means being deliberate about what an agent will extract if it only gets 30 seconds with your content.

Now, there’s a good pointer that the way of thinking about how you present yourself online now isn’t about beautifully designed websites and the nuanced meaning of words and phrases that are emotive about your products or service. It’s about data. And you need to appeal to — and this is a human emotive statement, which isn’t really applicable — but you’ve got to appeal to the AI agent. What will appeal to an AI agent? It doesn’t care, using the word “appeal.” It’s what will make it index you and include you when someone types, speaks, or whatever it is that produces search results, where the engine or tool you’re using goes out and pulls in all the information. You want to be sure you’re included in that. So it’s a bit of work to do.

Shel Holtz: Yeah. It’s not a one-to-one copy and paste from your existing website. In fact, we know that AI likes Q&A, for example. If your website’s not already in a Q&A format and the text can lend itself to that, you’re going to want to repurpose it that way.

We know that it likes what’s in the top 30 percent of the content, so you’re going to want to front-load the most important information. And frankly, if you’re not doing this, then your company is not going to show up in those AI summaries, or at the very least, it’s not going to show up represented the way that you want it to.

I do have two other issues to point out. One of these is the increased use of agents. You mentioned an agent, but it’s an agent that’s working on behalf of the model. Now I’m talking about the agent that’s working on behalf of me. In which case, that LLMS.txt file is going to become the routing layer to your official content.

And this is a big deal for regulated industries, because people have to find the canonical information. They have to find the approved information, not necessarily the one that was created for AI. So if you’re doing this, you do need to make sure that what you are creating for AI is official. It is going to present the information that the lawyers have approved, that is going to pass muster with the regulators.

The other thing is that if I can rewrite the text of a web page for the AI version of this that I’m creating, I can put something completely different on that page. And this becomes an ethical issue that we need to watch out for. People might try to game this. I mean, remember the early days of SEO and all of the gaming of SEO that we saw with tags that had nothing to do with what was on the page. They just knew that this was a popular search term, and it would drive traffic there.

We can see the same kind of gaming with this LLMS.txt routing to the AI-compliant pages, with information on it that’s there to get you into the search results, there to get you into the AI summary, not necessarily to give you the same information that you would have found on a web page. So this can be problematic, and I think it’s something that the industry’s going to be watching for.

Neville Hobson: I agree. So you compile your list of “these are the things we need to pay attention to,” but again, in the back of your mind, you’ve got this reality: The technical landscape is still forming. No one has the full picture yet. That would include what are the requirements of this, that, or the other. So assume it’s going to be like we currently have, although I would argue it’s not going to be like we currently have. It’s going to evolve, but that’s going to take time. So you need to do that precisely.

I would say pay attention to experiments like this. And again, this is a media company doing this, so they are publishers of content that might mention your brand. That’s why it’s important to you that you show up in these kinds of search terms as well. So a lot to pay attention to here, Shel, I think, really.

Shel Holtz: Yeah, and remember, The Economist isn’t doing this with their articles yet because that’s still behind the paywall. This is all that stuff in front of the paywall. I mean, eventually it’ll get there to some degree, although I don’t know how you do this behind the paywall when it’s just showing up on, you know, Google Search or…

Neville Hobson: No, again, the technical stuff is not worked out yet. They’ll find ways here.

The other thing to mention — I’m glad you mentioned The Economist. I meant to mention this, but the final point in Digiday’s piece, I thought, was really interesting. It’s a reminder, I suppose, that for all the experiments, The Economist sees AI as infrastructure rather than authorship: something to speed up research, workflows, and product delivery, not a shortcut to churning out more copy.

So AI is not going to be writing their content, they say. It will do all this stuff at the back end, if you will, that would enable its journalists, its human journalists, to do the stuff that they do well. That’s admirable. And I think The Economist has a track record of that kind of behavior in its history. So it’s great.

I think it’ll be interesting to see what other major properties do, including the big ones in the U.S., the big ones here, and elsewhere in the world. And I bet you they are working on stuff like this, too. They have to be.

Shel Holtz: No doubt. And that will be a 30 for this episode of For Immediate Release.

The post FIR #516: Your New Shadow Website appeared first on FIR Podcast Network.

Employees at the Pentagon have spun up over 100,000 AI agents. In the private sector, we’re seeing reports of 10,000 or more agents being deployed by employees at a variety of companies. The problem is that most organizations lack governance to address agents, and the problems this explosion of agents operating on employees’ behalf can cause are innumerable. In the long-form FIR episode for May 2026, Neville and Shel delve into the rise of agents, the harms they could cause, what companies should do to ensure these agents deliver benefits rather than problems, and how communicators can take a leading role in addressing the issue.

Also in this episode:

• AI copyright lawsuits are coming for communicators
• Google’s search overhaul could signal a post-citation era
• Placing your thought newsmakers, thought leaders, and subject matter experts on podcasts is becoming a standard media relations practice
• “I worked all weekend” is no longer an argument for the fees you charge
• Short-form video clippers are creating go-to content from long-form videos — including yours

Dan York outlines the big enhancements in WordPress 7.0

Links from this episode

• Slopaganda, the New Rules of Narrative Warfare
• AI Copyright Lawsuits Pose Growing Risk for Communicators
• Practical Considerations for Managing IP Risk in AI-Generated Content
• Best Practices for Mitigating Copyright Risks in AI-Generated Content
• AI in Litigation Series: An Update on AI Copyright Cases in 2026
• Powered by A.I., Google Changes Its Search Box for the First Time in 25 Years
• Google Search as You Know It Is Over
• Why Podcast Placements Are the New Press Coverage
• Including Podcasts in Your PR Strategy
• Podcast Guesting vs. Traditional PR: What Works in 2026
• U.S. Newsroom Employment Has Fallen 26% Since 2008
• U.S. Podcast Consumption Reaches Record High: The Infinite Dial 2025
• You Can’t Beat AI.
• Steve Rubel on AI, Media Analytics, and the Future of PR
• AI and the End of Billable Hours
• The Clipping Economy: How Short-Form Video “Clippers” Are Overrunning the Internet
• How Short-Form Clips Took Over the Internet
• The Clipping Economy
• The Case for and Against Clipping
• Inside the “Clipping Farms” Driving Fintech’s Marketing Boom
• Companies Have a New AI Problem: Too Many Agents
• Businesses Will Have Over 150,000 AI Agents by 2028, Says Gartner
• AI Agents Introduce a New Class of IT Management Challenges
• Why Most Enterprise AI Agents Will Fail — And What Leaders Are Missing
• How Smart Governance Can Contain Agentic Sprawl
• Six Capabilities Enterprises Need to Scale Agentic AI in 2026
• Pentagon Workers Vibe-Code 100,000 AI “Agents” to Use on Unclassified Networks

Links from Dan Yorik’s Tech Report

• Turn Your Blog Posts Into Podcast Episodes
• Google Search as you know it is over
• Google Changes Its Search Box for the First Time in 25 Years
• WordPress 7.0 Field Guide
• AVFTCN 040 – Returning From A Hiatus, and Plans for 2026

The next monthly, long-form episode of FIR will drop on Monday, June 22.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville Hobson: Hi everyone and welcome to the For Immediate Release podcast long-form episode 515 for May 2026. I’m Neville Hobson.

Shel Holtz: And I’m Shel Holtz, and we have six really interesting reports to share with you today. And not all of them are about AI. I’m not saying most of them aren’t, but a couple are on other topics of interest to communicators. Also have a really excellent report from Dan York looking at the latest upgrade to WordPress, a massive upgrade, one of the most significant upgrades WordPress has seen in some time, and Dan’s report is fascinating as he talks about this. But we are going to start by filling you in on a new podcast on the FIR Podcast Network. We haven’t had a new show on the network in a while. You know, we started this as just FIR and we needed a place to house multiple FIR shows. Those who have been listening a long time may remember FIR book reviews and FIR speakers and speeches. And we had a number of these. And then we had some people say, hey, can my podcast live on your network? And we said, as long as it has something to do with communications, sure. So all of them have pretty much faded except a couple that Chip Griffin continues to crank out, but now we have a new one. And the reason we have a new one is because I’m doing it as a new podcast by me and my longtime friend and colleague, Steve Crescenzo. And it is called On the Same Page. It is an internal communications focused podcast. We’re recording it twice a month, about 20, 25 minutes per episode. And each episode focuses on an element of the strategic internal communications framework that I developed. It was several years ago. It was actually before I took a job in the private sector again. I’ll have been at the company I work for now nine years in October. So yeah, I developed this a long time ago. Then I wrote 28 blog posts about it. Somebody said, turn it into a book. So I did. And I have found a publisher for that book. So the podcast and the book are companion pieces and the first episode of On the Same Page is out now. You can find it on the FIR Podcast Network or wherever you get your podcasts. For me, that’s Pocket Casts. For you, it may be Apple Podcasts or any of the countless places that you can get podcasts these days, including Spotify. So we’re there. We’re on Spotify, as is this podcast For Immediate Release. And speaking of For Immediate Release, Neville, we’ve done a few episodes since the last long-form monthly episode.

Neville Hobson: Yeah, we have published quite a few, four in fact, over the past month. All of them have listener comments and that’s unusual, pleasantly so, every single episode with comments. So let’s run them down. So in the long-form, episode 511 for April, on the 27th of April, our lead story was the tale of law firm Sullivan & Cromwell and how in spite of AI policies, training programs and guardrails, a public document went out filled with AI-generated errors. No human in the loop, seems, as we explored what went wrong and what it means for organizations trying to operationalize AI responsibly. We also talked about new evidence that AI is not leveling the workplace, why AI is raising the bar for PR, the rise of slopaganda, and much more. Comments, Shel, on that one?

Shel Holtz: Yeah, a couple, and a couple of responses to the comments. We’ll start with Wayne Asplund who said, for what it’s worth, I have a way I tend to think about this. Simpler work, smarter decisions, better customer outcomes. You should be focused on all three when working with AI. If not, you’re reducing quality and raising risk. The obvious example is what you talk about. If your only care is speeding up work and you’re not focused on the customer outcome, which describes many people, magnify the risk and reduce quality. You reduce the smart side of things too, because you’re not engaging with the content. The model works for all sorts of AI activities. And Neville, you had a response to that. You said, simpler work, smarter decisions, better outcomes is a good test of whether AI is actually adding value or just accelerating activity. The law firm example feels like a case where the first of those dominated: speed and efficiency, but the other two didn’t hold. And that’s where things break down. I particularly like your point about engagement. If people aren’t really engaging with the content, then the smarter decisions part never really happens. It just looks like it does. The balance you described feels like something organizations need to make much more explicit, not just assume. And then we had Philip Weiss who loved the word slopaganda. My thoughts on this are from an earlier post, and he shared the post that he wrote called Slopaganda, the New Rules of Narrative Warfare, and the link to that will be in the show notes. And Wayne Asplund replied to Phillips saying, “some of the creative terms coming out at the moment. One newspaper I read stoked the fear of job losses by talking about the job- the job-pocalypse. And another article I saw used the term e-idiots.” Definitely going to pinch that one.

Neville Hobson: Ha Excellent. Great to have.

Shel Holtz: And by the way, the idiot’s comment reminds me of a T-shirt that I saw the other day that said any fool can use a computer, many do.

Neville Hobson: Yeah, that’s good. Yeah, it’s good when we see that kind of engagement, comment discussion taking place in the all on LinkedIn, of course, right?

Shel Holtz: yeah, yeah, although I think I got one or two of these actually off of Facebook this month. I can’t remember which ones, but.

Neville Hobson: cool. Excellent. OK. So our next show, we asked the question, when does AI stop supporting decisions and start shaping them? In episode 512, published on the 4th of May, we explored a provocative idea that AI is becoming the new executive influencer. Some research suggests leaders are already relying on AI for the majority of the decisions. But what does that really mean in practice? We have comments, right?

Shel Holtz: Yeah, Vincent Brunel left a comment saying, accountability question is the one that keeps me up at night when AI shapes the outcome, who actually owns the decision? The algorithm suggested it can’t become the new, I was just following orders. And Dean Landaish, I’ve known Dean, God, I can’t tell you how long through IABC. Having done lots of genealogy work with AI, its use for policy and decision guidance frightens me. Even with safeguards in place, AI systems will eventually tell you what you want to hear rather than facts or even even-handed evaluation. Like the social media algorithms, it eventually feeds you what it thinks you want based on patterns. Aside from writing progressively more restrictive prompts to stop this, I’ve had to go so far as erasing and dumping my user history. I wonder how many neophyte business users will build in proper safeguards to prevent the bias creep that AI builds over time.

Neville Hobson: And next, a Harvard Business Review article about redesigning marketing organizations for the agentic age provided context in FIR 513 on the 11th of May as we explored what happens when AI moves beyond generating content to orchestrating workflows and organizational systems themselves. We discussed examples of using Claude CoWork and AI agents in communication workflows. Work that once took hours now happens in minutes. Comments, right?

Shel Holtz: David Bradfield from up in Toronto. Great episode, Shel and Neville. The shift to the agentic age is less of a tech hurdle and more of an accountability crisis. If a communicator can’t hold an autonomous agent accountable to the brand’s ethical and strategic guardrails, the operating model isn’t just slow, it’s dangerous. I’ve just launched Avalere Advisory, which is very much aligned to the opportunities you cover. The narrative code only works if you have the right cultural upskilling and governance to match. Moving managers from task supervisors to agent auditors is key. I love the focus on the narrative code. It’s something we were building into my last corporate role for our CEO and other executives. It’s the backbone of the unified model communication teams need to succeed. And another comment from Vincent Bruneau who wrote a year ago, this was demos. Now it’s workflow design and organizational architecture. The speed of that shift is what makes the experimentation point so critical. The gap between those who are testing and those who are watching is widening fast.

Neville Hobson: And finally, was Twitter a unique moment in media history rather than a repeatable model? That’s a question we asked in episode 514 on the 18th of May when we explored the growing arguments that text-based social networks are entering irreversible decline, not because text itself is disappearing, but because giant algorithmic public feeds may no longer fit how people want to communicate online. And comments, One comment.

Shel Holtz: One comment from Mark Hillary, who says, of smaller rooms is a good analogy. I have my friends and family on various private WhatsApp groups. This now serves the same function that social media used to perform. As social media becomes more more aligned with entertainment rather than connectivity to a network, everything changes for people who want to communicate to a wide group of people.

Neville Hobson: Brilliant. So now you’re up to date on FIR episodes.

Shel Holtz: Yep, that’s right, and 515 will get underway in earnest here in just a minute, but first I do want to let you know that Circle of Fellows is coming up, the May episode this coming week. It’s the second of our special episodes with Brad Whitworth moderating a panel to discuss the remaining chapters of the book, The 7 Cs of The New Communication Compass. The book’s author and editor, Dianne Chase, is going to join Brad along with the remaining fellows who wrote chapters who weren’t part of the last episode. That includes Zora Artis, who wrote the Cohesion chapter, Cindy Schmieg, who authored the Collaboration chapter, and me. And I actually was the moderator last month. This month, Brad is doing it so I can be there to talk about the chapter I wrote on community. The panel is going to be live streamed at 5 p.m. Eastern Daylight Time this coming Thursday, May 28th. Participants of the live stream will be able to ask questions, share comments, observations and experiences and be part of the discussion. But if you can’t join us live, there is, of course, the audio podcast and the YouTube replay coming up after the episode has been recorded. So as I said, we’re going to jump into our first report dealing with copyright right after this. I have two reports today that have nothing to do with AI at all. I hope that’ll be refreshing, but let’s start with AI. According to the law firm Davis and Gilbert, 99% of public relations firms are now using AI. That’s a pretty dominant number and it could be a rounding error that would take you up to a hundred percent, I suppose, or maybe there was one company that said, uh-uh, not us. But the top use cases should come as no surprise. They’re using it to write content, take notes and summarize meetings, spark ideas and monitor media. In other words, AI is touching just about everything communicators produce. And while we’re busy adopting these tools, copyright laws have been piling up. Now the specifics I’m about to cover are based on US copyright law, but copyright is a thing in most of the developed world. So even though these cases may not apply in the UK or Europe, principles, at least some of them probably do. So let’s look at an update from the law firm Norton Rose Fulbright published in March that walks through six of the most important AI copyright cases working their way through the courts right now. First, there’s Thaler versus Perlmutter. The Supreme Court denied cert on March 2nd, which means the rule is now settled, at least for now. Purely AI-generated work cannot be copyrighted in the U.S. Period. If a human didn’t, and here’s some core language to keep in your mind, if a human didn’t make a meaningful creative contribution, you cannot and do not own it. In Bartz v. Anthropic, the court ruled last June that training on copyrighted books was fair use, but storing pirated copies was not. Anthropic settled for $1.5 billion, the largest copyright settlement in U.S. history. The fairness hearing happened just this past May 14th, like a week and a half ago. The judge didn’t rule from the bench, she took it under submission, but observers expect final approval any time now, and authors are looking at roughly $3,000 per work. And by the way, I have applied for that relief for several of the books I’ve written. And if I get it, Neville, you’ll get a piece of that three thousand for the How to Do Everything with Podcasting book.

Neville Hobson: ha ha! Twenty years after we wrote it. Yeah, okay. Wonderful. Wonderful.

Shel Holtz: That’s right. But yeah, it’ll still earn you 1500 bucks if this all goes through. Yeah, in Kadrey v. Meta, we saw the same fair use conclusion on training, but a different judge in the same district disagreed with the Bartz judge on market harm. He basically told the plaintiffs they brought the wrong argument and signaled that a future plaintiff who actually builds on

Neville Hobson: Excellent. Look forward to it.

Shel Holtz: An evidentiary case on market harm, showing that AI-trained models flood the market and depress demand for the original works, that could win. Then there’s Disney and Universal versus Midjourney, where the studios are alleging willful infringement of their characters, Elsa, Shrek, Darth Vader, the minions, and ask for statutory damages of up to $150,000 per copyrighted work, plus an injunction and disgorgement of Midjourney’s profits. Now, here’s where the story has often gotten more interesting. Since Norton Rose published their article pointed to the Disney OpenAI deal, the one where Disney was investing a billion dollars and licensing characters for OpenAI’s Sora video app as a sign of two things. First, that formal licensing was becoming the path forward. And second, that every signed deal strengthens the case against AI companies that didn’t license because it proves a licensing market exists, which is exactly the kind of market harm courts look for in fair use cases. Well, on March 24th, OpenAI announced it was shutting down Sora. The consumer app went dark on April 26th. The API goes dark in September. And Disney’s billion-dollar investment, the so-called bridge between Hollywood and AI never closed. According to Deadline, a Disney insider said simply, the deal is not moving forward. So one of the most cited pieces of evidence that an AI licensing market was emerging, it just evaporated. That makes the legal terrain more uncertain for communicators, not less. You might still be tempted to say that’s an interesting legal story, but it’s a problem for the AI companies and law firms. Not for me, not so fast. A PRSA panel reported in a recent piece on the PRSA site titled, copyright lawsuits pose a growing risk for communicators, quoted Samantha Rothaus, a partner at Davis+Gilbert, who made the point that for communicators, accuracy in AI-generated content isn’t just about looking smart. Inaccuracy, she said, can be misleading and deceptive, and that creates regulatory and legal risk. Another panelist, Michael Lasky, said he sees significant gaps in AI policies and governance within the PR field, and those gaps create significant risk. And Debevoise & Debevoise & Plimpton article from early March written by attorneys who advise marketing and communication clients lays out exactly where that risk lives for us, for communicators, not for the AI labs. And it lives in three places. One, ownership. If your team uses AI to generate a campaign asset, a logo concept, a hero image, an ad headline, and there’s no meaningful documented human contribution, you can’t copyright it. You don’t own it, and a competitor could pick it up tomorrow and use it. And there’s nothing you could do about it. Two, vicarious liability. Debevoise, and I know I’m not pronouncing that name right. It’s the name of a law firm. They flagged this and it’s worth understanding. If you’re using a third-party AI tool, and you are, and that vendor’s model was trained on infringing material, courts have already held in non AI cases, that the company using the tool can be liable too. The rule is roughly this. If you benefit financially from someone else’s infringement and you had the contractual ability to control how they were operating, you can be on the hook, even though you didn’t do the infringing yourself. So the AI provider’s training practices and your indemnification language are now your problem too. Have you read your AI vendor agreement? Do you know what’s indemnified? What’s excluded? And third is output infringement. If an AI tool produces something substantially similar to a copyrighted work, even unintentionally, even because of a sloppy prompt, and you publish it, you can be sued. One analyst made the point that images are especially exposed. Reverse image search trolls are already a cottage industry, and AI is just going to feed them. So what should communicators be doing? Three things, none of them particularly unusual. First and probably most important is to build an IP clearance step into your AI workflow, the same way you’d clear a stock photo image or a freelancer’s contribution. Document the human contribution to anything you actually need to own and read your vendor contracts with your legal team, specifically for training data sources and indemnification carve-outs. The legal landscape is going to keep shifting fast, as the Sora story demonstrates. Communicators who get ahead of it now are the ones who will be in a position to keep using these tools aggressively. Everyone else is going to be playing defense, and defense is a really bad place to communicate from.

Neville Hobson: quite a picture. So I think that actually makes it adds a significant dimension to what we’ve already been talking about, not just you and I, but many, many people we have on this podcast. On the example I mentioned in an episode recently, we talked about the law firm that submitted, it was actually a document to a court in a bankruptcy case. It was a corporate bankruptcy, with big numbers involved, riddled with errors because no one had checked it. So that’s been the focus: human in the loop has got to check it. It’s deeper than that, There’s your kind of, in a sense, your obvious level of foundation of the things you need to do to ensure that you’ve done everything you possibly can to ensure that what you produce using an AI assistant to help you. It doesn’t contain errors that can trip you up because you didn’t check it. Here we got something that makes that even worse sounding, frankly. What troubles me most is the statistic in the PRSA’s piece, quoting the law firms report they did that 99% of PR firms are using AI. And we already know from other reporting that, you know, many of those, I’m not going to say most because I don’t know, don’t have the governance in place to protect them. Do they have a human in the loop? I’m getting very cynical about hearing that phrase because I hear it like it’s a nice thing to say and there’s absolutely no substance behind it whatsoever because we’re seeing evidence and prominent examples. So there are others that aren’t so prominent, but they’re happening where they don’t have a human in So we’ve talked about this before. You had an idea a few episodes back that you need to have verification specialists. Maybe it had got to come to that because not everyone has the ability to verify. And even if you do, what’s the legal implication of that person? Is that the responsible person on behalf of your firm? What if they’re the intern doing that? It’s got to have someone with authority to do that kind of thing. So 99% using AI, the top reason. writing content, 79 percent. That really does make this sound very serious indeed, that you could run into deep trouble. Now, I know you said this is the examples you gave under US copyright law and copyright law is jurisdictional and it’s based on geographies. It hasn’t changed any in hundreds of years in that regard. So if you’re a global firm operating in 20 countries, you need to understand the copyright law in each of those 20 countries. So the risks are high, I would say. And you painted some dark picture there of here are the potential consequences if you don’t do this. But this needs to be part of your thinking. You mentioned as well, reverse copyright trolling to find stuff that could land you with a request of payment for the copyright infringement, stuff like that.

Shel Holtz: Yeah, the digital version of ambulance chasers.

Neville Hobson: AI-generated images. Right. But that’s part of the landscape, I guess. So it’s something to take seriously. I did note the concluding part from the PRSA story where they’re talking about this same law firm did a survey last autumn. Thirty-seven percent of PR companies they surveyed were developing their own closed proprietary AI systems specific to their clients. Will that protect them? I think it gives a false sense of security because I doubt it would. So you can’t think, I don’t need to use, you know, take your pick, Claude, Gemini, chat GPT or whatever. Even I create a bespoke version of the LLM, I’ll build something unique. If you haven’t got safeguards in for this, it won’t matter. You’re still going to run into trouble. So this is huge, I think. I can see lawyers everywhere rubbing their hands with glee because this needs legal input to this in each of those different jurisdictions of the copyright law.

Shel Holtz: Yeah, I’ll give you two quick examples just from my own experience in recent days. And by the way, Chris Penn, this goes back a couple of months, but he did publish a post on LinkedIn that gave you a prompt to use in order to check your output for copyright violations. And he said, it may not catch absolutely everything, but if you’re sued, you could pull up the fact that you did that as a good faith effort to ensure that you weren’t violating copyright. It could stand in your favor in a lawsuit, but presumably most of the instances it would actually catch. But these two instances that I’ve experienced recently, was a search I was doing. I use AI for tech support all the time. I find it better than tech support. And I was trying to resolve something. I won’t get into what it was, but it said, if And the language it used was very specific, right? It said, if you’re losing sleep over this, stop. Google has said that they understand that this is a bug. They’re aware of it. All of us are experiencing the same thing. They’re working on it. It should be fixed in a day or two. So I was looking for more information on this. So I went to Reddit and I did a search and there I found exactly that same language in a post that somebody had left in the discussion for this particular software product. So it’s not that Google Gemini had taken that information and put it in its own words. It had actually taken those words from whoever wrote that post and shared it with me as a response to a prompt on Google Gemini. So yeah, if that were copyrighted, that would be copyright infringement right there. And if I quoted it, that would be problematic. The other I shared on LinkedIn, I thought this was a really interesting story about somebody who uploaded a photo, an image and said, is an AI-generated image that I created. I prompted the model to use the style of Monet. And there were thousands and thousands, according to the article, there were like 1.6 million people who said, this is AI garbage, it’s slop, it’s horrible. Here’s what’s wrong with the art, the lighting, the brushstrokes. And then it was revealed it was actually a real Monet. It was not an AI-generated. He just said it was. And I shared that on LinkedIn, got some interesting comments. But one of the comments I got said, all AI-generated art is theft, presumably because all AI-generated art was trained on real art, just like all AI-generated text responses is based on text that it pulled off of the net. But with people having that attitude and the person who shared that comment isn’t the only one I’ve heard say that. You know that there are people out there who are looking for this. As I said a minute ago, the ambulance chasers of the digital era are the ones who are out there looking for something actionable and then going out and finding somebody to represent in a lawsuit based on that. So copyright and AI, especially in this period we’re going through right now with AI backlash in general, I think you’re going to see this activity ramp up. And I don’t see any reason why corporate communications using AI can’t be caught up in some of these legal actions.

Neville Hobson: There is plenty to think about in that case. So moving on to our next topic, if you’ve been following how AI is reshaping search, and I’m sure many of you have, we’ve talked about this on FIR before, including our discussion a few months ago about GEO, Generative Engine Optimization. Then what Google announced at the developer conference recently in May is the next chapter. And it’s a significant one. Two pieces caught my attention amongst the many reporting about this. Sarah Perez’s writing in TechCrunch and a report in the New York Times by Tripp Mickle, Kate Conger, and Brian X. Chen. Both cover the same announcement, but from different angles. And together, they paint a picture that communicators really need to sit with. Here’s what Google announced. The search box, that iconic long slender bar that barely changed since 2001 is being overhauled for the first time in 25 years. It is getting bigger, more conversational, and more interactive. But that’s almost the least interesting part. What’s really changing is what happens after you type. Instead, or you even speak, actually, now, instead of a list of links as part of an AI overview, Google will increasingly serve you AI-generated interactive experiences, custom visuals, dynamically built mini-apps, and what they’re calling information agents that work in the background around the clock, tracking the web on your behalf and alerting you when something relevant changes. TechCrunch makes the point starkly. Searching the web will increasingly be done by AI agents rather than humans. People will focus on acting on the information those agents surface rather than clicking links themselves. The New York Times goes further with a quote from Richard Kramer, a financial analyst with Arete Research. He said, and I’m paraphrasing here, that Google is reducing everyone to raw data providers. The open web, said, is on the way out. Now think about what that means for communicators. We spend years adapting our content strategies, first for SEO, then for GEO, trying to get our content picked up and cited in AI-generated answers. But what’s the strategy when the answer isn’t a list of links, isn’t even a text summary with citations, but a dynamically generated interactive experience that never surfaces a source at all? That’s the question I want to dig into. We’ve moved from SEO to GEO, and now we may be moving into territory where there’s no citation game left to play. I’ve got some thoughts about that. But first, though, what do you think, this means for communicators? And what should they actually be doing differently right now?

Shel Holtz: Well, I think communicators need to adapt. We always have. So has the marketing world and the advertising world. When Google first came out, I mean, you know, first of all, I have to say, I think this idea that it’s the end of the open web is utter nonsense. If anybody can go build a website, then the web is open. The question is, can they be discovered? Well, depends on how well they play the game. It could be that they just want to notify their customers that this is where you come to do X, Y, and Z with this company that you have aligned with. So, you know, that could be easy. But, you know, if you want to be discovered more broadly, you’ll have to figure out how that gets done in this new world. There’s going to be a way. I do think some of the things that Google has introduced here are interesting. The agents… I don’t know, man. They just seem to me to be the AI version of Google Alerts. How many of the emails that I get every day are based on the alerts that I have set up for things related to the company I work for and the topics we discuss and the markets that we serve and things like that. I have all of these Google Alerts set up. I’ll be happy to make them agents now on Google, that’s just fine. I don’t see that that’s going to change my relationship with it all that much. It’s going to surface pretty much the same information. I do like the multimodal capabilities that they’ve introduced. You can now include multiple images and things like that in a search. But they’re not getting rid of those 10 links down at the bottom. You just have to scroll all the way past the AI overview, which is getting more sophisticated and occupying more of the page, but those 10 links are still going to be there. In fact, I was listening to the episode of Hardfork that dropped yesterday, and they interviewed Sundar Pichai, the CEO of Google. And he said, people are always going to want to be able to click a link, especially based on the kind of search they’re running. He says, that’s what Google does. We’re never going to abandon the links, but they certainly have pushed them farther down the page. And we know that the AI overviews have led to a significant decrease in the number of clicks that people are enacting on a search results page. So that kind of traffic has dried up to a large degree. So, you know, even though you can still get to those links, I don’t think that solves the problem for people who want to be discovered on the web. We’re going to have to figure this out. I would continue to look at the AEO slash GEO approach to visibility and also keep your eye on what the search experts tell us as they gain experience with this revised version of Google.

Neville Hobson: Hmm. I’m not sure that the links and AI overviews are going to stay as they currently are. I’m certain they’re not. One of the, in fact, three of the articles I read that go into quite significant technical detail about this are based on people’s interpretation, a lot of what was announced at the Google Developer Conference. So there’s no clean answer yet where this is going. We don’t know yet. But there are a number of directions worth looking at from a communicator’s point of view, I believe. So I’d start with if it means that the citation game, as I might describe it, was fundamentally about being findable. So you are signed, your content is cited, it shows up in AI overviews. So getting your content into the results that humans would click on. But that’s not what’s going to happen here. If agents are doing the searching and humans are acting on what the agents find, the question shifts from how do we get found to how do we become part of the information environment that AI systems draw from? That’s a subtler form of presence. It’s less about individual pieces of content, they’re more about sustained authoritative voice over time. This won’t happen quickly. So if you’ve not done prep, for overviews, you’re behind the curve on this because when this lands and becomes clear how it’s going to work, I think it’ll attract huge interest, particularly thinking about, we have a, one of the topics we’re going to talk about is on this AI agents topic, that you’ve got AI agents out there while you’re fast asleep or traveling or doing something else when you come back to your computer the next day or the next hour or whatever it might be. It presents you with the answers to your question. And they are unlikely to include links. They will be, in the sense, for many people, the complete answer to your question. You won’t need to click anywhere. That will appeal to a lot of people. So this is doing all the donkey work, as we might say. I think it suggests that owned channels will matter a great deal, not less. So a search as a distribution channel is being restructured. Direct relationships with audiences like lists, communities, podcasts, trusted platform, et cetera, become more valuable precisely because they don’t depend on being surfaced by an algorithm or an agent. The communicators who’ve invested in direct audience relationships are going to be better positioned than those who relied on search-driven traffic. And I think you made the point that AI overviews started something which was lessening click-throughs. And that is going to get even worse with this from a website owner’s point of view. There’s no click. So that whole model will have to change. So I think that said, there’s some other issues here too of interested communicators. The human moment is still the moment that matters. Even if an agent does the searching, a human makes a decision. So the communication challenge shifts to the point of decision rather than the point of discovery. That’s a more interesting place to operate, closer to the outcome, closer to the relationship. One of the downsides to thinking about this is you are still, let’s say, the human in the verification loop. You’ve got to check all this stuff. Do you just believe what the agent’s surface is? We’re not at the stage yet where it’s a bit like, self-driving cars, know, you’re 99% sure it’s not going to crash and kill you. Well, you actually need 100%. You don’t need 99%. And this could be in that kind of area.

Shel Holtz: I’m not even 100% driving my own cars 100% safe.

Neville Hobson: Well, you got it. You’ve got to conduct that. So the uncomfortable conclusion that I take from what I looked into was that some of what communicators have been doing for years, content volume, keyword optimization, SEO-driven publishing may genuinely be losing its value. So the honest conversation with clients and employees is we need to redirect effort from producing content for search algorithms to building genuine authority and direct audience relationships. That takes time and you will not see the benefit of that for a while. In the meantime, you’ve got utter disruption to the existing model heading our way.

Shel Holtz: Yeah, we definitely need to do that. But I do need to reiterate that as they were making the announcements about the brand new updates to Google search at I.O., they did double down on including reference links. Here’s what I got. Even though Google is transforming search from a simple box of links into an agent driven conversational interface powered by Gemini 3.5 Flash, they are integrating reference links deeper into the experience. There’s going to be a split over how that’s done. If you’re getting AI overviews doing a regular search, they’ll still appear at the top of the standard results page and you’ll still get the blue links visible right below it. If you go into AI mode, click the AI mode button. It completely replaces the traditional results page. Within AI mode, websites are either explicitly cited as references within the AI response or they receive no visibility at all as there are no traditional link positions under.

Neville Hobson: Right. So that comes into the area of there’s no clean answer yet. So you’ve got the kind of vested interest vendor saying, well, the link’s going to be down here. I’ve not seen that prominently mentioned in any of the tech reports that I did read. Indeed, I’ve seen one that said this is not prominent at all. Where does that fit then? Or is it going to be parallel? If AI agents are going to be out there on your behalf doing all the search work, you don’t have to do anything. How does that fit? Is it going to be like, well, if I don’t want to use AI agents, I can still use the old search. Is that how it’s going to be? No one knows yet. So this will emerge, I’m pretty sure. And it may well be right that they’re scroll, scroll, scroll, scroll. There are all the links I’m used to seeing. I don’t think that’s actually how it’s going to be, matter what they said in the I.O. Developer Conference. We don’t know.

Shel Holtz: or contextual.

Neville Hobson: So you and I are actually doing what everyone else is doing, speculating what it all might be. But there’s some clear direction, such as what I mentioned just now, for communicators in particular to pay attention to here, that you could see wind of change is blowing here. And whether there’s still going to be links at the bottom of the page or there’s an AI overview somewhere else on the page, you need to pay attention to this, what’s changing. The AI agents are going to really change things here.

Shel Holtz: Well, here’s the thing, even if there are links contextually in the narrative of the AI overview or there are still blue links below, people aren’t clicking them anyway because they’re getting that AI overview answer. And we talked about this when they first released the Amazon Echo because they talked about the issue of one. Yeah, we talked about the idea that this is one true answer. It’s not 10 links. And if you don’t like those, you can scroll to the next 10.

Neville Hobson: There you go. Four years and years ago, yeah, yeah.

Shel Holtz: It’s one true answer and you don’t know where they drew it from.

Neville Hobson: Yeah. But I think it’s shifting, though, the behavior, how it works, the idea that you enter a search term, either speak it or whatever it might be, and on the fly, will create interactive visuals. It will do all sorts of things that search can’t do until now. And so you won’t get an AI overview that’s just text. It might have links to it. Whenever links to visual, it’ll have the content there and then, as I understand it from what I’ve read, but hey, it’s not clear yet. So I’m sure we’ll see voices of people that you pay attention, that I pay attention to with some sensible analysis. There’s not enough information to go on yet. The trick I’ve been thinking most about is how do I separate, I won’t call it slot because it’s not, it’s just someone’s opinion that’s not informed from the stuff that is informed. So I’ve got a handful of the latter, and probably scores of the former. I’m trying to ignore all that stuff. To me, this is like the first alarm bell. Alarm is the right word, the first alert bell that you are about to see as a communicator, a fundamental shift in how search works. And it’s going to be, I think it’s going to be pretty. It’s going to be chaotic. So you need to pay attention to this in the right way. And that’s where you need I guess really to find voices that are authority that you can trust to hear how they’re described.

Shel Holtz: Well, let’s switch gears and put AI behind us for a few minutes and talk about something entirely different. Ron Culp’s blog, Culpwrit, ran a guest post earlier this month by Anuj Agarwal with a title that’s hard to argue with, Why Podcast Placements Are the New Press Coverage? By the way, I met Ron when he was running comms at Sears when I was doing some internet consulting there. Ron is one of the senior dons of public relations these days, and I do pay attention to what’s on his blog. In his post, Agarwal isn’t talking about having a podcast. He’s talking about pitching your client, your executive, your subject matter expert as a guest on someone else’s podcast. That distinction matters because that’s where actual news is happening. And if you’re not doing this yet, you’re already behind. Now let’s start with the audience math because the case is overwhelming. According to Edison Research’s Infinite Dial 2026, which they released back in March, 80% of Americans 12 years old and older have now listened to or watched the podcast. 58% listened in the past month. That’s about 167 million Americans, an all-time record. 45% listened in the past week. And among 18 to 34 year olds, weekly podcast reach is now equivalent to broadcast television. I want you to let that sink in. I’m going to say that again. Among 18 to 34 year olds, weekly podcast reach is equivalent to broadcast television. For your client trying to reach younger adults and appearance on the right podcast has the same weekly reach as TV. Now compare that with what’s happening on the traditional media side. Pew Research found that newsroom employment in the US dropped 26% between 2008 and 2020, and that bleeding has continued ever since. Newspaper newsrooms alone are down 57% from 2008. There are just fewer journalists and fewer editorial slots, which means there’s been a shift around where people are finding their content. Your clients still want earned media that channels worth pitching have just expanded. So why is a podcast appearance more valuable than a traditional press hit? Four reasons, and they’re all important reasons. First is the time on stage. A press mention is a paragraph. Even a feature article gives you a few quotes wrapped around someone else’s narrative. A podcast guest appearance gives your client 30 to 60 minutes of uninterrupted access to an audience that chose to be there on a podcast that’s likely to be specialized as opposed to generalized. That’s not just longer, more time. It’s a fundamentally different kind of exposure. The audience hears how your client thinks, not just what they say. They hear the tone, the texture, the moments of pushback. And Nielsen research has shown that listeners trust long-form audio hosts more than they trust traditional advertising. And that trust transfers to the guests by association. Number two is durability. A press hit spikes for a day and then disappears. A podcast episode lives forever. It sits in someone’s back catalog. It shows up in search. It gets clipped and shared on LinkedIn six months later. That’s the long tail at work. And Neville, I don’t know if I’ve shared this with you, but I get emails from the host of our archive site. Remember, we switched sites at one point and we archive the old site where many, many, many episodes of FIR live and I’ll get emails saying your site has exceeded the allowed site traffic for the month and won’t be accessible until the first of the month. Who’s going and visiting this site that hasn’t been active in 11 or 12 years and listening to those old episodes of the show? 40% of podcasts

Neville Hobson: Hahaha AI agents, I would hazard a guess. Yeah.

Shel Holtz: No doubt. I don’t know if they’re listening to the episodes, but they could well be scraping that site. But 40 percent of podcast consumers now name YouTube as their primary podcast platform, which means a single guest appearance can produce both an audio asset and a video asset all from one hour of your client’s time. And that’s content you can repackage and repurpose for multiple channels. Here’s the third reason. And this is the one I really want you to hear, because most communicators miss it. A podcast appearance is increasingly how you earn traditional press coverage. Newsworthy things your client says on a podcast get picked up by the mainstream media. And now they’re not just quoting your client, they’re amplifying an extended quote in context and they’re sending readers back to the original episode. Look at Jamie Dimon. A couple of weeks ago, he gave a podcast-style interview at a Norway Sovereign Wealth Fund conference where he warned that the private credit market is worse than people think. That’s a quote, worse than people think. That single cent has generated coverage across Bloomberg, Reuters, Fortune, the Financial Times, and erased about $500 billion in alternative asset manager value the same day. The Jensen Huang interviews, a Stratechery, BG2, the All-In podcast, those generated Fortune and CNBC stories, and they do every single time. Howard Lutnick told the All-In host that, “‘Boeing executives follow me around like puppies,’ because of how he was structuring trade deals.” Fortune quoted that line word for word in a feature on the Trump administration’s trade strategy. The podcast appearance was the press strategy. That’s the model. You’ve got to, you’re not just earning a podcast placement, you’re earning the cascade of secondary coverage that flows from it. And finally, there’s depth of message. An interview lets a thought leader actually develop an argument. In a newspaper, quote, complex ideas get compressed into one sentence, usually the wrong sentence, often the safest sentence. On a podcast, your client gets to explain the why behind the what. That’s where credibility gets built and where the next sales conversation gets seeded. Now, the catch, most podcasts aren’t worth pitching. Agarwal makes this point bluntly, and he’s not wrong, of the millions of shows indexed across hosting platforms. A lot of them haven’t published in months, maybe years. They have audiences that are minuscule. They’ll book anyone with a pulse. Recommending a weak placement to your client costs them time and signals that you’re vetting is getting lazy. So here’s what to actually look for before your pitch. Publishing pattern, you know, at least two episodes a month consistently over the last six months? The guest history, scroll back through 20 or 30 episodes. Who were the guests that they booked? If your client doesn’t fit that pattern, walk away. Niche depth over raw audience size. I mean, a 600 listener show inside your client’s exact industry will outperform a 50,000 listener general business show on almost every metric that actually matters. The host social application. A host who promotes new episodes on LinkedIn extends your client’s appearance for free. And listen to one recent episode before pitching. If the host lets guests just monologue without pushback, the audience tunes out and your client gets nothing. And here’s what your client needs to bring to the conversation. Not talking points, a point of view. One counterintuitive take, one framework that actually gets used, one honest answer to a hard question. Hosts are really good at spotting the difference between a guest with something to say and a spokesperson reciting messaging. The appearances that earn return invitations, produce shareable clips, and get picked up in tomorrow’s Wall Street Journal are built on original thinking. The PR pros who figure this out now, who develop a real podcast pitching practice the same way they developed a media relations practice 20 years ago, are going to look like wizards to their clients.

Neville Hobson: That’s quite a landscape you described there, Shel. I think I look at it in a way when I hear all these statistics being rattled off about how many people listening to podcasts, what’s up, my kind of eyes glaze completely because, you know, and then I hear about, you know, the Jamie Dimon of this world and these kind of, you know, CEOs of mega corporations and how they move markets with a phrase that gets picked up in the mainstream media. That doesn’t happen with, you know, as a matter of course for everyone. Amongst all the rubbish podcasts you outlined, where you’ve got minuscule audiences or not, I think about this, if I were advising someone to start a podcast and it was a business that’s a medium sized business, maybe a thousand employees in a not particularly exotic sounding industry, but they’re the market leader and they’ve got some interesting stories to tell. How is that relevant to them, knowing that the CEO of JP Morgan or whatever said something in the FT report, completely and utterly irrelevant and not likely to happen to them. So what would be of interest to a company like that who has, as I say, stories to tell, it’s worth doing something with them for a podcast for all the reasons you mentioned that how they get noticed and so forth. So that’s the trick, finding that relevance to convince them or to pitch them if you like on doing it. I mean, you’ve got some good stuff mentioned in this piece you’ve been discussing that are valid, absolutely would make a good way to pitch it. I did like the section in this piece where it talks about how to build the skill before a client formally asks. And totally. And it got me thinking, it says, identify 15 podcasts relevant to their audience and industry. It got me thinking immediately and this is where AI comes into play. Okay, so we can’t escape. This is about AI as well in this topic. Your example, where the experiments you’ve been doing, where you’ve been looking for podcasts in your industry that take guest speakers or all the criteria you’ve done and you’ve been sending out agents who’ve returned with amazing data. That is exactly how you would approach this. And you get your results fast and probably well, they’re probably definitely more accurate than if you do this manually yourself. So where the guidance talks to you about filter, filter these findings you get from shows that publish consistently and actively except guests, there’s your AI tool to do that for you. So you could apply that to this. And it would probably, again, depending which AI tool using, I would say it will give you things you didn’t know to ask even because you didn’t know, but it would find something. So AI would help you do this without doubt. But that’s not the main trust the story though. The thing you’ve outlined is very good. I agree. I just think you need to make it wholly relevant to your client who is not a Fortune 500 company, who is just a regular business, even with 20 employees, they still got a story to tell. If they’ve got a valid story to tell, there’s a chance that they’ll get noticed. And your goals might well be very different from that mega corporation example that you had, where it might be resistive temptation to listen to the sales guy to say, we want leads generated and I want to get 10 sales out of this. You might, but your goal might be different to that. Still got to be a measurable goal. So talking about our CEO will do a thought leadership piece. No, please don’t go there. You need a measurable goal. It could be leads, it could be something less esoteric than that, quite simple, raise awareness of something that you can measure. Now that then puts it in the area of something that you as a communicator know about measurable objectives. So I think scaling it back to reality is a way to do this, not just look at how many millions of people listen to podcasts, because that to me is a total eye glazing metric for most people.

Shel Holtz: Well, I think the point to be made from the number of people listening to podcasts is just that it’s a valid medium now as influential as TV. You know, if we had made this case 15 years ago, it would have been a tough sell. But the fact is that a lot of people have started to recognize the importance of being a guest on a podcast. I mean, this is something the Republicans figured out during the last presidential election cycle. So I think the scale is important, especially as you look at the numbers of diminishing opportunities in the mainstream press. In terms of Jamie Dimon, yeah, I mean, I made the point that he’s going to get on a general business podcast with 500,000 listeners, a million listeners and make news. But I also noted that for some CEOs, for some thought leaders is probably the podcast that has 600 listeners that’s in that niche area. Now, you mentioned the agent I created. It’s actually a skill. I created this in the Hermes agent platform that is configured when I launch it to go out into the podcast space and find the podcasts that meet these criteria that I listed there. But I’m able to enter the information about the person or the story or the thought leadership concept that I’m trying to pitch. So if, for example, I want to get a story out there with this guy that we have who works on all our airport projects. He is a design manager. He is really good. And he has some thought leadership ideas on construction work airside in active airports. Now, do I want to get that into the construction trades? No. What do I care if builders hear that? I want to get that into the airport trades. where the people who are hiring the builders will hear it and hear about his expertise and hear his thoughts and go, yeah, I think I want to talk to them about our next airport project. So yeah, I have the AI agent set up to go do that. It even does a first draft of the pitch based on what it knows about the host and the content of each of the podcasts it returns to me. But yeah, I have to go out there and make that pitch. I’ve got one going on right now based on a sustainability story. So it’s interesting finding the places where we want that story to go because everybody’s looking for more sustainable ways to build, but how does that affect our reputation? So I have to cue this skill in the Hermes platform to find ones that are going to be of high value for me. But I have no doubt that if my CEO gets on a podcast or that design manager gets on a podcast, and says something really interesting, it’s going to find its way into ENR, the engineering news record, which is the primary trade publication or building design and construction or any of the other trade publications out there. And that’s that trickle down. It’s not the scale of what Jamie Dimon can do, but it’s the same concept exactly just at that smaller, more niche scale.

Neville Hobson: Yeah, I think that’s the key thing for me, certainly, is you’ve got to scale it relevantly to your client and their audience and market. The only other thing I’d add to that, Shel, is to let our listeners know, listen to episode 513, where you’ll hear Shel talk at length about the experiments he’s been doing with Claude CoWork, particularly, but also this example that you kind of touched on, as it were. It’s worth the listen.

Neville Hobson: Well, Dan, that was quite a report. I think an excellent assessment of what’s happening with WordPress. I know you’re a big, long time fan and user of WordPress. And so you did a good job in conveying the significant changes from WordPress 7. I think it’s worth emphasizing, and you did make the point, Dan, that a lot of the cool stuff that you talked about, AI-driven in particular, is only on the hosted WordPress, the WordPress.com service, not self-hosted WordPress. So no doubt that will be coming at some point, I would imagine. So

Shel Holtz: Yeah, the FIR site is on WordPress, but it’s the hosted version. And I’ve read about this. I upgraded it to 7 and went looking for these cool features and they weren’t there. that’s Dan’s report clarified that for me.

Neville Hobson: No, exactly, exactly. It’ll come, I’m sure. So thanks a lot, Dan. That was a really, really good report. So our next topic, going now back to the AI, the AI themes we’re following. Yeah, we can’t escape AI at all these days. We know that. But here’s the thing. There’s a Substack writer I follow called Ruben Hassid. He writes a newsletter called How to AI. And if you don’t know his work, He’s worth your time. There’ll be a link to that in the show notes. He publishes twice a week. He’s prolific. He’s sharp. And he has a gift for making uncomfortable ideas land without making you feel attacked by them. His latest piece is called You can’t Beat AI. And before you groan at the title, it’s not the piece you might expect. It’s not a doom narrative. It’s actually a remarkably clear-eyed analysis of what’s happening to the value of knowledge work and what to do about it. The core argument is this, AI is getting cheaper at a rate most people haven’t fully absorbed. We’re talking about a 6,000-fold reduction in the cost of AI intelligence over four years, he says. Not twice as cheap, 6,000 times cheaper. And crucially, not just cheaper, smarter too. So your salary is now being compared to a subscription, and the subscription is winning on price. The part in his newsletter that really got my attention was his argument about what this means for how we communicate our value. He makes the point that for decades, time spent was a proxy for quality. If something took six hours, that effort signaled something. It signaled expertise, diligence, craft. But now that production is cheap, now that a decent first draft, a market scan, a slide outline, a campaign proposal, can be generated in 30 seconds or less, time spent is a weak argument. He puts it simply, nobody cares that it took you all weekend. The better argument he says is, I understand the real problem. I knew what to ignore. I found the missing risk. I made the decision easier. I saved the team from doing the wrong thing beautifully. And that lands differently for communicators, I think, because most of us would say, Yes, of course, we’ve always sold judgment and strategic counsel. But here’s the uncomfortable follow-up question. Do our clients and employers actually pay for that? Or do they still at some level pay for the volume of deliverables, the press releases, the reports, the campaigns, and just assume the judgment comes with it? That’s what I want to explore, because if intelligence is becoming a commodity, communicators individually and as a profession, need to get much clearer and much more explicit about articulating what they offer that can’t be replicated for $20 a month. What do you think about all that, Shel?

Shel Holtz: Well, I think we’ve been talking about this issue of pricing in the AI era. Yeah, I mean, yeah, even well before AI, the hourly model has been questionable. I’ve shared this anecdote, I think probably five or six times on the show. But for those who haven’t heard it, when I was a consultant, Mark Schumann was visiting us. He worked for the same consulting firm I did at the time. And he was talking about value-add, which I opposed. I said, that should be baked into our hourly rate. And he said, okay, how does the hourly rate factor into the fact that I bring 35 years of experience to this job and I have a brainstorm in the shower that took me 20 seconds. I wouldn’t have had that brainstorm if not for that 30 years of experience, 35 years. But it still produces tremendous value to the client that may make them millions of dollars or save them hundreds of thousands of dollars. How do you value that 30 seconds it took to have that brainstorm? And I haven’t come up with a really good answer to that yet, except pricing based on outcomes. That said, I have to say that when I was an independent consultant, which I did for 21 years, I had some clients who wanted me on a retainer. And the retainer was so that they could call me when they had an issue. They didn’t want me to produce anything. They didn’t want me to write anything. They didn’t want me to create anything. They just wanted my counsel. They wanted an hour on the phone. So for X dollars, they got 10 hours a month and they could call and say, we’re going through this. What do we do? Or what’s your experience with that? I think you can still charge based on time for that. It’s the time you’re spending on the phone with them. So I don’t I don’t think charging. based on time vanishes, you just have to figure out what your formula is, where you can factor in how much time it took, where you can factor in what the overall value of it is. And then finally, what you can factor in as the ROI that the client has gotten from this, that you can come up with a formula to make a reasonable return for your effort. But the old model of saying, you know, he’s billed out at $375 an hour. It’s going to take 25 hours for this project. Here’s your price. And by the way, if there’s scope creep, we’re going to come back to you and tell you how many more hours that adds to the project. That’s not going to wash anymore. Not not under these circumstances. Especially by the way, I have to say, not when they can go to AI and ask the question before they come to you for that hour of phone time.

Neville Hobson: I think that’s a key. I mean, that’s the key thing. A client might say when you present the proposal, you just outlined, for instance, that they’ll say, you use AI. You’ve told us consistently how you’ve developed skills in how to use the tools that give you the results that you’re seeking. and you have expertise in that area. So if that’s the case, how come you’re me, you’re quoting me X hours at so much per hour at this price? So what’s your AI doing? better question might be, well, what are you doing in that case? So that’s something that we need to figure out. And I don’t think anyone has yet how to kind of Deflect that into the real questions that need to be answered, the real discussions that need to be going on, which is not that. That’s kind of an emotional based one. But the point you made, absolutely true. Anything you’re thinking of doing, the client’s already also looking at the AI tools they use to find an answer to that. They’ll be second guessing you even. In which case, that cannot sustain itself, that kind of model. It’s not going to suddenly vanish, I’m sure. And to your point too, I agree with you that some elements of service that you might provide that others will buy from you will probably still be time-based, such as the retainer model, for instance. Then again, I think your competition is the smart AI model, frankly. So that comes down to the, you you’re being… Your salary has been compared to a subscription now by the client in particular. So again, more change coming to communicators and not just communicate, but I’m focused on this. What does it mean to communicate? I’ve written about this a few times in my blog. I’ve had many conversations with people about this recently on how to counter the interference of AI. As one person put it to me, you can’t, it’s not interference. you’ve got to learn to live with it. You’ve got to learn to make the most of it, ally yourself with augmenting your ability and not just churning out stuff from a chat bot effectively and do what we’ve been discussing in previous episodes about hallucinations and setting out stuff riddled with errors because you didn’t check it. So your methodology has got to shift. And it doesn’t address the base question, which is how do we migrate as a general rule from time-based charging to value-based charging? I hear that question being asked a lot. I’m asking it too, and I don’t have an answer to that yet. So that requires some exploratory analytical conversation. And it may well turn out that there isn’t a single model that works for everyone, although there must be elements that will be consistent no matter what you do. And it requires significant input from your client or your employer to have that conversation that give you as the communicator, the kind of intelligent input you need from the business owner or your employer. What are they looking to achieve? What do they want exactly? And how can this help them achieve that? So we haven’t started any of that yet, other than asking the question. So we need to pay attention to this sooner.

Shel Holtz: yeah, we definitely do. But there’s one other factor to keep in mind here. And this is a phenomenon that I have been experiencing in addition to reading about it. And that is that for all of the productivity gains that we’re seeing from AI, and we are, mean, the time it takes my agent to go out and find those podcasts for me to do that would be hours with the agent. send it off. I go do something else and it delivers the results when it’s done, which is not hours and hours. It’s usually 15 or 20 minutes. But does this mean that I now have an extra day that I can take off because I have got all of my work done? No, I’m as busy as I ever was. I am filling those hours with work for my employer. I believe it is high value work and it’s either figuring out how to do something with the AI or managing those AI agents or creating the agents or skilling up the agents or it’s doing the things that AI can’t do. So yeah, I’m busier than I’ve ever been. And by the way, I know we’re not talking about the job-pocalypse here as we heard it referred to earlier. But increasingly, I know there are some things that AI can replace customer service agents being a great example. I think for most of the jobs out there, there’s elements of your job it can replace. It can’t replace the whole job. And for all of the companies I’ve seen announcing layoffs that are tied to AI, I have not heard one of them say, and these are the full-time agents that have replaced the full-time work that this person did. I think they’re just anticipating what they’re going to be able to do with AI. So I don’t think we’re going to see the job loss that we’ve been talking about because I think they’re going to realize that what we still need 70% of what this person does and he can devote more time and attention to that because the 30% that the AI can do, we’re going to let the AI do. So, the reason I point all this out. is that while companies may for a while think, who needs an agency? I can do all this with AI. They’re going to realize what they’re producing is slop, that they don’t have the time to put in the effort required to generate the right kinds of agents, the right kinds of skills, the right kinds of prompts. They don’t know how to assemble all of this together in a strategy. And they don’t want to take the time to work with AI to do it. my God, this is what agencies are for. I’m going to have an agency do it. So the agencies will take advantage of the AI, but I think there’s still going to be work for agencies to do. There just may be a valley as companies say, the AI can do all this for us before they realize, well, you know what? It really can’t.

Neville Hobson: Yeah, yeah, yeah. So I guess a short answer to this, there’s no single answer to how you do this. There’s no real way of proving value in an age of cheap intelligence, I suppose, to use that way of describing it. It’s going to get cheaper, according to Hassid, I agree, based on what he said in this, but also what I’m seeing others say. But there’s still a role. I think we need to be clear-eyed about what the future holds in this regard. And you’ve mentioned a good point, I think, about job losses and the way in which that’s currently playing out, where people are taking a big hammer to a tool that doesn’t require that at all to solve, where you’re just laying off people. You’ve got some leaders in some organizations who truly do not deserve to occupy those roles in those companies, the way they treat employees in this regard. But you’ve also got some very sensitive, empathetic people who don’t do it like that. So that’s the landscape we’ve got. In the meantime, I would say communicators need to think about this more. Even if you don’t know the answers, OK, compile all your questions then. Then do some research yourself even to find out what others say. Hey, I will help you do that. And at least get this on your radar before it’s too late.

Shel Holtz: Yeah, and it is definitely going to get cheaper for no other reason that the AI companies don’t want to be spending the money they’re spending on their infrastructure, especially the power requirements. So they are investing in finding solutions to this, not because they are all deep, sincere environmentalists, but because it’s costly to them and they’d rather not be spending that money. All right, well, we’re going to shift away from AI again for a moment. Here’s a phrase that was,

Neville Hobson: There you have it. I bet you we’re not. Go on then, go on, go on, go on.

Shel Holtz: Yeah, bet we are. Well, maybe not entirely. Here’s a phrase that wasn’t in my vocabulary three months ago, the clipping economy. And if you’re a communicator, you need to know what this is because it’s already changing how content about your client or your employer travels and whether you have any control over where it ends up. Here’s what’s happening. A piece from NPR featured a 24-year-old guy in Antwerp, Belgium named Emre Bayraktar. He was working three part-time jobs. He was cleaning cars. He had the night shift at a warehouse. He was making sandwiches at Subway. But in his spare time, he’d take long influencer interviews, edit them into short clips and post them. And one night he got a notification telling him that he’d earned $12 on one of his clips. Two weeks later, he’d earned $2,500. Today, he runs a network of 40,000 freelance clippers. 40,000 individuals like him who take long-form content and turn it into short term clips. He runs that network and that’s the model. Streamers, podcasters, brands, even political campaigns post bounties on marketplaces and anyone with a phone and a video editor can claim the bounty by chopping up the long-form content into vertical clips and posting them to TikTok, Reels, Shorts, X. You get paid per thousand views. NPR cited recent bounties at a dollar per thousand views for Major League Baseball clips, $25 per thousand views for an AI startup. Forbes reported that generating a million views through one of these clipping companies, a company literally called Clipping, founded by a guy named Anthony Fujiwara, cost somewhere between a hundred bucks and a thousand bucks. Fujiwara’s company did $7.7 million in sales in 10 months. MrBeast started his own clipping company called Vyro late last year. Ed Elson, who co-hosts the Prof G Markets podcast with Scott Galloway, calls this the clip economy. And his argument is that the clips are no longer promotion for the show, the clips are the show. The live streamer, Hassan Piker, pulls about 33,000 viewers for a typical live stream. His average clip pulls 700,000 views. The clip is 20 times bigger than the thing it was clipped from. And Adam Rosenberg from Digiday says clips used to be the byproduct. Now they’re the product. And yeah, I’ve got personal evidence. This is true. I think I mentioned just recently someone I work with told me she does not watch Saturday Night Live. She’ll watch 10, 12, 13 clips so that she can see the funny bits. So what does this mean for those of us who work in PR and corporate communications? Let’s answer three questions. First, do you need to watch for Clippers using your client’s content out of context? Yes, always. The streamer Tim “TimTheTatman” Betar, told Digiday that being clipped out of context is now the cost of putting anything long-form online. Clippers will pull the line that fits a narrative they want to push and let it travel. If your CEO does a 45 minute podcast and says something deliberately provocative at minute 31 to make a setup payoff at minute 33, you should assume someone is going to clip the provocative part and lead the payoff on the cutting room floor. So monitor for that. Set up alerts on your client’s name across TikTok, Reels, and Shorts the same way you set up Google Alerts 20 years ago. And have a response plan ready because by the time you find the clip, it may already have a million views. Second, should legal get involved? Sometimes, but think carefully before you reach for the cease and desist. Frank Poh, who runs a creator focused law firm called Poh Law, told Digiday that publicity rights are the live legal issues here, meaning your client’s image, voice and likeness being used to sell something without permission. If a clipper is making money off a clip of your CEO, and especially if the clip is being used to promote a product, you have an argument. You have a case, but if it’s a clipper just monetizing views on a clip that’s substantially accurate, you’re probably stuck with the Streisand effect. Going after a clipper with 200 followers can generate more press than the clip ever would. Pick your battles, save the legal response for clips that are deceptively edited, false or commercially exploitative, and let the rest go. Third, and this is the one I expect people are going to push back on, Should you pay clippers to amplify your clients content? Well, here’s the case for it. The Digiday piece pointed to a clip of John Hamm dancing in a club on the Apple TV show, Your Friends and Neighbors, that went so viral that Hamm got asked about it on The Tonight Show. That’s an enormous earned media outcome from a 30 second clip. Spencer Pratt, the former reality star now running for mayor of Los Angeles, is openly running paid clipping campaigns through the Whop platform, and disclosing it, which is the right way to do it. fintech, crypto, and AI companies are the heaviest users of clipping farms precisely because they operate in legal gray areas. Most clippers don’t disclose payment, which means almost every paid ClipYou commission is potentially a violation of the Federal Trade Commission rules. Here in the U.S., the U.K. Gambling Commission has already taken enforcement action. X has flagged the practice publicly. Instagram CEO Adam Masseri put out a video on April 30th saying the platform will crack down on re-uploaders. And if you go down this road, you disclose every payment every time, no exceptions, and you assume that the FTC is reading your contracts. The clip not going anywhere. The question isn’t whether your client gets clipped, it’s whether you have a strategy for when it happens.

Neville Hobson: Well, that’s quite a story, I had not heard about this either. The whole idea of clipping farms or as the NPR piece wove its story quite skillfully in an interesting way about the kid in Antwerp, how he started out and then the other examples in there of these people. There’s not really a flip, but a kind of an alert note. that the ones who are making the real money on this, they say, tend to be the middlemen clippers rather than the original creators. And that, think, is clearly a sign that this is ripe for intervention by regulators and others in companies who see this kind of behavior. So this may well be one of those things that is temporary in the context of that. The idea of clipping, though, isn’t. And so some big names are going to move into this soon, I would guess. But the thoughts occurring to me when listening to what you were saying was, how does this kind of start? How do you find companies to do this? You asked a question not rhetorically on should your employee or client get involved in this? It’s high risk, but potentially significant. return on that risk, think, like the example you mentioned of the John Hamm example. I heard about that. I heard about that without realizing it was because of a clip. So this probably is one of these natural evolutionary steps in the stage of where we are with online communication, online digital communication that is generational largely, isn’t it? If it’s what is it, Gen Z or even Gen Alphas, I suppose. sharing all this stuff to TikTok and other places. I’ve not seen it on Facebook, but again, I wasn’t knowing what to look for. I’d make a point of looking. How do you identify this stuff, particularly if they’re not disclosing anything? Does it look kind of like, that guy posted this really interesting clip, not knowing that he’s a clipper? Definitely, you need to be aware of this from a communicator’s point of view. you’ve looked into this more than I have. So it’s difficult to say more than that at this point. But I think this is a phenomenon. It’s growing. We’re going to see more of this before, in some way or form, it gets regulated. I don’t mean that from an official point of view. But I don’t see this as remaining unfettered for much longer.

Shel Holtz: No, not at all. I mean, figure the kid in Antwerp has what, 44,000 clippers working for him. Then there’s the Whop and the one that MrBeast started. So you’ve got to figure there’s tens of thousands of people out there creating these clips. How many can they create in a day? Probably hundreds, if not more. And they’re probably flooding the Internet. We just don’t see them because we’re not looking for things that are associated with the tags.

Neville Hobson: Incredible.

Shel Holtz: that they’re putting on them. you’ve got to figure that people are seeing clips and assuming that they’re something that was created because somebody was passionate about it or it came from the company or it came from the person whose image you’re seeing when that may not be true at all. Somebody else is making bank off of that without permission and may not even need the permission. But I got to tell you, the next time I get somebody at my company placed on a podcast, assuming there is a video, I will be clipping that podcast and sharing those clips both to raise awareness of what our spokesperson said, but also to drive views back to the podcast so they can see everything that she said. I’ll be doing that. In fact, for on the same page, the new podcast I started with Steve Crescenzo, I am using a service called Opus Clip. That’s what they do is clips. You upload video. And it creates clips with the closed caption type of, here’s what they’re saying in real time. And I’ve been sharing those on TikTok and Instagram and Facebook and not X because I’m not there. But we’ll see if it drives any more awareness of the show, but this is real and it’s big and it’s important.

Neville Hobson: Yeah. Yeah, I did like the concluding part of the NPR piece where they where someone said, suddenly I realized clips aren’t the promotional material for the content. Clips are the content. And that, to your point on your colleague you mentioned, who doesn’t watch Saturday Night Live, watches the clips so she could pick out the best bits to her, the best bits that she finds on TikTok without watching the show. I do that with some stuff and I never realized, hey, that’s clipping. I didn’t realize that. So that actually got me thinking a bit about something I was doing not long ago with a podcast. I was contracted to a company to produce podcasts. We did a little bit of that. And I remember thinking at the time that the clips are getting more traffic than the actual podcasts. Okay, I have learned from, I should have. paid more attention to that. So maybe we could try with FIR. I mean, I’m thinking one of the reasons we don’t do that, I’m sure from suddenly speaking, just from my point of view in in producing content is I didn’t have time to do it. I have no time to spend. So I would look for something that made it dead simple to just create five clips and I could look at each say absolutely spot on with the have to say Now I need to edit the front and the back. No, no, Just give me that content. Riverside offers that tool, as you know. And maybe there’s something we can do there. But that might be interesting.

Shel Holtz: Yeah, I haven’t been terribly impressed with Riverside’s clips so far. That’s why I’m paying separately for Opus, which I think does a better job. And they do give you some editing tools that makes it really easy to clean them up quickly. I picked on the same page for this for two reasons. One is every other week. So I have plenty of time between episodes to do this. And the other is it’s brand new. So we’re trying to gin up awareness. So we’ll see how it goes. I don’t have any metrics on this yet, but I’m really curious to see how it pays off.

Neville Hobson: Yeah, definitely worth paying attention to. The Wall Street Journal ran a piece on the 15th of May in its CIO Journal section that I think deserves more attention from communicators than it’s likely to get because on the surface it looks like an IT story, but underneath it’s something more interesting than that. The headline is, Companies have a new AI problem. Too many agents. The reporter is Belle Lin and she spoke to technology leaders at companies including Lyft, Davita, GitLab and FICO about a phenomenon they’re calling AI agent sprawl. We’ve talked about AI agents quite a bit recently. So this really is relevant to those conversations. Here’s the situation. As agentic AI platforms have become easier to use, we’re talking tools that even non-technical employees can operate without involving IT. People across organizations are spinning up AI agents. at a remarkable rate. Agents that summarize emails, write briefs, analyze data, automate workflows, conduct research. One company told the journal its 3,500 employees are creating dozens of new agents every single day. Another has over 10,000 agents already deployed. And Gartner projects that within two years, so that’s before 2030, the average Fortune 500 enterprise will be running over 150,000 agents. So that’s the average enterprise multiply that by how many Fortune 500 enterprises around the big numbers. The IT problem is obvious as the journal costs cybersecurity duplication conflicting outputs, but only 13% of organizations believe they have adequate governance in place. That’s a striking gap. Now here’s where I think communicators need to lean into this. When you have agents operating across an organization. often invisibly, often without central oversight, summarizing information, drafting communications, automating decisions, who is responsible for the accuracy and consistency of what those agents produce. And when something goes wrong, when an agent produces conflicting results or surfaces inaccurate information that makes its way into a client communication or an internal briefing, who owns that? Who manages the narrative around it? We’ve talked about agentic AI on FI before, as I mentioned, mostly from the angle of what agents can do and why communicators should be paying attention. This story is the flip side of that conversation. It’s what happens when agent adoption outpaces the governance frameworks needed to manage it. And it’s a reminder that governance isn’t just a technology challenge, it’s a communication challenge. Somebody in these organizations need to be thinking about transparency, accountability, and trust. And that sounds a lot like a communicator’s job to me. So we talked about some really serious errors made by organizations by not having the human in the loop, as I mentioned earlier. This, in my view, fits into that conversation, except it’s now become much more urgent if you’ve got this scale emerging, which clearly it is emerging. And what do we do about it? So is it the communicator’s job? I kind of qualify my own question. Sounds like a communicator’s job, but is it? What do you think?

Shel Holtz: Well, partly. I mean, there’s definitely a role here for IT and definitely a role for the senior leadership of the organization. But we absolutely have a part to play here. So I think one of the problems is that a lot of organizations establish their governance, their guardrails, their policies in the early days of Gen.A.I. as they saw employees starting to use it, say, hey, we’d better have some rules, especially since we’re going to adopt it. So many companies that had Office 365 contracts suddenly had access to Copilot and wanted to make that available. So they came up with their governor’s policies before anybody was talking about agents. Well, I can’t say nobody was, I’m sure they were at the AI labs and…

Neville Hobson: Yeah, just one, don’t forget, the journal says only 13% of organizations believe they have adequate governance in place, which means what? 70%, 80%, something percent do not have it in place at all.

Shel Holtz: Yeah, that’s the thing. And I think those are the ones that may think that they do because they spent a lot of time, they worked with legal and HR and they came up with their governance and now we have agents and the governance doesn’t address them. It addresses prompting and it addresses things like don’t put sensitive information into a prompt in a public AI model, doesn’t say a word about agents. So I think one thing we need to look at is revisiting the governance policies that we have. And here’s where a communicator can come into. into play is look at your governance and then propose some enhancements in the great language that only we communicators are able to craft and send those to legal NHR and IT and say, you know, we need to update our governance. Here’s a proposed draft and at least get things kick started. But, you know, not all of the agents are ones where we need to be thinking about them. Although, I think there are some issues where we could get involved, but I work in a construction company. I have project engineers who are doing things like submittals and RFIs, stands for requests for information. And a lot of that, I think there’s potential for agent work. I’m not that interested. If you can be more productive in your job as a project engineer, more power to you. I think where there’s some process. One of the things to worry about is I don’t want every PE spinning up their own agent independently. If somebody comes up with a good agent to do this with, we should be sharing that throughout the organization so that there are consistent results and we don’t have to be worrying about any mistakes or misphrasing or anything that went into the creation of an agent, creating bad results over here on this project, but great results on this one where they knew what they were doing when they spun up their agent. So there should be, I mean, that should be baked into the governance that says we don’t want, you know, 175 agents created by different people all doing the same thing. Where I get worried though is where they’re answering emails, where they’re answering customer service questions. You know, who reviewed the voice and the tone of those responses and who approved the messaging guardrails? When an agent says something wrong publicly, who’s accountable for that? If it’s a marketing message, is marketing responsible or is it IT? Nobody has the answers to this yet. And I don’t think anybody’s asking. We should be the ones who are asking.

Neville Hobson: Yeah, it’s, what a chaotic landscape we live and work in. I think, I agree, Shel. I mean, communicators have a clearly significant role. I could conclude as we’re saying, hey, communicators, this is another thing you’ve got to be getting involved in now in this. This is how you prove your value. So add this to the 15 new things that you now have to get to deal with. Here’s number 60.

Shel Holtz: Yeah, I was just blown away when I read this, that the Pentagon staff used Google’s agent designer on genai.mil to vibe code more than 100,000 AI agents in about five weeks. They hold IL-5 authorization. That means they can deal with sensitive, unclassified data. Their chief deputy for intelligence said at a symposium, I’m on team go fast. So if corporate communicators think this is a problem, they have time to think about the military, the actor we’d expect to be most cautious is already at 100,000 agents and accelerating. Now, Gartner in step two of the process they outlined, step two is to build a centralized agent inventory. And that was framed as a technical thing. It tells IT what’s running. But for communicators, we need to know what they’re all saying, what tone, what facts, what brand voice. Our job is to make sure the inventory captures the messaging dimension, not just the technical one. Otherwise you end up with 150,000 versions of your brand voice. And once we’re able to analyze what they’re doing, we’re able to go back and say, hey, you need to adapt, adjust that agent. So it does this instead of that.

Neville Hobson: Well, interesting times ahead as always from everything we talk about here. I would say conclude, buckle up. That’s what I said.

Shel Holtz: Absolutely. Also something else to buckle up for, we have an interview coming in June. We haven’t had an FIR interview in a while, but we have a returning guest. He hasn’t been on the show in something like 15, 16 years, but Pete Blackshaw is joining us based on some work he has spent a couple of years doing about the role of AI in customer service. it’s going to be a great conversation. I was just blown away by what I read in his post about this. So looking forward to that. Yeah, watch for that in June. Also, of course, coming in June is our next long-form episode. We are going to record that on Saturday, June 21st and make it available on Monday the I’m sorry, it’s going to be on Monday the 22nd. We’re going to record on Saturday the 20th. I think I got that right. Anyway.

Neville Hobson: Yeah. Yeah. Looking forward. Yep. No, we’ll be recording actually on Saturday, yeah, the 20th, you’re right. And publishing on the 22nd,

Shel Holtz: The 20th. There you go. Yeah, I can count squares on calendars. But watch for that. It’s coming. And of course, we will have our midweek episodes in between all that. We do hope that you will comment on anything you’ve heard that raised a thought from this episode or from any of the upcoming shorter midweek episodes. You know where to share them. Send them to us at [email protected]. Attach an audio file if you like. You can record that audio file right off of the FIR website, in which case you don’t even need to email us. We get that. There’s a tab that says send voicemail right there on the side of the page. You can leave comments on the post for this episode at FIRPodcastNetwork.com or wherever we share news about the latest episode dropping, LinkedIn, Bluesky, Threads. We’re everywhere. We’re everywhere except TikTok. right now, and maybe we’ll fix that if we start clipping. And that will be a wrap for this episode of For Immediate Release.

The post FIR #515: Agents Everywhere appeared first on FIR Podcast Network.

There’s a concept circulating in Platformer, the Reuters Institute, and Nieman Lab: the text-based social networks that defined the last 15 years of public communication may be in irreversible decline. Apptopia reports that Bluesky’s daily users are down 96% from January 2024; Threads has lost users in seven of the past eight months (down 61% from its October 2024 peak); and X has been “culturally altered.” At its peak, was Twitter less a replicable product category than a unique moment in media history? The mass audience has moved to short-form video, algorithmic feeds reward attention over the social graph, and platforms increasingly refuse to be referral engines.

Text still thrives in newsletters, Reddit, Discord, WhatsApp, LinkedIn, and AI chat interfaces — what’s collapsing isn’t text, but giant algorithmic public feeds. Neville and Shel look at what this means for communicators: the promise of scale is giving way to relevance, trust, and consistency — a shift that requires a different approach to brand presence on social. Get details in this not-so-short midweek FIR episode.

Links from this episode:

• Are the Twitter clones in trouble?
• Pew: Americans’ Social Media Use 2025
• Pew: Social Media and News Fact Sheet
• Reuters: Mapping news creators and influencers in social and video networks

The next monthly, long-form episode of FIR will drop on Monday, May 25.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville: Hi everybody, and welcome to For Immediate Release episode 514. I’m Neville Hobson.

Shel: And I’m Shel Holtz. Communicators devote a fair amount of time to social media management. It’s no different where I work. We’re a smaller team in the construction industry, so we don’t have any dedicated social media resources. But whether it’s a company like mine, where it’s part of the job that somebody does, or a global brand like Wendy’s or Starbucks with a full-blown team, everyone’s trying to make an impact on social network users. The strategy behind those efforts may need an overhaul, though, to address the decline of text-based social networks. Platformer’s Casey Newton wrote about this recently, focusing on Threads, Bluesky, and X — but I think it’s fair to throw Facebook into the mix. Depending on whose numbers you believe, Threads has lost momentum, Bluesky never became the Twitter replacement that political journalists or media folks had hoped it would be, and X is, well, shall we say, culturally altered.

Meta and Bluesky dispute some of this third-party data, so I don’t want to overstate the precision of the numbers, but we shouldn’t shrug off the larger point. This isn’t about whether Threads beats X or whether Bluesky can recover, but rather about whether that old Twitter model can be rebuilt at all. And increasingly, the answer looks like probably not. Twitter at its peak was a real-time public layer for news, commentary, expert reaction, and professional visibility. Journalists, politicians, academics, CEOs, and PR people were all there reacting to each other in public. That gave communicators something we had never really had before: a live dashboard of what influential people were saying, what stories were breaking, and how publics were interpreting events in real time. The problem is that this depended on a specific set of conditions — a text-first interface, a public follow graph, a tolerance for public argument, and a shared assumption that this was where you went to see what was going on. Even with a small subscriber base compared to Facebook and a lot of other networks, Twitter was where news broke, and it was frequently cited in the mainstream media’s reporting.

Well, those conditions have changed. The mass audience has moved heavily toward video. TikTok, Instagram Reels, and YouTube Shorts are now the primary discovery platforms for younger users in particular. News and commentary arrive as video, personality, remix, and clip. In fact, I was talking about this recently with someone I work with who said she doesn’t watch Saturday Night Live — she watches 10 or 15 of the clips that Saturday Night Live shares on YouTube so she can catch the funniest bits.

At the same time, the logic of the feed has changed. The old social feed was built around who you followed. The new algorithmic feed is built around what holds attention. A post on early Twitter spread because of the social graph. A video on TikTok spreads because the system thinks it’ll keep people watching. Now that changes the incentives. It rewards performance, emotion, personality, and visual fluency.

It’s also why the link-in-the-post model is fading. Social platforms don’t want to be referral engines. They want the content consumed inside the platform. You can’t conflate social engagement and site traffic anymore. For brands, this requires a pretty significant rethink. Today, social is less about sending people somewhere else and more about creating native moments of value right there, inside the feed.

The implication for communicators is that we can’t just ask, “What should we post?” We have to ask, “What role does each of these platforms play in our communication ecosystem?” Some platforms are for discovery, some for reputation, some are mostly listening posts — environmental scanning, sentiment tracking, intelligence gathering. Some platforms may not be worth the effort at all anymore. We also need more human voices.

The logo account is not adequate anymore. Trust attaches to people — experts, leaders, practitioners, analysts. That doesn’t mean every executive needs to be dancing on TikTok. In fact, please, no. But organizations do have to get better at helping credible people communicate in platform-native ways. The decline of the old public square forces us to build more durable relationships. What matters? Newsletters, podcasts, owned communities. LinkedIn still matters for professional audiences. So I’d resist the lazy conclusion that text is dying. Text is everywhere — in newsletters (which, by the way, is where I latched onto this story, in Casey Newton’s Platformer), in captions, in scripts, in search results. What’s dying is something more specific: the idea that a text-first social network can serve as the default global town square.

Twitter may have been less a replicable product category than a unique moment in media history. For communicators, the job is no longer to master the town square. The job is to understand the map after that square has gone to seed. Neville, is this what you’re saying?

Neville: It’s a lot. There’s a lot going on here to kind of zero in on a handful of potential responses, I suppose. But one thing does seem to be quite clear from all that you’ve outlined, which I believe is the case: Twitter probably was historically unique. And I think the issue, or an issue, is that everyone doesn’t think like that. They think it’s repeatable, it’s replicable. And it’s not. I think you could also see AI maybe accelerating the decline. Content abundance — so much of it. Authenticity is getting really difficult to judge. And everywhere is noisy. And that’s not what many people want.

So I guess, to crystallize it in a sense — you know, we’ve got all these elements you mentioned. The paradox of Bluesky: it hasn’t grown. Threads has got scale, but it doesn’t really have a big identity. It’s kind of part of Meta. What does it all mean for communicators? We’ll come back to that, I’m sure, in a bit. But I wonder — the thought that keeps recurring in my mind from everything I’ve read about this is that the decline may not be about text at all. That’s not to say it’s because they’ve all migrated to YouTube and video platforms. I don’t believe that’s the case either. I think, as you pointed out, and that’s obvious to all of us, the text itself isn’t disappearing. People talk about the decline of text-based social networks. But the audience hasn’t vanished. They’re just dispersed. They’re elsewhere. They’re not in a central place. There is no public square — no global public square.

No matter what the likes of Silicon Valley folks seem to think, there is no global public square. Or rather, there are places online that are accessible globally, but they’re not controlled by big platforms solely. Text isn’t disappearing. It thrives in the examples you mentioned — newsletters, Reddit, LinkedIn comments, Discords, WhatsApp, AI chat interfaces — probably 20 more examples you could give. So perhaps people still want text, but they don’t want giant algorithmic public feeds. I think that plays a big role in this. They don’t want the manipulation of algorithms to tell them, “This is what we think you want to read.” And that is still very strong on most big platforms today. They want to keep you on the site. They want to keep you engaged. They want you to keep doomscrolling, or whatever it might be. I think that shifts the conversation we should be having. It’s not about the decline of text — it’s that people don’t want giant algorithmic public feeds. What do they want? Well, they don’t know what they want, do they? They don’t.

You mentioned LinkedIn — definitely the interesting exception. I think there are multiple reasons why it remains one of the stronger environments for text discussion. I guess I’ll put it as questions, basically. Is it because having something like this — that’s kind of affixed to your professional identity — changes behavior, or rather reinforces “corporate” behavior that is different to how it would be on Facebook or on TikTok or other video platforms? Accountability is higher, reputational stakes are visible, audiences are contextual rather than anonymous. It works, I think, partly because people behave differently when attached to their real professional identity. Hence — you know, I saw one today, someone complaining about Facebook-like posts being published on LinkedIn: “Please stop doing that. This is a professional place.” That kind of comment. So I suppose you could say a useful contrast — to this point about people behaving differently — is outrage-driven engagement everywhere else compared to reputation-aware participation on LinkedIn. I thought that was quite a good contrast to make.

Outrage-driven engagement is very strong everywhere you look, but not on LinkedIn. People really don’t put up with that. So I think you have these elements that take part in all of this — that maybe smaller communities may be replacing mass audiences. And I think there’s a lot to be said in that. The smaller communities that would comprise, for example, niche communities, private groups, curated networks, newsletters, podcasts, creator ecosystems, if you like — much more than these algorithmically driven giant public feeds. I don’t follow much on the other traditional social networks. I prefer my own discovery on things like that. So maybe we’re moving from broadcast social media to relationship-based digital communities. And maybe this is just the start of that — which clearly has major implications for communicators and publishers. We can get into that, I suppose, in a bit. That broadly is what it looks like to me.

I’ve read the Reuters Institute’s report, I’ve read Nieman’s report, which actually are much wider than just talking about the decline of text-based networks, but there’s very much in there. There was an interesting article in Fast Company which looked at Bluesky’s limitations. I thought that was intriguing — what they wrote. The contradiction, they said, was that journalists and intellectual communities (and they define what they mean by that) still love Bluesky, but broader mainstream adoption remains elusive. And that is absolutely true. Little blips here and there of this group or that group moving away from X. Good example in the media: Deutsche Welle, the German broadcaster, announced, I think last week or the week before, that they have multiple — maybe 20 or 30 — Twitter handles, and they’re all going to shut down, and they’re moving it all to Bluesky. And I think — brave. Reality: X is still the place where significant public announcements are made. Look at what’s happening, for instance, in Iran — in the Iran war with Israel and the US — the propaganda battle between the two is taking place on X, mostly.

Trump, well, he’s got his Truth Social network, but I see people on X quoting Trump a lot. So governments, big corporations — but large groups tend to still be on X, in spite of what many, me included, would see as that being definitely a place not to be because of its toxicity. Deaf ears for that message, frankly, to people who are invested big in a place like that.

But what does it all mean for communicators? Well, you’ve touched on a few things, and maybe that’s something that we could explore a bit. But broadly, that’s what I’m seeing.

Shel: Well, I agree with you about the dispersion. I’m going to stick with the idea that the text-based social networks are in decline. You mentioned the German magazine that is sending all of their handles over to Bluesky.

Neville: Well, no — they are the big broadcasting and media organization, not a magazine.

Shel: Well, I’m sure Bluesky is going to be very happy about that. According to the Platformer article, Apptopia said that Bluesky has effectively collapsed as a competitive threat, with daily users down 96% from January of last year. That’s a pretty steep decline. If you look at Threads, which is turning three years old in July, Casey Newton writes that daily active users on the platform have declined in seven of the past eight months. After peaking in October 2024, just before the US presidential election, daily users are down 61%, and global monthly users have held up better at 388 million, but that’s still down from an estimated 400 million at the beginning of January of this year. So these are declining.

Elon Musk took over Twitter and turned it into X, and people saw what he was doing with it. They talked about, “I’m going to Threads,” or “I’m going to Bluesky.” A lot of them were saying they were going to Bluesky, and they ended up being the left-leaning people who couldn’t tolerate the rise of hate speech and rage-baiting on X. So they went to Bluesky and effectively turned it into a left-leaning echo chamber, with left-leaning people arguing among themselves. Meanwhile, one of the people Casey quoted in Platformer said on X that the extreme right is pulling everybody else further to the right because there’s nobody from the left there to argue with them. So that’s what’s happening in the text-based social networks. They are losing monthly and daily active users at a pretty alarming rate.

Where are they going? Well, the numbers for Instagram and TikTok and YouTube Shorts — the video platforms — are all on the rise. This is where people are going, and in particular, young people. I think one of the reasons LinkedIn is succeeding, in addition to the fact that it has a focus on business, is the fact that their algorithm, at least in part, focuses on who you follow. Whenever I log into LinkedIn, I see people that I know — not everybody who’s posted, but I do see people that I know. When I’m scrolling through Instagram, it’s a couple, but really, it’s trying to send me stuff it thinks I’m going to stick with rather than things from people I have connected with. So that makes me less interested in scrolling through Instagram. But I do find myself, when I see something — a video that’s in Reels and I tap it, it takes me over to Reels (this is on Facebook too, by the way) — I’ll find myself stuck there for a while as I’m looking at more of those short video clips. So they’re onto something in terms of what’s going to grab attention. But what does it mean for the public square? I think that — and we talked about that a lot in the early days of social media — you may remember the book Here Comes Everybody by Clay Shirky. I loved that book. It was so optimistic about the public square and about the ability for people to connect with each other. And I just think that’s fading. That’s not what social media is about anymore. And I think the algorithms have driven a lot of that.

Neville: I have a copy. Well, I think you’re right. Indeed, as far as Casey Newton’s piece goes, that same argument — Twitter traffic has collapsed effectively. Many other people are saying the same thing. The Reuters Institute talks about that. Nieman talks about that. Fast Company doesn’t quite say it in the same words, but talks about the tensions there. I totally agree with that. So I think, though, that it has more to do with the shifts in how people see public spaces to connect with other people. I think it reinforces some views you see about users turning their backs on social media. We’re seeing that a lot, and much of it generational, I’m sure. Commentary about fewer platforms and deeper relationships — I think they mean to say fewer platforms that interest people, or that matter to many people.

Industry reporting is showing that short-form video is dominating attention, as you mentioned — TikTok, for instance, and YouTube Shorts, stuff like that. Great, I get all that. But I think much of it might depend on how you use these networks yourself. And I’m not a big fan of algorithmically driven content at all. And, you know, I use LinkedIn maybe not the same as you. Indeed, my whole approach to it is different to what it was even two years ago, where it was important to connect with others and have people like your content and share it widely and all that. Those things aren’t important to me anymore. What is important to me, using LinkedIn as an example, is to connect with a handful of people whose opinions I really do value. And they are literally a handful — maybe two handfuls. I don’t really care about, you know, stuff getting shared 50 times and all that. So I pay attention to some of the metrics, but not a lot.

And I think, you know, Bluesky — I’m there since the beginning. Threads — I’m there since the beginning. And I’ve definitely noticed a change. I look at it this way, Shel: great, less noise, because there are fewer people in there now, which is wonderful — and new people. In fact, I use Bluesky in particular as a way to surface new people to connect with. Threads, that’s a different crowd entirely. And in fact, Threads is far more of a burgeoning community, let’s say, than other places are. But I think, you know, you kind of pause and look at the big picture and say, “Is this getting wider attention around, you know, for everyone?” Well, of course it’s not. It has very high attention amongst tech journalism, in the media and publishing industry — although both of their needs are different. Tech journalists, and I see this myself, actively discussing Bluesky, Threads, X fragmentation. You see a lot of commentary about that. We’ve just been talking about that. Media publishing industry — they’re concerned about traffic collapse and discovery. Comms professionals — attention is growing. And so our conversation today might be quite timely, especially among people who relied on Twitter professionally, says one bit of research that I pulled up. The general public — it’s very low. Yes, of course it is, because this is not mainstream. It will get there eventually, I suspect. Most users migrate without discussing the macro trend. They just move out of Twitter and go someplace else, or whatever it is. Academia and research are beginning to grow around Bluesky behavior and community structures elsewhere.

So people feel the change, but they’re not talking about it in a way outside these niche areas. Insiders are analyzing the change. That’s what a lot of the discussion is now. But generally speaking, no one’s really fully named the change yet. Is it a decline of text? Is it a decline of social network interest? Is it — whatever, whatever, whatever you think it might be. So I think the question, or rather the focus, isn’t or shouldn’t be on whether Threads or Bluesky are failing or collapsing. That’s the wrong perspective. It’s more, as we’ve mentioned earlier: were Twitter and others like it unique historical moments that you cannot recreate? In which case, are we looking at, you know, swathes of people looking for a home to go to? No, I don’t think so. I think most have found a place.

There is no longer a single place that’s a giant public square. There are multiple, and they’re all small. So are we going to see a kind of movement of thousands, maybe millions, of teeny niche communities? You could actually argue: what a nightmare for communicators. So that’s where your audience goes. But that’s what it looks like to me a bit, Shel, frankly. What I hear some people talk about, what I read — my own behavior, even — I’ve navigated myself to small communities. So I’m still a member of a number of LinkedIn groups, although, to be honest, most of them have really diminished in quality of content. A lot of people pimping their wares, full of consultants doing that kind of stuff. No value there at all. It’s almost like Facebook in that sense, where every other post you get interrupted by ads and stuff like that. So that’s going on too, by the way.

So is it fatigue? I think it could well be generational. I mean, I’m not looking right now at any research that supports that opinion, but it seems to me that, you know, if you’ve got migrations to video channels that are typically the home of younger-oriented people — younger-oriented communicators, I’d say — that’s where it is. I take a look at TikTok now and again. I’ve been on TikTok forever and I don’t contribute. I watch. It’s not my cup of tea, the kind of stuff there. The new successor to Vine, called eVine — I’ve got an account. I haven’t yet posted anything, because I couldn’t figure out what I want to say there. But I’ve looked, and there’s lots of interesting stuff. So I’ll keep that kind of perspective. On those, I’m a consumer as opposed to a creator and a sharer. On others, I create, and most of that’s text-based content. I don’t believe that’s always because I’m a boomer — far from it. That’s my preference. But I frequent these other places as well. Is that typical? I have no idea. I don’t think it might be. But who knows?

But I think this is definitely a conversation to have in terms of what it means for communicators more than anything. So, you know, audience building is going to become harder for communicators. That’s how I see it. Owned channels are becoming far more important. Trust matters more. We know that from all the conversations we’ve been having and what we can observe ourselves. Community matters more than reach. I believe that. Personality matters more than scale — I don’t see evidence of that, but maybe that is true. Consistency matters more than virality. I think that’s going to become increasingly more important, and that’s connected with trust.

So the old social media promise was scale, right? The emerging model may be relevance and trust. Are we seeing an emerging model? I’m not sure, but that’s where the direction of travel looks like it’s going, it seems to me.

Shel: Well, it starts with, I think, brands needing to adopt the philosophy that you just articulated with your subscription to eVine — not posting anything because you haven’t figured out what to post there. I think a lot of people in social media create content, they reformat it appropriately for each channel, and throw the same piece of content into all those platforms. That won’t work anymore. I think you need to look at each platform, look at the people who are there, and figure out what you want to share that’s relevant, that will resonate, that will produce some kind of movement of the needle. And for those where you can’t answer that question, maybe it’s time to abandon those networks as a brand. As an individual, go where your people are. We’re not talking about what network you like personally. For people who love being on Bluesky, I’m not suggesting that you stop. Go wild. Have a great time on Bluesky. But is it a place to share the content that you’re sharing from your organization? Is there a way that you can craft content that’s going to work there? Maybe, maybe not, but it’s time to reevaluate that based on these trends.

Neville: Yeah, I think so. I think so. I mean, I think things like human voice, the individual voice, the demonstration of judgment — making judgments about something, your own lived experience and the intentional participation you exhibit by being on a certain platform — I think are increasing. Not scaled. I mean, this is small beer compared to the scaling that used to be the important thing. But they’re all connected with what we’ve been talking about: trust, authenticity, oversight, the evolving role of communicators in AI-shaped environments, to quote a phrase. I mean, it’s all in there too. Whatever — we think communicators need to be paying close attention to this in a macro sense, and indeed maybe the micro sense without getting into the weeds or anything — just be aware. So I pay attention to a lot of this stuff. I may not read every Reuters Institute report or Nieman report — not all. I’ll read the executive summaries. It’s a bit like, you know, the TikTok clips — I’ll look at the clips to get interesting stuff out of it and reflect. Do I think this is something that people I’m connected with will be interested in knowing? And that drives a lot of what I write about on the blog, for instance, which is AI-heavy these days. I think this is important to pay attention to. Perhaps public social networking is in decline. Maybe it is. But the social behaviors in small groups are definitely not. And that’s the thing to keep in mind — those two contrasts.

Shel: Yeah, and that requires a whole different approach than we’ve been taking — to this blasting this content out. And that will be a 30 for this episode of For Immediate Release.

The post FIR #514: Was Twitter A One-And-Done Phenomenon? appeared first on FIR Podcast Network.

Neville and Shel dig into a provocative Harvard Business Review article that argues most marketing teams are structurally unprepared for the speed and scale that agentic AI now enables. The bottleneck, the authors contend, isn’t the technology; it’s the operating model. Neville and Shel connect the piece to conversations FIR has been having for the past year: AI as orchestration rather than automation, professionals shifting from supervisors of tasks to directors of systems, and 2026 increasingly framed as “the year of the agent.”

At the center of the Harvard piece is the idea of a “brand code” — a machine-readable knowledge system that lets specialized AI agents continuously create, adapt, test, and optimize marketing in real time. Communications urgently needs its own equivalent: a “narrative code” containing executive voice profiles, message hierarchies, sensitive-topic guardrails, and escalation rules. Whoever builds it first, he warns, will inherit the agentic stack, and if marketing gets there first, comms will be stuck with a system never designed for crisis, controversy, or stakeholder complexity. The episode also includes some concrete examples and early thoughts on Hermes, Wispr Flow, and where human judgment still has to win.

Links from this episode:

• Redesigning Your Marketing Organization for the Agentic Age
• The Year of the Agent: What it means for the future of communications
• Google Summary: The Year of the Agent: What it means for the future of communications
• If you work in PR and you’re unsure how AI agents will help you, this should help.

The next monthly, long-form episode of FIR will drop on Monday, May 25.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Shel: Hi, everybody, and welcome to episode number 513 of For Immediate Release. I’m Shel Holtz.

Neville: I’m Neville Hobson. Over the past couple of years, we’ve heard countless conversations about how AI is changing marketing and communication. Most of those discussions tend to focus on tools — faster content creation, better personalization, workflow automation, synthetic media, analytics — all the things AI can supposedly do more quickly and at greater scale than humans. A new article in Harvard Business Review published last week takes the discussion somewhere much bigger.

Its argument is not simply that AI will improve marketing productivity. Its argument is that AI may fundamentally redesign how marketing organizations themselves operate. The article is called “Redesigning Your Marketing Organization for the Agentic Age,” and the authors argue that most marketing teams are structurally unprepared for the speed and scale AI now enables. The reasoning is interesting; we’ll look into this in a minute.

AI has already accelerated software engineering and product development dramatically. Products, updates, campaigns, and features are being developed and shipped much faster than before. But marketing organizations, they argue, are still largely built around sequential workflows, siloed teams, approval chains, meetings, handoffs, and coordination-heavy processes. So even when AI speeds up individual tasks, the organization itself still moves slowly.

In other words, the bottleneck isn’t necessarily the technology, it’s the operating model. What struck me reading this article is that in many ways it feels like the continuation of conversations we’ve already been having on FIR over the past year. About a year ago, Shel demonstrated some of the early agentic AI capabilities we were beginning to see emerge — systems that could move beyond simple chatbot interactions and actually take actions across workflows, tools, and platforms.

At the time, it felt experimental, slightly futuristic, and maybe just a glimpse of where things might be heading. Since then, we’ve repeatedly returned to related themes on the podcast: AI as orchestration rather than just automation, and managers becoming directors of systems rather than supervisors of tasks, to name but two. Recently, the wider communications industry has been framing 2026 as the year of the agent, a fundamental shift from generative AI, which creates content based on prompts, to agentic AI, which acts autonomously to achieve long-term goals. The rise of such autonomous agents requires a focus on agentic orchestration, with professionals acting as AI engineers who guide, manage, and audit these digital employees. As we discussed on this podcast last year, communication departments will adopt a hybrid structure where humans focus on high-level strategy and creativity while AI agents handle high-volume procedural communication tasks at machine speed.

We’re already seeing a marked impact on marketing and public relations. The Harvard piece explains how companies such as HubSpot and AWS have begun putting this model into practice. They say organizations are achieving measurable gains, with marketing materials adapted up to 98 times faster, unit costs reduced by 80%, and click-through rates increased up to 17 times. Research from BCG has demonstrated these benefits at scale.

Organizations embedding agentic AI into marketing workflows, the research has found, can achieve up to a threefold increase in ROI, campaign speed, and content volume. That’s why this Harvard article feels so interesting to me. It doesn’t contradict any earlier conversations; it complements them. It takes many of the ideas we’ve been discussing conceptually and places them inside a concrete organizational model. The authors propose something they call an agentic marketing organization — essentially a system where humans and AI agents work together continuously across multiple layers of activity.

At the center of this idea is what they describe as a brand code: a machine-readable knowledge system containing brand strategy, customer insights, messaging frameworks, business rules, governance structures, and operational guidance that both people and AI systems can understand and act upon. Once that foundation exists, specialized AI agents can continuously create, adapt, test, distribute, optimize, and report on marketing activity in real time. It’s a vision of marketing that starts to look less like a department and more like an operating system.

But what really caught my attention wasn’t the technology itself so much; it was the shift in the role of the marketer. Because beneath all the platform architecture and workflow diagrams is a much deeper question: if AI increasingly handles execution, what becomes the real value of marketers and communicators?

The article argues that value shifts away from production and toward judgment — setting intent, evaluating outputs, interpreting signals, shaping governance, and guiding how the system evolves. And that raises some fascinating questions for communicators. But first, Shel, your demo of those early agentic capabilities was about a year ago now. As I mentioned earlier, it felt experimental and slightly futuristic then. So what’s changed since then?

Shel: It feels like ancient history now. If I were to look at that, I’d probably shake my head and say, “my God, that’s pretty primitive.” The way it worked was, it took a screenshot of every site it visited and then acted on the screenshot. So it was a very slow and tedious process. The video that I shared, I edited out all of the waiting time for it to go through all of this, because it showed you everything. And those days are long gone.

That was clearly a demo. I don’t remember which of the AI models offered that — I think it was Anthropic — but it was just tedious and not all that functional. It did what it was supposed to do in the end, which was to create a spreadsheet with the information I’d asked for. It was some open-source spreadsheet that it used.

I ran a similar exercise just last week using Claude Cowork. And this was for a piece somebody in our sustainability department wrote. It was about two projects that had achieved world-first certifications for zero waste, which is kind of a big deal in the construction industry. It’s one of the biggest contributors to landfills and the like, the industry is.

So I’m looking to place this article. And what I did was, I told Claude Cowork that I wanted four subagents working: one to look at construction and AEC publications — that’s architecture, engineering, and construction; AEC is the category for the industry. Another one was going to look at sustainability publications. And there was one other, but I also had it look for podcasts where the authors of this report might be invited for an interview.

I said, what I want you to do is find the publications and podcasts based on their previous content that are most likely to be interested in something like this, and then create a spreadsheet with the name of the outlet. And of course, divide it into these categories — right? AEC, podcasts, sustainability-focused publications, and the like. Mainstream media was the other category. But I also wanted the URL, I wanted the name of the appropriate person to pitch the article to. And then, based on what that person has written — that particular reporter or editor — I wanted a pitch that was personalized to that person.

And I came back in about half an hour, and there was a spreadsheet ready to go. And I had started acting on it. I don’t copy and paste the pitches; I go and take a look at that reporter’s writing and review the pitch and then make some tweaks to it. But my God, can you imagine how much time that would have taken for me to go out and do this on my own by way of research? That would have been hours and hours.

And instead the agents went out and did it, and then Cowork assembled all that information into a spreadsheet. I was doing other stuff while it was doing that. I wasn’t sitting and watching, because there frankly wasn’t that much to watch. I mean, you could watch the agent tell you, “now I’m going to go look at this.” But, you know, that’s kind of boring. Let it do its thing.

Neville: Yeah. So a question I have related to this, I suppose, is to put it into one practical area, which is: people might think of this in the context of the interaction you have with prompts and the old-fashioned way of doing things that is still prevalent. So how did you — the agents went off and did their thing, and then you came across what they produced and so forth, and it saved tons of time — how did you gain confidence, let’s say, that it was accurate, that there were no hallucinations, no errors? Or is that not the issue anymore with this kind of development?

Shel: I believe that hallucinations would still be an issue. It’s still a model at some level doing this work. I mean, it’s Claude with Claude Cowork. I did install Hermes over the weekend. We’ll talk about that in a bit, but it’s an agent platform, an agent framework, and you create the agents to do things.

For example, I created one over the weekend that I set up to be a weekly job, and it’s going to go out and look at construction industry news to find things based on our areas of expertise where I work, where we have subject matter experts and thought leaders, to find the top three articles that are ripe for newsjacking. If you remember David Meerman Scott’s newsjacking — things where we can get some stuff out there quickly.

Neville: Yeah.

Shel: And take advantage of the fact that this is something that people are looking at and gain some traction over it. So every Monday at eight, it’s going to run this job, and by 8:30, 8:45, it’s going to give me the results. And all of this is through Telegram, or WhatsApp, or whatever app you choose to use to interact with the bot. It still starts with a prompt. The difference is that you’re not prompting a question in order to get an answer; you are telling it what task to perform.

And in the case of the one that I set up on Hermes, it’s now a weekly task. And the interesting thing about Hermes is that it learns as it goes. It continually self-improves based on the more it knows about you and the kinds of tasks that you’re asking it to perform. So I’m looking forward to seeing how that goes. But so far, I just have the one agent running there. But it’s still a prompt at the end of the day.

And in fact, I used — I think it was Gemini — to help me craft the prompt to get the best results I could. I said, here’s the list of requirements, turn it into the best prompt that Hermes will understand and act on most effectively. And it did that. It did a great job. And I’m very satisfied with the results so far. I ran one test of it, so I liked it.

Neville: Yeah. So Claude Cowork is kind of at the heart of this. I’m experimenting myself with Claude Cowork — with Claude generally, Cowork sort of. Nothing like you’re doing with this, I hasten to add. But one of the things that I’m very impressed with about Claude is the way in which you tell it the things about you, who you are and what you’re doing, all this stuff — your preferences in how it conducts what you’re asking it to do — in a way that, unlike ChatGPT for instance, where you have to, in a sense, include in a prompt stuff you’ve already told it for something previously, but you’ve got to do that again. It doesn’t kind of remember that in the same way. Claude is different, though.

So your setup — I mean, I guess what I’m asking basically is, when you set this up, did it require that level of preparation that is probably desirable to do that? Or was there anything special that you had to do that was outside of what you would normally do with Claude Cowork?

Shel: Well, for the byline piece that I was looking to pitch, that I set the subagents out to do their thing in Cowork, I did in the prompt explain what my goal was and what the organization was. I had it look at our company website to get a good sense of who we are and what our areas of specialization are. I gave it some additional information.

But then something I do with all of these now — not every prompt, if I’m just in Claude or ChatGPT, but especially with the agents, with deep research projects and things like that — I’ll say, “ask me questions before you go out and do this.” And it usually asks some very salient questions. It’s very good at deducing what it doesn’t know. And the answers factor into the results you get, which is really interesting to me — that it can, if you ask it to, understand where there are gaps in the prompt that it could use this information in order to deliver really excellent and pertinent results.

Neville: Got it. So thinking about our listeners listening to this, to how you’ve explained all of this — is it kind of credible and within the reach of anyone literally wanting to do this? Or do you need to have some kind of mental preparedness or knowledge technically to do this? Could anyone just dive in and start something? Right.

Shel: Well, I don’t know about diving in. With Hermes, for example, I watched a couple of YouTube videos. I watched one that actually walked me step by step through the installation process and then had a whole section on use cases. I’ve watched more. There’s one on 99 use cases for Hermes that I watched, which was pretty good. So it helps you get in that mindset. But in terms of, can anybody do this?

In the world of communications, anybody better be able to do this, because you’re not going to be sent out to look for these sites and assemble a spreadsheet anymore. You need to be able to orchestrate these agents. And that means knowing how to prompt it to get the results that you want. And that’s different, again, from prompting ChatGPT for an answer to a question, right? You are giving it a task, and it could be a recurring task that somebody on your team does.

Now, in communications, I still don’t see this replacing a communicator, because every communicator is going to have the human-only or human-required elements of the job. I cannot see one of these conducting, say, an employee focus group. There’s so much that we do. I mean, you know, in public relations, the word “relations” always stands out to me, and maintaining those relations is not something a bot can do.

But in terms of what that Harvard Business Review article was talking about, you can swap marketing for communications. I think it’s more true in comms. Comms workflows are more coordination-heavy than marketing. We have legal, we have HR, we have the C-suite. We have to make sure everything’s consistent with the brand and maybe get some brand representation approvals. They’re the owners of the channels that we have to deal with.

If marketing needs a brand code — and this was a concept I really liked in that article — communications needs a narrative code. You know, a machine-readable positioning, machine-readable executive voice profiles, message hierarchies, sensitive-topic guardrails, rules for escalating things that emerge that need to be taken up a step in the hierarchy or maybe up to the C-suite or the CEO. I don’t know anybody who’s built a narrative code.

Whoever builds this first in your organization, by the way, is going to end up owning the agentic stack. If marketing builds it first, we in communications are going to inherit a system that wasn’t designed for crisis communication, wasn’t designed for controversy or reputation damage or stakeholder complexity — it was built for marketing. And that’s the one we’re going to end up having to work with. You probably remember, Neville, in the early days of social media, Richard Edelman was out there sounding the drum that PR needed to own social media before marketing and advertising got their hands on it, because they would turn it into something inauthentic, right? It’s the same thing here.

Neville: Yeah. Yeah.

Shel: I think we in comms are going to have to build out the narrative code and let marketing take advantage of the agentic stack that we’ve built. But we need to be in the room when those decisions are being made.

Neville: So another challenge for communicators, and I can see that. I think the overall structure of the Harvard piece, as I mentioned in the introduction, is on the organization as a whole. And I think there are examples where that’s in work — I quoted a couple, and then there’s the BCG research, which I found quite interesting. But that’s… restructuring is a way away yet on an organizational level, I would say, for most companies. But the individual actions, such as experimentation you’re doing, are definitely right in front of us, literally right now.

And it prompted a thought in my mind, looking at this overall picture, about some assumptions in the Harvard piece that I think are worth looking at for a minute, where the article assumes that strategic judgment remains human, not AI focused, but execution becomes agentic. So I think, okay, then — though history suggests automation rarely stops neatly where people would like it to and where they would expect it to.

So perhaps a question that’s relevant to address in this context is: if AI systems — agentic is part of that — increasingly assist with strategy too, which is what they’ll be doing, where exactly does human value migrate to? That’s a broad question, but for communicators specifically, how would we address that one?

Shel: I think, first of all, if you’re going to look to the agentic system to assist with the development of the strategy, I would sit down and map out a game plan for that. I wouldn’t just say, “hey, you know the company I work for, come up with a strategy for us.” I would say, first of all, what is this strategy…

Neville: Ha ha ha.

Shel: …going to be designed to achieve? What do we know about the direction the company’s going and decisions that have been made? I would certainly use it to go out and say, research the marketplace and research our competitors and identify, to the extent that you can, what their strategies are. I would develop the strategy myself, but I would give it to the AI to stress-test.

And by the way, some of this is agentic and some of it is just querying a chatbot. I mean, let’s just take crisis communication as an example. No CEO is going to go into a boardroom with an answer from an AI system telling a leader something they don’t want to hear. That is amplified by the agentic stack. If we go in as the crisis counselor and say, “look, I know you’re not going to like this. Here’s my judgment. And I’ve got this information that came from the weekly analysis of sentiment in the marketplace,” so I think it can bolster your argument. It can’t replace your argument. You’re going to walk into that boardroom as a human and make a case.

Same thing, maybe, with focus groups. When passive signals in social media, for example, and message boards get gamed, sitting in a room with 10 employees becomes the truth that the dashboards that are out there — the agents that are out there looking at sentiment — get checked against. So when a dashboard says that morale is great and the focus group says it isn’t, I’m going to pay attention to the focus group. I’m going to pay attention to those 12 people in the room before I listen to an agent that says, “well, we’ve been analyzing all the sentiment in Slack and email, and everything is just dandy.”

So I think it’s the same with strategy. I think I would never abdicate strategy…

Neville: Mm.

Shel: …but I could certainly develop it faster and be more confident in its viability by using agents and chatbots.

Neville: Yeah, I agree. And it makes me think of, I guess I would say, what’s coming, which is already here in ways that lead to even greater — well, integration, I suppose, is the right way. I’m thinking what you said at the beginning of this segment we’re talking about now, which is, you don’t hand the whole thing over to the AI and say, “hey, go and develop a strategy.” You would do…

Shel: And you know there are people who are, right?

Neville: Yeah, they will. They will. But it seems to me that this is really, in a sense, the fulfillment of an expectation — a promise — from artificial intelligence tools like this, that you would have a conversation with it in the same way you would with a human being who might be an external consultant or a colleague who’s a subject matter expert or whoever it might be, that you would explore with that individual: we’re developing a strategy for next year, let’s look at how we’re going to do this.

You set the framework for how you might start that conversation with your AI assistant. And as you said, this is not specifically agentic; it’s the whole spectrum of what the tools are. And you set it on course to go and research this. And that’s probably what an agentic tool will do. And that to me is the excitement of where this is going — that you can get to that stage, which then I think would address some of the skepticism and indeed alarm bells by some in organizations when they see unfettered technology going all over the place or being asked to do stuff. This, though, makes it credible and gives it some legs of credibility.

Which leads me, I guess, to possibly the final question here. We’re seeing this, as you’ve explained, this is light years ahead of the demo you gave a year ago, which gave a signal, a strong sense of what’s possible, where this could go. We’ve seen that fulfilled. It is eminently possible. And you don’t need to be a rocket scientist, as you might have expected you would have to be a year ago. This is doable. And the more people experiment with it in simple ways, like you’ve outlined as a real-world example, they will want to do that in that case.

So the question then, therefore, is: okay, fine, a year on from last year, you’ve explained something you’re doing that delivers value quite readily every Monday morning, let’s say. So what’s next, do you see, in terms of developing technology and the developing value people will get from it that would accelerate probably its uptake? How do you see it?

Shel: I think that the next thing we’re going to see is an evaluation of every role and where an agent will fit. This is something we went through a couple of years ago. Ethan Mollick was talking about it in his book, Co-Intelligence, before we were even talking about agents — talking about inviting AI to the table and figuring out where you could work it into your workflows. But it was still the chatbot. It was still the, “I’m going to ask you a question and you’re going to deliver some kind of answer.”

I think we need to do that again and look at agents. What tasks are we performing, and which ones can we hand off to agents? And I think there are probably roles where this is going to be even easier to do, where you’re going to see more opportunities than in communications. I mean, you know, engineering, for example, I think is wide open for this sort of thing.

So I think that’s what’s next — as we do hand off certain (and I’m going to call them) mundane tasks, because this is not the high-level strategy and the human-touch stuff that is so important in so many jobs. But as we hand these off, and it now takes an hour instead of a week, what does that do to the rest of our workflows? What does that do to our organizational structure?

One of the things that I was reading over the weekend was the expectation that middle managers are going to be a thing of the past, because what do they do? They handle the flow of information up and down between the people who report to them and the people that they report to. They handle a lot of mundane tasks that might now be handed off to an agent. Agents, according to — I don’t remember who this was who was saying this. It was somebody noteworthy. It might’ve been Dario Amodei at Anthropic, but I honestly don’t remember for sure — but middle managers can be replaced by agents by and large.

So what does that do to organizational structure? Certainly flattens it. But now, in terms of those executives who have a lot of people reporting to them, what part of that reporting structure can be handed off to an agent? So I think this is sort of a cascading situation where everything we do leads to a reconsideration of something, that leads to, well, what else can we do with the agents, which leads to further reconfiguration?

I think that’s what we’re looking at. And I don’t think it’s going to happen overnight, because, as you alluded to, the technology may be moving fast, but organizations tend not to, particularly when it comes to issues of structure and governance.

Neville: I think this is so exciting, to be frank — the idea of the changes we can see coming that will be painful for many. But is it more structural change? It’s a constant in our lives, is it not, with all of this? Something we should embrace emotionally and logically, that we can control this. And I don’t mean control the tech — we can’t do that. But we can control the risk and the benefits of something like this by not reacting to something that’s coming, by, in a sense, embracing it and experimenting with this and learning it. And as you said, if we don’t do this, the marketing guys will. And so we can’t have that. I think…

Shel: And then we’re stuck with theirs.

Neville: I think it’s something to really pay attention to. So this has been a useful, interesting discussion, Shel, getting your thoughts on this in particular. So yeah, I think we’ll come back to this conversation unquestionably at some point in the future.

Shel: No doubt, as we see developments. In fact, as I say, I just started working with Hermes over the weekend, and it was an eye-opener, and I expect, as I work with it more, I’ll have more thoughts about it and my thinking will evolve. I should point out that I did install this on a personal virtual server, not on a company computer. I’m not taking that kind of risk. And it’s my personal account.

One other thing I thought I’d mention — you talked about the idea of having a conversation with the AI, and I think that’s becoming more of a focus. And I’ll give you two quick examples. One I already mentioned is with Hermes: you don’t go to a terminal and engage with it or go to its website. You do this through WhatsApp or Slack or, in my case, I’m using Telegram — just like I’d be having a conversation with a person in that same app.

But on, I think it was Thursday, I did a half-day webinar that was offered by the Marketing AI Institute, Paul Roetzer’s organization, and it was on AI for writing. And it was very interesting. Chris Penn was among the speakers; he did a great job, as always. But one of the folks there talked about, you know, have the conversation with AI for real — do it with your voice, not with your keyboard. And she talked about a tool, which I haven’t used it yet — I have installed it across my personal computer, my laptop, and my phone — called Wispr Flow. It’s an AI tool. Have you…? It’s pretty cool. I mean, in any tool you’re using, you just click it and talk. And it doesn’t go directly into the chat box; it interprets it…

Neville: Yeah, I’ve been using it. Yeah. Yeah.

Shel: …and then puts the best prompt based on what you just said into the box. And that’s what you use to prompt the model. And I’m looking forward to giving that a try. And it’s called Wispr Flow, by the way, because if you’re in the office in an open-space format and you don’t want to disturb the people next to you, it understands what you’re saying when you whisper to it.

Neville: Yeah, it is interesting. I’ve got a hurdle to jump with it, though, which is getting accustomed to speaking what I want things to be done and how, rather than typing them. You know… yeah, and I haven’t got across that hurdle yet. That’s limiting my use of it. So I’m reverting to the, well, I’m more comfortable typing, I can type fast and all that kind of stuff. But, reality, this is faster than that. And it is…

Shel: Yeah, same.

Neville: I recognize the benefits of it. I can see this. Not everyone will be used to this. This is not dissimilar to the argument we could have about voice notes. I know people who love voice notes; I don’t. And I know more people who don’t like it. It could be a generational thing, I think to myself. But it’s part of the communication landscape. So you need to get accustomed to these developments.

Shel: Yeah. And I hear about voice notes being preferred by some reporters who are being pitched, because it’s evidence that it wasn’t AI slop that’s pitching them.

Neville: Yeah, yeah, yeah. Yep, yep, yep.

Shel: And that’ll be a 30 for this episode of For Immediate Release.

The post FIR #513: Why Communications Must Build the Narrative Code for the Agentic Age appeared first on FIR Podcast Network.

While there’s no evidence that business leaders are outsourcing the most important decisions to AI, there are reports that many executives are relying on AI to make many — in fact, most — of their decisions. The implications for communications could be huge.

Links from this episode:

• AI Is Changing More Than Work, It’s Rewiring Executive Decision-Making
• Inside the C-suite: How AI is quietly reshaping executive decisions
• AI and the future of human decision making
• C-Suite Executives Dominate AI Decision-Making as Strategy Becomes Priority
• Decision-Making by Consensus Doesn’t Work in the AI Era
• How AI Is Transforming the Way Executives Lead
• Leadership at a Turning Point: How AI Is Shaping Executive Decision-Making
• Can AI Make Executive Decisions?

The next monthly, long-form episode of FIR will drop on Monday, May 25.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville: Hi everybody, and welcome to episode 512 of For Immediate Release. I’m Neville Hobson.

Shel: And I’m Shel Holtz. The inspiration for this week’s report came from a post Brian Solis wrote recently. In it, he argued that AI isn’t just changing work — it’s rewiring how executives make decisions. Once Brian put that in my head, the trend started standing out in other things I was seeing. I’ll summarize the numbers and what they mean for communicators right after this.

The numbers Brian pulled together are honestly alarming. A Confluent study of UK private sector leaders found that 62% of executives now use AI to make the majority of their decisions. That’s not some — it’s the majority. 70% say they second-guess themselves when AI disagrees with them, and 46% say they rely on AI more than their own colleagues.

On the U.S. side, SAP’s research found that 44% of C-suite executives would reverse a decision they had already planned to make based on AI input. 74% place more confidence in AI advice than in the advice they get from family and friends. Meanwhile, McKinsey reports that 92% of companies plan to increase their AI investment over the next three years, but only 1% — 1 percent — describe themselves as mature in deployment. The money to pay for AI and a sort of blind trust in its abilities are racing ahead of the internal competence to use it. Now, I want to be clear before I go on. I’m not anti-AI, Neville — you know this. Anyone who listens to the show knows I’ve been beating the drum for AI as a tool for communicators and for business in general for a long time.

AI as a thinking partner, a research assistant, a stress-tester for ideas — that’s enormously valuable. But there’s a meaningful difference between using AI to inform a decision and using AI to make the decision. And Brian puts this well: AI is becoming the new executive influencer. The problem is that it hasn’t earned that role, at least not yet.

So let’s talk about what this means for those of us in communication, because the implications are everywhere. Start with employee trust. The implicit deal between an organization and its workforce is that the people at the top got there because they have judgment and experience and pattern recognition that the rest of us don’t have — or at least they’ve been able to employ it really well and get noticed by the people who promote you into those leadership decisions.

That’s the story leadership tells, and it’s the story employees buy into. Now imagine the all-hands where the CEO announces a major restructuring, and somewhere in the Q&A, or worse, on Blind or Reddit a week later, it comes out that the decision was essentially handed to a chatbot. What happens to confidence in leadership? What happens to engagement? What happens to the social contract that says, follow me because I know where we’re going?

You can’t credibly ask people to bring their full selves to work, as they say, while you’re outsourcing your own judgment to a language model. Now extend that to external stakeholders — investors, customers, regulators, the board. They’re paying, and in a lot of cases they’re paying a lot, for executive judgment. If a strategic call goes sideways — and you know that happens — the explanation that the AI suggested it isn’t going to land well.

It’s going to sound like an abdication, because it is an abdication. And from a crisis communication standpoint, “we trusted the algorithm” is one of the worst defenses I can imagine. I don’t expect that anybody’s going to say that, but it doesn’t mean it’s not going to come out. Just ask anyone who’s worked an aviation incident, a financial services failure, or a healthcare AI misfire. Imagine the reaction when either the leader tells people, or they learn through a third party, that the afflicted stakeholder hears, “Well, that’s the decision the AI told me to make.”

And there’s a third implication that I think communicators need to surface inside our organizations: the erosion of dissent. I find this particularly interesting and disturbing.

Confluent found that 65% of leaders say decision-making has become less collaborative since adopting AI. The Harvard Business Review just ran a piece arguing that consensus is dead in the AI era. That may be — but debate isn’t consensus. Debate is the friction that exposes bad assumptions. It’s what didn’t happen at that auto manufacturer — I think it was Volkswagen with their emissions standards. They didn’t have the psychological safety to feel safe in dissenting against the decisions being made. In this case, we’re not even looking forward at the leadership level in some cases. If AI is pushing aside the colleague who would have pushed back, whatever process your organization had for dissent just stops functioning. And when dissent dies, so does the early warning system communicators rely on to spot reputational risks before they get out of control.

So what do we do? A few things. We push for governance — and if you already have a governance model, push to revisit it. Your governance needs clear declarations of which decisions AI informs versus which ones it actually makes. We coach our executives to talk publicly about how they actually use AI, with appropriate humility, before the question gets asked for them.

We build the internal narrative that human accountability is non-negotiable, no matter how good the model gets. And we keep reminding leadership that machine confidence isn’t the same as strategic clarity. Brian’s right: AI is a test of leadership. It’s also, increasingly, a test of communication. Neville?

Neville: Well, just to set my position clear on this, too — I’ve been a drum-beater for AI as a research assistant, as a useful tool, since GPT first came out. The initial kind of hysterical enthusiasm was tempered over time, but I use the tool every single day in what I do for work, or for pleasure for that matter. So it’s something I believe strongly in. But I’ve got this, how could you say, in the back of my mind always — this thought that I don’t accept blindly anything the AI assistant tells me. If I’m researching something, for instance, I’m going to make a recommendation about something, let’s say, or I’m writing a report or even something relatively simple like an article for the blog. If I felt I wanted to say this and it’s telling me that, that’s a simple decision: I’m either going to follow it or not. Typically when that happens, I’ll ask it questions to further that angle. But this is something else, what Brian writes about. And The Register — I’ve read their piece — tempered with a bit of hysteria, it seems. I mean, this is a very alarmist piece, or argument, you could say. If it’s saying, as it is — the survey that The Register reports on — 62% of leaders of private sector companies, and according to The Register that’s owners, founders, CEOs, managing directors, the C-level leaders of various types of companies. They didn’t say sizes. But they use AI to make the majority of the decisions, which leads to some of the alarm bells ringing that you outlined. What if it gets out that the AI made a decision when something goes south? You could flip that. What happens if it gets out that an amazing decision that led to the company being massively successful was actually made by an AI?

I think it’s inevitable you’d have that sort of focus on it alongside more sane arguments, perhaps. You could argue, well, that CEO is pretty smart that he used an AI to help him do that — as opposed to the other side, which is, gee, we’ve got to fire this guy, he used an AI and it went wrong. So you’ve got to put some balance there. Also, I think you mentioned this earlier, and I agree with you, that there are two angles to every question we might ask about this. One is internal, within an organization, and the other is external. So it is an interesting point. And one thought I had in my mind, the pragmatic question: if a leader changes a decision he or she has made because the AI assistant suggests something different, who actually owns that decision in the end? In fact, whether he changes his mind or not, if the AI said, “I recommend you should do this, and here are the 10 reasons to support that idea,” that are different from what the leader was going to do, and he or she made the changed decision based on that — who actually owns that decision? Or, as I asked myself, is that really the most important question to be answered? But it’s still a natural one to arise. And yes, we could run through a long list of the implications in this scenario for the employees of the organization, other stakeholders, and the external audiences. But I have to say Brian’s arguments are well made. He sets the scene — the executives are relying heavily on AI. From there it goes more into the alarm function.

Judgment being reshaped — the judgment exercised by a leader is obviously so flaky that it can be reshaped by the AI assistant. In other words, that individual is willing to let that happen. I wonder whether this is all part of, perhaps, the speed with which people are expecting decisions to be made. Indeed, something I was doing this weekend — we’re on a holiday weekend here, by the way, so I had time to do this — that was nothing to do with work. It was a personal thing I was involved with that required analyzing a document that had a lot of financial information in it. I asked my AI assistant, in this case Claude, as part of my experiment with Claude, to summarize it and pinpoint the key aspects. It did that in about 20 seconds. And that was enough for me to know what questions I would need to ask it next, to develop it the way I want, rather than starting from scratch trying to do that. So there’s the benefit. But I think treating AI like a trusted advisor, to me, makes a lot of sense. And I’m trying to balance that thought with the alarmist approach — you know, this is a bad thing, all these terrible things are going to happen, and it will all come out. So how does that gel with treating AI like a trusted advisor? Although your point, I agree, it hasn’t earned the trust in the context of this conversation. So does it mean leaders are willing to override their own decisions or instincts based on AI input? Well, according to The Register, 62% have said they are, I suppose. If that’s true, I think we’re in trouble already, before this gets any further. So the real challenge — I think you’ll agree with this, Shel — is not the tech at all. It’s the leadership aspect, the human behavioral aspect of this, as is so often the case. When people talk about the relationship between the human and the AI and they just talk about the tech, it’s not — it’s a human issue. Cut through all the alarm bells and pluck out something which to me is extremely important, that really doesn’t get much airtime in Brian’s report at least: isn’t this really about the whole point of judgment? That someone in a leadership position in an organization is in that position partly because he or she is very good at exercising judgment in the work they do or the decisions they make. Are we saying that judgment is so fragile that an AI could just overturn all of that in an instant and lead all this? I guess my point is that I’m noting this. I listened to what you said. I haven’t read all the surveys you mentioned, or the other reports — the Harvard Business Review, for instance — I will. But I find this literally the worst-case scenario, and that’s being pitched as, you know, this is upon us, based on The Register, which, by the way, has a — let’s call it interesting — reputation over the years for some of their reporting. But this is very factual; their own report is actually quite well written. So what do we make from this then? Should we be worried? I don’t think we should, if we see this as simply something to note and look at as a communicator — let’s say the role you’ve got in ensuring that the CEO isn’t going to have his or her judgment completely overwhelmed by an AI. I just find the idea of that frankly ridiculous, in the sense of, well, not implying or even saying that this is the norm. It’s a result of surveys. There’s other research also supporting some of this, I think. But we should put it in perspective: this is, I guess, an inevitable discussion point that’s emerging at this stage in the development of AI and organizations. We’ve reported recently on this podcast how leaders are taking ownership of the AI deployments in their organizations. That doesn’t mean to say every company is doing this, because they aren’t. But we’re seeing that, and then we’re seeing other reporting we’ve commented on — that employees and other stakeholders related to an organization are unhappy with what’s happening with AI rollouts in their organization. So you’ve got all these mixed messages coming left, right and center, and now this. It doesn’t mean we should — oh my goodness — stop doing this, or have a meeting with the CEO and say, “What are you doing?” No, I don’t think so. But we need to note this nevertheless. I don’t believe this is something we should all get terribly alarmed about, to be honest, as long as we apply our own common sense to observing what’s going on and making sure we understand the CEO we’re supporting as communicators — let’s say the leadership teams — that this isn’t happening.

Shel: Well, I don’t think this is the most important issue we’re facing with AI. I do think it’s a time to worry. Now, I will say I don’t imagine that the CEOs leading the world’s biggest companies — the Jamie Dimons, the Josh Domaros, the Tim Cooks of the world — are using AI to make important decisions. And you have to wonder, because I don’t think they asked, in the survey they did, what types of decisions these CEOs are making. Are they the game-changing decisions, the most important decisions they have to make, or are they lower-level decisions? We talk about AI taking all that drudge work off the table. Are they allowing the AI to make decisions associated with that kind of work? But I think, as people — and CEOs are people — as they get accustomed to letting AI make decisions, it might get easier and easier to turn bigger and bigger decisions over to AI as time goes by. With any luck, AI is going to get better and better and may earn that trust. But this would cause that decision-making instinct that leaders have, based on their experience and their judgment and the other things that got them to that level, to atrophy. I mean, atrophy is happening elsewhere as a result of AI among some groups of people — the ability to write your own thoughts down, to craft your own email, to conduct your own research. As far as CEOs making good decisions with support from AI, I think support from AI is going to become table stakes. I think CEOs who don’t know how to use it are going to become dinosaurs in fairly short order — not necessarily the ones who have the job now, but I don’t think you’re going to see people getting promoted into that position, or hired into it, if they don’t know how to use AI for decision support and the other things we see AI being used for very effectively at leadership levels. And leaders are using AI, according to most of the research I see. I wonder, though, if they start turning more and more decisions over to AI, what is the board or the owner going to see as the value of the CEO? If most of this work — or much of this work, the majority according to that Confluent study — is being done by AI, does that mean the enormous salaries being paid to the people at the top of the organization are going to decline? Or does it mean that the role changes altogether, or maybe even ceases to exist in favor of some other model? And by the way, I’d love to see the same question posed to people at other levels of the organization, because this probably is not something confined to the C-suite, this turning decisions over to AI. I wonder how much it’s happening in middle management. I wonder how much it’s happening among frontline workers. If it’s at the same level, then it’s a company-wide issue that needs to be addressed, because there are going to be some problems that emerge if we don’t — I mean, along the Volkswagen lines with their emissions scandal. Dieselgate, exactly. Yeah.

Neville: That was Dieselgate, as it was dubbed. I mean, it’s a good point you make. I agree. And the point you made earlier, too, is actually a critical question: what kind of decisions are we talking about here? Is it on the scale of, let’s proceed with the merger with this company rather than that one? Or is it something like, should I fit in a stopover in this city on my way to that city to meet with these people and so forth and achieve these things? Is it that? Or is it even something more prosaic? You know, what do I get my wife for her birthday next week? I’ll have my secretary do it — but the AI could tell me. I mean, that’s ridiculous, actually. But it’s significant to know what kinds of decisions we’re talking about, because I’ve not seen it referenced. It’s implying — and people are jumping, obviously, on this — that these are the kind of organization-affecting major decisions that are suddenly at risk because an AI is doing it. I find that ridiculous, to be honest. So we need to know what kind of decisions.

Shel: Yeah. I mean, in my industry, there’s a go/no-go decision on pursuing a project. I cannot imagine, in my wildest imagination, in my organization, anybody turning that decision over to an AI. But what if somewhere in the industry they do, and end up pursuing a project that ends up being more trouble than it was worth? Somebody in the organization at that leadership level, who was involved in the previous discussions, would have known for various reasons, but the AI didn’t have the experience and the insight that that individual had. That could be a financial problem for the organization.

Neville: So the role of the communicator in all of this — and this is not to say that the communicator who works closely with the leadership teams, including the CEO and others in the C-suite, is involved in every single thing they’re doing. No, that’s not realistic, because they’re not. But the communicator’s role in preserving human judgment is the right question to ask. What is it in this context? Where do communicators fit in helping leaders balance AI insights with human insight and judgment and experience? Where do they fit in doing all of that? So the two angles I notice: internal comms — communicators act as sense-makers, ensuring context, ethics and human impact remain part of decision-making. Externally, they help articulate how AI is used responsibly in the organization, which is increasingly central to trust and reputation. That addresses the point you made about when it leaks and it gets out that AI did something. I think increasingly we’re going to see that point — articulating how AI is used responsibly in an organization — because the impact can be huge if rumor builds, which it would do: “the AI is making all the decisions in this company, and why do we need the CEO and all that?” So that’s a good role for a communicator to take on, and to be seen to be the person who is the “yes, but” person and the key advisor to leadership in these things, which strengthens the communicator’s role, in my view. So there are things we can do to address this. If this is as big a problem as these articles make out, I don’t believe it’s something we should lose any sleep over right now in the context of everything else that’s going on in the organization. But nevertheless, we’ve got respected sources — Harvard Business Review, we’ve got Deloitte talking about it, and others that we pay attention to because they’re credible publications talking about this.

Shel: Well, yeah.

Neville: Brian seeded an interesting discussion point, it seems to me.

Shel: Yeah. And let’s look at a very plausible scenario. Let’s say somebody sues the organization over a decision that the CEO made, or that leadership made, that affected them badly, and they feel they deserve compensation for that. In the U.S., anybody can sue anybody for anything. And we have seen some recent lawsuits. Look at the lawsuit that we’re seeing play out right now between OpenAI and Elon Musk.

Neville: Yeah.

Shel: And look at the records, the emails that have been surfaced in discovery. Look at the trials that have been held over lawsuits brought by the parents of children who killed themselves because they got encouragement or assistance from ChatGPT, and who sued OpenAI over that. What they got in discovery was access to the kids’ entire ChatGPT history. So you have a shareholder or a customer who sues the company, and in discovery, all of these things come to light — and that’s how it gets out. So I think even decision support has to be balanced with other input that you can demonstrate in a courtroom influenced the decision that was made, so it doesn’t look like the decision was completely outsourced to the AI. I think that’s an entirely plausible scenario in a lawsuit. So yeah, it’s something we need to consider. And as you say, and as I said, there are things communicators can do about this. One is making sure people are aware of the potential for this situation. And then, as I said, influencing the governance model so that it incorporates decision-making — if it doesn’t already have decision-making and decision support in the governance document, it needs to be added. And then making sure the leaders are talking about how they’re using it, so it never comes up that they’re using it to make a decision of importance in the organization — that it’s focused on using it in very effective ways.

Neville: Yeah. I mean, I think the picture you painted — lawsuits and stuff like that — are very possible, particularly in America, where, as you said, anyone can sue anyone for anything, usually for amazing sums of money, in the billions. So maybe what needs to happen in organizations that would address this, among other things, is keeping records. So that, for instance, in an organization that has deployed or rolled out AI tools such as chatbots — let’s say maybe their own version of something based on ChatGPT, whatever it might be — it needs to be known that those record anything you interact with on an AI. Whatever level you are in the organization, there’s a record kept along with anything else: emails, internal reports, you name it, they’re monitored and tracked in most organizations. And the fact that you could add to that picture even some of these automated note-takers, like Otter and others, that are commonly used in intrusive ways in Zoom meetings — and you hear stories of private Zoom meetings —

Shel: AI transcripts of Zoom meetings in which the decisions were made.

Neville: — where the outcomes are disclosed or leak out publicly because someone used one of these tools that summarized things, including the recommendations or suggestions if they were made by anyone. If that gets into a law case by the plaintiffs, that’ll be shown out of context — you can be sure of it. So, right.

Shel: Yeah. And that’s why a lot of organizations are saying to their employees, you can’t record these kinds of meetings.

Neville: Right. But someone will, and it’ll happen. So you need to head that off the path, as it were, and have your own structure in place and your communication surrounding it. So, for instance, you have to have very clear narratives around decision ownership, for example, that would help you in crisis situations. That’s the internal focus. Externally, you’ve got to communicate the kind of structure you have for human accountability — not “the algorithm said we should do this.” We can laugh about it, as I am at the moment, but imagine the reality of something like that happening. So I think these are all things that are plausible, I do believe, particularly in the U.S., I have to say — but hey, could be anywhere. It isn’t complicated to work out a plan of how you would prepare for things like this. But I’d rather look at it not as preparing for worst cases, although you need to. It’s just a switch — flip it over a bit and look at the benefits of all of this. And again, not solely the communicator: the individual leader has to be willing to go along with this, has to be willing to share some of the thinking he or she is doing and the discussions with the assistant, whether it’s an AI or anyone else, to realize that you can’t do this without full transparency, at least to your advisors, including the communicator.

Shel: Yeah, absolutely. And we will be back with a follow-up episode when the inevitable headline surfaces of a company that gets in trouble because it’s revealed that the CEO abdicated a decision to AI. Until then — actually until next week — that’ll be a 30 for For Immediate Release.

The post FIR #512: The AI Shift in Executive Decision-Making appeared first on FIR Podcast Network.

The policies are clear and well communicated. The guardrails are firmly established. Every last employee has been trained. And someone in your organization still releases a public document riddled with AI-generated errors. What went wrong has nothing to do with technology and everything to do with internal culture and accountability. In this long-form April episode, Neville and Shel examine a company that seemingly took all the right steps yet still had to apologize publicly for a court filing riddled with hallucinated citations. Also in this episode:

• Gartner predicts that, by 2028, 75% of employees will rely on an internal chatbot to get the news that matters to them. How will internal communicators need to rethink their role to ensure everyone knows and understands what they should in order to achieve strategic alignment?
• One of the promises AI executives have made is a leveling of the playing field, giving lower-level employees the opportunity to excel and rise through the ranks. According to one new study, exactly the opposite has been happening.
• PR hacks have been accelerating the pace at which they churn out press releases and pitches. That has raised the bar for what it takes to earn a journalist’s trust (and journalists do still rely on press releases, according to a survey of reporters).
• Apple’s announcement of its CEO transition offers communicators a clinic on how to announce a new top executive.
• “Slopaganda” from Iran has proven remarkably effective, which means it is undoubtedly coming for your company or clients soon.

In his Tech Report, Dan York outlines big changes coming with WordPress’s next update.

Links from this episode:

• Elite law firm Sullivan & Cromwell admits to AI ‘hallucinations’
• Sullivan & Cromwell law firm apologizes for AI ‘hallucinations’ in court filing
• Letter re: In re Prince Global Holdings Limited, et al., No. 26-10769
• Sullivan & Cromwell Just Put Every Firm on Notice. And S&C Advises OpenAI on Safe AI Use.
• An AI Screw-Up By… Sullivan & Cromwell?
• LinkedIn search results for Sullivan & Cromwell AI
• AI, Trust, and the Reinvention of Corporate Communications: Inside Gartner’s 2026 Playbook
• Does your intranet still matter in an AI-first workplace?
• Chatbots in Internal Communications: Game-Changing Wins
• How AI Chatbots Are Redefining Internal Communications?
• The future of internal communication: How AI is changing the workplace
• High earners race ahead on AI as workplace divide widens
• Sarah O’Connor: One early view about AI was that it would share…
• How AI is forcing journalists and PR to work smarter, not louder
• What journalists want from AI-assisted PR pitches
• Journalists Trust Human-Written Pitches Over AI
• Journalists Reject AI-Generated Press Releases As Untrustworthy
• What communicators can learn from Apple’s CEO transition announcement
• Tim Cook to become Apple Executive Chairman; John Ternus to become Apple CEO
• Iran’s Meme War Against Trump Ushers In a Future of ‘Slopaganda’
• Iran’s ‘slopaganda’ team uses AI Legos to flood social media
• Slopaganda wars: how and why the US and Iran are flooding the zone with viral AI-generated noise
• Slopaganda Comes of Age
• Alberta separatist leader unconcerned about influence of YouTube ‘slopaganda’ videos

Links from Dan York’s Tech Report

• WordPress 7.0 Source of Truth – Gutenberg Times
• WordPress 7.0: Real-Time Collaboration Arrives in Core
• WordPress 7.0 Release Party Updated Schedule

The next monthly, long-form episode of FIR will drop on Monday, May 25.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Shel: Hi everybody and welcome to episode number 511 of For Immediate Release. This is our long-form episode for April 2026. I’m Shel Holtz in Concord, California.

Neville: And I’m Neville Hobson, Somerset in England. We have six great stories to discuss and share with you this month and to delight and entertain you, we hope. Topics range from the consequences of not following company guidance on AI use, chat bots, employee use, and the workplace divide, using AI to work smarter, what we learned from Apple’s CEO transition announcement, and the future of slopaganda. Lovely word, that one, show. Plus, Dan York’s tech report.

But first, let’s begin with a recap of the episodes we’ve published over the past month and some listening comments. In the long form episode 506 for March, published on the 23rd of March, our lead story was on Anthropic’s view that AI will destroy the billable hour, a topic we’ve talked about before on FIR. We also explored digital monitoring of employee work, Gartner’s prediction that PR budgets will double next year, the escalating misinformation crisis, and Cloudflare’s prediction that

bot traffic will exceed human traffic by 2027. That’s next year, by the way. On LinkedIn, you’ll find no shortage of posts stridently deriding the notion that anyone should ever use AI to write them. In FIR 507 on the 30th of March, we rejected roundly that idea and looked at the actual trends in using AI for writing. And that prompted some comments from listeners, right?

Shel: Yes, it did. Starting with Susan Gosselin, who’s actually with a client of mine back in my consulting days. She writes, there are many types of writing that I think AI is great for interpersonal communications, summaries, et cetera. But for marketing writing, that’s another thing. There are issues of copyright to consider and what you’re feeding into the channel.

This article from Jane Friedman, and she’s linked to it, and we’ll include that link in the show notes, is aimed at authors, but it does have implications for marketing writers too. For instance, I work for an American IT MSP, that’s a managed service provider. Let’s say that an MSP in Spain that does our line of work sees our website and our authoritative blogs and e-books and likes it. They decide to run our whole English website into Spanish using an AI translator.

then make a few tweaks and publish. There’s not a lot to stop them. There’s also the issue of being able to defend your copyright overall. The law is not yet fixed and the risks are real. Then Steve Lubetkin writes, I find AI particularly helpful for rote tasks like organizing lists, transforming Excel spreadsheet columns, and summarizing interview transcripts. It’s also great for brainstorming ideas when it suggests perspectives I hadn’t thought of.

but ultimately it comes down to using it as a tool for further human intervention, not less. Neville, you responded to that saying that’s a great way of putting it, Steve. Those rote tasks are exactly where AI seems to shine, the kind of work that takes time but doesn’t really benefit from deep human creativity. And I agree on brainstorming too. It could be surprisingly good at surfacing angles you might not have considered. I do this a lot.

Your last point really nails it though. It’s not about removing human input. It’s about focusing it where it matters most. Used that way, AI doesn’t diminish the work. It can actually elevate it. And finally, we have a comment from Yorma Mananan who writes, AI can help people escape from the writer’s block. So why not use it to get started?

However, writers must own all content created with or without AI so the content doesn’t sound like you, you shouldn’t publish it. The challenge is to learn to speak machine English with AI. Define clearly why you were writing, what you want to say, and what you want your readers to do after reading your content. Without your strategy, AI can’t produce quality content that sounds like you. Strategy first, AI second.

And Neville, responded to Yorma. You said, I like how you framed this using AI to get past the blank page is a very practical use case. That starting friction is real for a lot of people, and AI can lower the barrier quite effectively. Your point about ownership is key too. If it doesn’t sound like you, it isn’t really yours, regardless of how it was produced. Where I’d add a layer is around your machine English idea. I see it slightly differently. Rather than learning to speak machine,

I think the real shift is learning how to think with the machine, using it to clarify intent, test structure, and challenge assumptions. But I agree with your conclusion. Strategy first, AI second. Without that, you’re just generating words, not communicating. And Yorma responded to you saying, agree. Machine thinking is a better way of describing the conversation relationship with AI.

Neville: Good comment!

Great. It’s excellent to have that. interesting, Shell, that it illustrates to me something. It’s not a trend at all, but I’ve noticed recently in other posts I see on LinkedIn that address this kind of topic. Increasingly, there’s people leaving comments that are basically saying that you own it, not the AI. And AI assists you in communicating, not creating the final stuff, essentially, which is what some of these comments are alluded to.

Maybe people are waking up to that more than they have been in the past. It won’t silence the big critics. We’ve already seen that because, you know, it’s going to be criticized no matter what. But the more people who talk up the reality of what we all talk about, which is this is an assistant. It’s a tool to help you and communicate more effectively. It enhances your ability in that context. And then you’ve got Steve’s talking about, you know, doing stuff with Excel and all this kind of thing.

I’m in the middle of an experiment on the middle. I’m still at the 10 % start of experimenting with Claude Pro that I know you’ve been using for a long time, but I’m taking this very, very one step at a time. And it’s very non techie my focus. But one thing I have noticed is comparing as I have been doing, let’s say a prompt in the simplest form, a chat prompt to Claude compared to one to ChatGPT.

The differences are truly startling in many cases. Claude typically is richer and deeper in its content with the same prompts. Now, of course, there’s variables at play here. ChatGPT knows a huge amount about me. Claude does too, because there’s a nifty tool that imports everything from ChatGP to tell it. But I’ve also added stuff. So it’s done that. It’s missing on some levels, though. And that’s probably because it doesn’t yet know totally enough about me to do this.

This is something you notice when you do this kind of thing with different tools. So I mean, that’s not the main thing about Claude that wow’s me, I must admit. Cowork and some of the other tools I’ve been touching on, but Cowork I’ve spent quite a bit of time on. But so I’m sure we’re have lots more conversations about this as we talk topics. Let’s see what comes out of today’s menu of topics. So thanks for those comments, everyone.

Let’s see, bad, no, that’s not the right one, this one, when workers lose their jobs, many turn to gig work to earn income while waiting for new opportunities. Increasingly, companies are hiring gig workers to create content and train AI systems. This raises various communication and ethical issues. And in FIR 508 on the 8th of April, we explain what’s happening and discuss the implications.

Then when bad actors use AI tools to clone a musician’s voice and upload synthetic versions of their songs, they can then file copyright claims against the original artist’s content. In FIR 509 on the 14th of April, we break down how this scam works, why it matters to communicators, and what you should be doing right now before an incident forces your hand. And you have some comments here.

Shel: We do two of them, one from Eric Redicop who identifies himself on LinkedIn as an entertainer and artist wrote that AI cannot use my work because it’s not posted online anywhere. I have to do it this way because YouTube allowed bogus copyright claims on my work and shut down my channel five times. then Ray Baron-Wolford, who is a CEO at a charity organization said, this is why it’s so important that every artist signs up.

to all copyright protection services.

Neville: Yeah, that’s a good point. I think the the first commenter though talked about a genuine issue, a genuine issue. And I’d wonder if, you know, if he’s saying that this can’t happen to me because none of my content’s online. I wouldn’t rely on that 100%. Actually, I wouldn’t. No, I wouldn’t. And that second comment.

Shel: Mm-hmm.

No. I mean, you have to be producing

the kind of content where you can have some success as an artist or an entertainer without having your content online.

Neville: Yeah.

Yeah, exactly. So the second commenter about signing up for every copyright protection that you can find is probably well, not probably it is a good idea, although I’m not sure that everyone would want to do that. And therein lies one of the issues about copyright. It depends on the jurisdiction. It’s a geographically based protection. Creative Commons is a good place to have established as a…

reserving your rights or some of your rights if you want to enable things to happen for others to use it. And that’s an international thing. So that’s a peace of mind, I would say. There hasn’t really been many, I’ve certainly not seen any legal court case tests since Adam Curry back in 2009 when he, I think it was in the Netherlands, he sued somebody who’d used a photo of his daughter.

and won the case. And it was not a Pyrrhic victory, but there was no you didn’t get any money out of it. But he got the legal ruling that these people had infringed on his copyright. But I’ve not seen any sense. So nevertheless, it’s worth doing. So yeah.

Shel: Yeah, it goes back to Susan Gosselin’s

comment, too, about any organization that does the same thing you do can take your content from your website, translate it into their language and publish it. And what do you do if your content is not copyrighted? There’s nothing that you can do.

Neville: Yeah, that’s incredible.

Reminds that reminds me shell back in the 2000s website scraping was it was a huge deal when when blogs suddenly came to the fore and you found that people were stealing all your content. I remember being in a lengthy email exchange with someone I think based in Romania or somewhere not a hope in hell. He was going to desist from doing that. Eventually it stopped though. So but I mean and that wasn’t from a copyright perspective. It was theft of content, which is well related, isn’t it?

So yeah, lots to learn from that. And finally, in FIR 510 on the 20th of April, we revisited the topic of shadow AI, the situation where employees ignore company approved AI tools and use their own preferred tools and not tell anyone. We discuss how one company approaches problem and how communicators might advocate for a version of this approach to aiding in AI adoption and speeding up productivity gains. And now you’re up to date on FIR episodes.

Shel: Also want to let you know about Circle of Fellows. We’ve had a fascinating discussion just this past Thursday on Circle of Fellows. It’s part one of a two-part discussion, and it’s all based on this. If you’re watching the video, you can see it’s a new book by Diane Chase, former chair of IABC and a great communicator called The Seven C’s of the New Communication Compass. And what Diane did here was…

outlined these seven points and find words that started with C to label them. And she wrote one chapter and then basically had IABC fellows write the rest of the chapter. So Diane is the first non-fellow to appear on Circle of Fellows, but it’s her book, so it made sense. And in the first installment that we recorded on Thursday, Diane was joined

Joining the panel were, of course, me — I moderated the session, Jane Mitchell, Ginger Holman, and Brad Whitworth. Next month on the May episode of Circle of Fellows, Brad Whitworth will be the moderator, and I’ll be a panelist talking about the chapter I wrote about community, and I’ll be joined by Zora Artis and Cindy Schmieg, IABC fellows who wrote the other chapters.

It’s a really good book. I recommend it for communicators. But we talk about some of the issues around these chapters and Diane explains why these were the topics that she said in the new communication environment. These are, you know, your North Stars, as it were. So definitely worth giving a listen to this month’s Circle of Fellows, which you will find on the FIR Podcast Network.

at firpodcastnetwork.com. And we are going to take a short break now for a sponsor message. We will be back to dive into our six topics right after this.

Neville: Here’s a story that on the surface looks just like another example of AI going wrong. In mid April, one of the world’s most prestigious law firms, Sullivan and Cromwell, had to apologize to a US bankruptcy court after submitting a filing that contained multiple AI hallucinations, fabricated case citations, misquoted legal authorities, even references to cases that simply don’t exist. The errors weren’t minor.

They were significant enough that the firm had to send a formal letter to the judge, acknowledge what had happened, and submit a corrected version of the filing. And just to make it more uncomfortable, these mistakes weren’t caught internally. They were identified by the opposing legal team. Now, if you stop there, it’s easy to frame this as just another cautionary tale about AI, unreliable tools, hallucinations, the risk of automation in high stakes work. But that’s not the story here.

This firm didn’t lack guidance, quite the opposite in fact. They have formal policies governing the use of AI. They require lawyers to complete training before they can even access these tools. Their internal guidance explicitly warns about hallucinations and tells lawyers to verify everything before it goes anywhere near a client or a court. In fact, their own language is very clear, trust nothing and verify everything. And yet, in this case, those policies were not followed.

A document that should have been scrutinized at multiple levels made its way into a courtroom with fundamental inaccuracies baked into it. The failure here wasn’t the technology. It was a failure of process, behavior, and accountability. Human in the Loop only works if there is an actual human who is clearly responsible for checking the work, not in theory, nor in a policy document, but in practice, at the point where a decision is made to send something out into the world.

And what this case suggests is that in many organizations, that loop is more notional than real. If AI is being used to accelerate work, where are the safeguards that ensure quality isn’t being compromised in the process? And are those safeguards actually being followed or just assumed? Having a policy is one thing, embedding it into how people actually behave, especially under time pressure, is something else entirely. And I think that’s where this story really matters beyond the legal profession.

Behavior is moving faster than governance. People are experimenting, they’re finding shortcuts, they’re integrating these tools into their daily workflows, often quietly and informally. The risk isn’t only that AI gets something wrong, it’s also that humans stop checking as rigorously as they should, or assume that someone else had, or trust output that feels authoritative, even when it hasn’t been properly verified. So when we talk about responsible AI or human-centered AI or governance frameworks,

This is what it comes down to in practice, not whether you have a policy, but whether at the moment that matters, someone takes responsibility for asking a very simple question, is this actually correct? And if this case tells us anything, it’s that answering that question consistently is still much harder than many organizations seem to think.

Shel: Boy, isn’t that true. And the first thing I thought when I read this story, because it seems like the organization did everything right, is the question about what are people rewarded for in this organization? I wrote a post about this on LinkedIn probably a couple of months ago, that process speaks louder than any message that you send.

through communication channels. And I included this story that I’ve probably relayed on this podcast 20 times just because it is such a good analogy to really make this clear. The company, the logistics company that was experiencing a lot of breakage of its packages in one of its distribution centers and…

the company just kept sending messages about how important it was to be careful and take the time when you’re loading these packages so that you’re not just throwing them around and you’re not breaking things that customers are expecting to receive unbroken. And the breakage continued and they brought a consultant in, communications consultant I might add, who looked at all of this and found that the reason this was happening was because people were actually being rewarded for productivity.

and not for quality. That meant that they were getting money for doing this quickly. So as long as you were going to pay them more to get this stuff out quickly, the breakage was going to continue. You had to shift the rewards mechanism. So it was rewarding quality. Now they’re going to slow down and make sure everything’s unbroken. Of course, you’re going to lose some of the speed there.

So when I hear the story about this law firm, the first thing I wonder is, yeah, we have all of these policies and we’ve been through all of this training and the governance is in place, but I’m being rewarded for getting this done quickly. And therefore I’m not going to take the time to review the citations that the AI cranked out. I don’t know that this is the case in this organization, but it was the first question I asked.

Neville: Well, that’s an interesting point, Shel, because that was in my mind, too, but it didn’t make it into my notes that they’re known to be, they are, I think, the second biggest law firm globally. They’ve been around 150 years, long and well established, highly credible, super reputation, all that. But they have some of their lawyers charging out at around two and a half thousand dollars an hour for their services. That’s serious money.

And that kind of adds to your point about speed is the focus here, not the quality. Now, to repeat what you said, we do not know if that’s the case in this law firm. But it could be, and your other point you make about that an individual might be saying, I don’t have time to check all this stuff because I’m being rewarded for getting stuff out fast. They have to address that if that’s true. They really have to address that because that perpetuates this. If it turns out that it’s true.

But it brings, I suppose, to my mind, the reality, all these policies, and there’s a lot of reporting on this that’s around if you look for it, talking about some of the training courses, the fact that a lawyer, no lawyer can even use one of their AI tools unless they are certified to have done this, this and this training program or that video they have watched, all that. And yet this happened. So there’s something out of loop here. There’s something not

not working properly. Could it be as simple as the person who signs off on this, i.e., this piece of work is going to that client or this filing is being submitted in this bankruptcy case in New York? This was a major bankruptcy case by not individual. I believe it was a financial company in the Virgin Islands, the British Virgin Islands for that matter, in the Caribbean. It was high profile. But could it really be as simple as that? All this

work going on at speed. that’s probably somebody thought that’s fine, because we’re going to check it all. And yet nobody did. So it signals something we’ve encountered before. And I’m reminded of a case we reported last year about Deloitte. And the issues they had was something similar, but it was not a legal case in a court. It was a report they prepared for a client, which happened to be the government of Australia, and another one in the government of Canada have six figure fees.

riddled with hallucinations and other things. So somebody didn’t check it in that case. I have no knowledge of what training they have in place. In this case, we do have knowledge of what training they have in case. Could it really be as simple as somebody, an individual, isn’t the known responsible partner in the law firm for the authoritative voice on this is okay to send to that client or to that court or whatever.

Even if 15 other people have been involved in checking stuff before, that one person has that responsibility. They obviously don’t have that, suspect. Maybe that’s the solution to this kind of thing. you I know you have some strong views on, you know, having a verifier in place in organizations. You want to talk a bit more about that?

Shel: Well, yeah,

I mean, I’ve said this before. I think that one of the jobs that AI is actually going to lead to the creation of is a verification specialist, somebody who is accountable and knows they’re accountable. They are baked into the process. It gets passed to them and they verify the entire document. I don’t care if it’s an 80 page filing. You know, there was another law firm.

that found itself in this trouble recently. It was in Oregon and the court of appeals there sanctioned the lawyer involved for the AI errors that were in the law firms filing. And the court in its finding…

emphasize that AI isn’t a lawyer and it can’t replace professional judgment or accountability. And that principle travels pretty well. AI is not a communicator. It’s not a strategist. It’s not a lawyer. It’s not an HR expert. It’s not a subject matter expert. It’s a tool. The professional has to be accountable. So for communicators, that means we can’t outsource accuracy. We can’t outsource context. We can’t outsource tone or ethical judgment to a machine.

We can use AI as aggressively as we can find that it helps us do our job, but we still have to verify ruthlessly and we have to make sure that other people in the organization know that that’s part of their remit too.

Neville: Yeah, it’s a very tricky one, think, Shel, given what we know currently about the developments that are happening with artificial intelligence, particularly in generative AI, particularly with tools, particularly like Claude and ChatGPT and Gemini, where and Steve Lubetkin in his comment to one of the episodes from last month alluded to that when he talked about how it is great to, know, deciphering columns in Excel spreadsheets.

Here you’ve got a tool that can actually generate the AI spreadsheet and perform literally everything in analytics, pivot tables that works literally in 20 seconds. And so you suddenly find that here you have a tool that is able to generate content that the traditional way of prompting would have taken considerable to and fro.

And, you know, changes here, editing there, you turn the AI, not this, that’s coming back to all that sort of stuff. And then someone checking it. And here you’ve got the situation where this is accelerating, it can do these things, arguably, and again, depending what it is, it can do things that until now, people would say AI can’t do that. And I’m thinking that, you know, AI is not a communicator. No, it’s itself is not at the moment.

So I mean, this will take us down a rabbit hole if we get into this, which we’re not going to do. But it’s a point worth noting that sooner or later, we’re going to have an AI tool of some type doing something only before a human could do. And then where are we? So again, that’s all a bit in the future and maybe sooner than we think. I don’t worry about that in the sense because there’s no point, Shel. It’s not happened yet. But I do worry about things like this because

This is an easy one to get right, it seems to me, that you got all these policies, et cetera, you got to not so much enforce them. That’s not really the right word to use. It’s to ensure that people follow those policies. Therefore, it’s a communication issue. It’s an educational issue. It’s not a training issue, but it’s education, awareness raising, and getting people to buy into why they should do this, in which case, you’re likely going to have to change your model of rewarding people in that case.

That’s big deal. So this isn’t something that you can idly do except on the kind of surface, i.e., you do all this stuff, you’ve got one person who’s got the responsibility and the consequences will fall on that person if it turns out no one followed the stuff. So that’s probably what would help here.

Shel: Yeah, and I think it’s also worth noting that it’s going to get easier to assume that AI got it right. I mentioned that AI currently isn’t a subject matter expert, but it’s becoming one that we have. OpenAI is creating one that’s just for doctors and Anthropic just signed a deal with a law firm to create a legal specific version of Claude. So, you know, I think when you

look at what happened here with this law firm. We should look at this as sort of a dress rehearsal for AI related crisis response. The law firm did the right thing, right? They acknowledged the problem, they apologized to the court, they filed a corrected version, but at that point, the reputational damage had already been done because that narrative…

had found its way into Reuters, The Guardian, Business Insider, Above the Law, LinkedIn, and all the legal newsletters. And that’s how AI failures will unfold for other organizations, whether it’s out of the legal department or elsewhere. You’re going to have the operational error, then the public narrative, and then people are going to pile on. Communicators should already have holding statements, internal FAQs, and escalation protocols for AI-generated errors, especially

In high stakes content like a legal file.

Neville: Yeah, plenty to think about on this. although the kind of advice I would give is, yes, you’ve got all your policies and so forth, as we’ve been discussing at the beginning, but have you got the human genuinely in the loop to take responsibility for what you’re giving to a client or to a court?

Shel: Well, let’s stick with the AI theme. Hey, that should be no surprise. Gartner is predicting that by 2028, 75 % of employees will rely on chatbots to get relevant internal communications. That’s not the distant future, folks. It’s the year after next, and that should stop every internal communicator in their tracks. Not because chatbots are coming for the intranet or the newsletter or the manager cascade. That’s just

Too simplistic. The bigger shift is that employees are moving from browsing to asking. They’re not going to hunt through the intranet and a stack of emails to get an answer to a simple question. They’re just going to go to the chat bot and ask, what has this changed for me? Do I need to do anything by Friday? Why is my department being reorganized? And they’ll expect an answer in seconds and probably get one. The Gartner prediction is based on a very real problem.

information overload. According to Gartner’s report, employees who report high information overload are 52 % less likely to report high intent to stay with their organization, so it’s a retention issue, and they’re 30 % less likely to report high strategic alignment with the organization. Gartner also says chatbots will provide personalized curated answers for pull communication and customized alerts for push communication.

That’s a major shift in the employee communication model. Now there are real benefits here. A well-designed internal chat bot can give employees faster answers, reduce HR and IT ticket volume, provide 24-7 support, support multiple languages, and cite authoritative sources so employees know where an answer came from. It can also deliver information within the flow of work rather than forcing people to go somewhere else to find it.

But here’s the part communicators are going to need to wrestle with. An AI answer is not the same thing as communication. An answer can tell an employee what

An answer can tell an employee what changed. It may even summarize why it changed. Will it preserve the intent, the nuance, the context, and the emotional intelligence of the original communication? There’s no guarantee it will. Take changed communication, for instance. We frequently write detailed articles explaining the rationale for a change because employees need more than the transaction. They need to understand the business context. They need to know what

options leaders considered and which options they discarded and why. They need to hear what’s not changing. They need some sense that the decision was made thoughtfully and not arbitrarily. But what happens when no one reads the article? What happens when the employee asks the chat bot, what’s changing in our benefits plan and gets a clean, accurate three sentence answer that strips out the rationale completely?

This is where internal communicators have to evolve from being message producers to knowledge architects. The intranet still matters. It may be less of a destination and more of the trusted knowledge later that feeds AI. Frank Wolf made this point really well in PR Daily. AI doesn’t eliminate the intranet’s jobs. It changes how pull, push and people centered communication work.

The intranet becomes the foundation that makes chatbot answers reliable. If the knowledge layer is messy, if it’s outdated or written in a way that AI can’t interpret or interpret well, the chatbot’s going to sound confident, still be wrong. This means we have to consider an expansion of the internal communicator’s job. Yeah, we still need to write, but now we also need to structure. We need clear source of truth pages and metadata.

We need FAQs that anticipate employee questions, and we need version control, expiration dates, and more. We need to decide which information can be answered directly by a bot and which question should trigger a human response. And we need to design for narrative preservation. That means writing source content with AI retrieval in mind. If the rationale for a change matters, don’t bury it in paragraph eight. Make it explicit. Label it.

Repeat it in a concise why this matters section. Smart brevity writing would be a great approach to adopt here. Create approved answer blocks that the chat bot can draw from and test the bot by asking questions employees are likely to ask, then check whether the answers reflect not just the facts, but the intended meaning. This also has implications for measurement, by the way. Page views and open rates become less useful

if employees are getting answers without opening an article. We’ll need to measure the questions employees ask, the quality of the answers they receive, the content gaps the bot reveals, and whether employees understand the strategy, the change, or the policy after interacting with the system. It’s a lazy conclusion to say employees won’t read anymore, so let’s just give them chat bots. The better conclusion is employees are changing how they access information

So we need to make sure the organization’s knowledge, context, and narrative survive that shift.

Neville: Hmm. Yeah, this is a huge topic, Shell, because what struck me listening to you was in a sense of continuity of what we just talked about in the previous topic is the verification of content that an AI produces for you. How are we going to deal with that? We talk about putting in place, you know, trusted sources for all this information. So, you know, let’s say I’m an employee, I’ve asked a question on something, it’s given me an answer.

I need to check that. So what do I check that? And how do I know that it’s accurate? So you project that out to the kinds of stuff people deal with daily. And this is a huge undertaking, I would say, because looking at that article, talking about this, it has an interesting piece in there about the safeguards that CCOs are mentioning here specifically.

will need to put in place to mitigate the risks of hallucinations, misinformation, and a fragmented landscape that comes with AI, they say. CCOs will need a greater emphasis on information quality, as well as an optimizing intranet content for AI searchability. You mentioned that point. They must also partner with IT, HR, and legal to establish robust governance to ensure that chatbots responses are accurate. That’s the bit. How are they going to do that? Because something internal,

It surely isn’t going to be only producing answers based on what it finds on your internal networks alone. It must be looking out onto the wider landscape. How do you verify and check all that? That’s a major debating point for taking this further, it seems to me. So it’s a huge undertaking.

Shel: Yeah, think one of the things, and I sort of breezed through it pretty quickly, but I think that we’re going to need to figure out is how do we monitor and assess what questions employees are asking that produce an answer that’s drawn from internal communications content, whether that’s in an email that went out or something that was posted to the intranet. How do we monitor the questions that are being asked and the effectiveness of the responses so that we can make adjustments?

So that we can report that, yeah, we can determine that there is alignment on why this change was made, or we can say, gee, people are just getting an answer that tells them what the change is, and they don’t have any understanding that we looked at alternatives and we tried to find a better solution. And this was the best we settled on. And here’s why it’s good for employees or here’s how to cope with this in your department or whatever it may be. And to do this without

necessarily surveilling employees, right? We don’t want to know who asked the question. I think it would be great if we could say, wow, look at this. 70 % of this particular point of confusion that was illuminated by the questions that we’re asking are coming from people in our operations division and not other divisions of the organization. That would be useful, but we don’t want to be able to say John Doe asked this question. What an idiot.

It’s a serious issue. And I think the guidance that we need to have the information in multiple places where the AI can see it so that it realizes that this is an important topic because it is in several places and that we have it in several formats, the FAQs, the answer blocks. This is repurposing the original content in ways that will help ensure

that the AI inside your organization is delivering information with context, with those other elements that’s so important for employees to understand to create that alignment. And by the way, I mentioned the seven Cs of the new communication compass. One of them is congruence, which we’re arguing goes beyond alignment, that there is congruence in the organization. So.

If we want that, and it is important, it’s one of the reasons there is an internal communications function, we really need to start rethinking what we’re doing and how we’re doing it.

Neville: Yeah, agree. I think surveillance is a very, very slippery topic and a slope. Because you’re going to have to have some kind of process in place and it probably surveillance is the correct label. Otherwise, you have really struggled to to find the answers you will need when you if you roll out something like this. So I think

You know, we’ve reported recently on keystroke logging and other ways organizations are now requiring in place to monitor whether employees are working or not. And it’s still making news headlines in the tabloids here, a case recently about someone had this wheeze of having something that touched his keyboard every now and again to show it was working. The trouble is that the employer was savvy enough to have the software could tell which key and it was the same key all the time.

So things like that, probably going to have to rebalance this privacy versus being able to see what people are doing algorithm, let’s say. And that’s going to be difficult, given the history, I suppose, of some organizations not respecting employee privacy. Look at the China model, and that’s not what we want to have here.

state surveillance on everyone’s daily lives is pervasive in urban areas, not necessarily throughout the whole country. So do we want that? We may not actually have the ability to say no to that, given what they need to do. So that’s part of the issue to include, I think.

Shel: Yeah, and I think one other thing we’re going to have to do is more asking. We’re going to have to survey after a change and ask employees if they understood the reason for the change. And part of the problem with increasing the number of surveys, and I’ve made this argument for years, that people will take surveys all day long if they see the results of the surveys and they see that things are going to change.

If you’re asking people, did you understand this? Did you understand the rationale for it? Do you agree with it? It’s hard. I mean, you can report the results, but what’s going to change? You’re going to change maybe the way you’re producing content. That’s not going to be visible to employees. So it’s going to be a challenge to ask those questions frequently without producing that kind of survey fatigue that we hear so much about.

Neville: Big topic. OK. OK, so there’s a widely held idea about AI that’s been around almost since the beginning, that it would be a great leveler in the workplace. It’s kind of continuity of what we’re talking about here, The thinking was that if you give everyone access to powerful tools that can write, analyze, summarize code, and generate ideas, then people with less experience or fewer formal skills

should be able to close the gap with those at the top. But what we’re starting to see in the real world looks quite different. In fact, it may be doing the opposite. The Financial Times has just published new research based on a survey of 4,000 workers in the US and the UK. And the findings are pretty stark. More than 60%, a 6-0, of higher earners say they use AI every day in their work. Among low earners, that number drops to just 16%, 1-6. That’s a pretty big gap.

So instead of leveling the playing field, AI adoption is heavily skewed towards the people who are already ahead, better paid, more experienced, often in more knowledge intensive roles. I think it makes sense because using AI effectively isn’t just about having access. Most people have access. It’s about knowing what to do with the tools. It’s about having the confidence to experiment, the context to apply them to real work, and the judgment to assess whether the output is actually useful.

And those are things that tend to come with experience, with education and with the kind of roles where you have a bit more autonomy over how you work. There’s a line in the research from one economist that really captures this shift. The more intelligent the technology becomes, the more your own intelligence matters. If you already have expertise, AI can make you faster, more productive, maybe even better at what you do. But if you don’t yet have that foundation, it’s much harder to extract real value from it.

There are other factors at play too. The research points to corporate training as one of the biggest drivers of AI use at work. So organizations that actively support and encourage adoption are seeing much higher uptake. And interestingly, the heaviest users of AI aren’t the youngest workers, as you might expect, but people in their 30s with more experience behind them. So again, this isn’t a generational story so much. It’s about how AI fits into the structure of work itself.

If AI is boosting the productivity of higher earners more than lower earners, then over time you’d expect the gap to widen in output, in value, and potentially in pay. And there’s a second order effect that’s a bit more subtle, but potentially more significant. If AI starts to take on some of the routine or entry level tasks that junior staff would traditionally do, then where do people build the skills? How do you develop expertise if the work that teaches you the fundamentals is increasingly being handled by a machine?

So instead of AI acting as a ladder, help people climb, there’s a risk it starts to pull away some of the rungs. And this is where it connects directly to leadership and to communication. This isn’t just about who has access to AI tools. It’s about who feels able to use them, who is encouraged to use them, who is trained to use them well, and who is supported in making sense of what they produce. So this is about culture, not technology.

If organizations simply roll out AI and assume the benefits will spread evenly, they may find the opposite happens, that they’ve unintentionally widened the gap inside their own workforce. So perhaps the real question here isn’t whether AI will level the playing field. It’s whether leaders and communicators advising them are actively shaping how that playing field is changing or just watching it tilt.

Shel: Yeah, that training point is, I think, really critical. And a project manager, an accountant, field supervisor, an HR business partner, and a communication specialist don’t need the same training. They also don’t need the same examples delivered by communications. So the communicator’s contribution here is translation. Here’s what this means for your role, for your task, for your team, and your day.

That includes, you know, surfacing success stories from unexpected parts of the organization. I would love to find an example of, a foreman on a construction project site using AI. I don’t want to just report on what the IT department is doing and the other, you know, tech forward departments are doing. You know, the goal shouldn’t be everyone becomes an AI expert. The goal should be that nobody’s quietly excluded from

the next operating model because they don’t see how AI fits in their work.

Neville: Yeah, we’ve talked about this multiple times, Shell, in various episodes, which is who feels able to use such tools. And that comes, that stems from the leadership communication, in my opinion, that has to encourage people to do this. They feel they’re being empowered. They feel they’ve been given permission to do this. And they know they can count on help to when they get stuck with something. That, is

hardly uniform in just any organization, frankly, and this isn’t about, we’re going to create the special department to do this, this needs to permeate across an organization. So you’ve got leadership at the very highest level, filter that down to your local manager, your line manager, or whoever you report to has got to encourage you as well. And I’m sure that’s that happens in many organizations, but to make this really work, that doesn’t result in the gap widening between those who are

naturally excited about this and have the experience and the knowledge and the expertise to know how to get value out of these tools. You’ve got to have something in place that helps everyone else who isn’t like that. And there’s a challenge for communicators without any question.

Shel: Well, for the whole organization, I mean, as I talk to people in other companies, it seems that we’re still in that experimentation phase that I think most organizations should be beyond by now. But the way it’s working right now in a lot of companies is the curious employees can try the tools and cautious employees can wait and everyone else will eventually catch up. that’s not going to work. mean, if this is becoming a material productivity and capability layer,

Neville: As well.

the

Shel: We need to implement intentional adoption strategies. That means making role specific examples and approved tools and safe use guidance and peer demonstrations. And we’re trying to do this where I work is get peers showing other employees what they’re doing. Psychological safety, plain language explanations of what employees are supposed to be doing with this. So all needs to be put in place and communications has a role to play here, but

If we don’t, adoption is going to follow the path of least resistance, and that’s toward the people who have the power, the time, and the digital fluency, and then you’re going to end up with that gap.

Neville: Work to do here.

Shel: Yeah, and by the way, there was another part of the FT’s reporting that I found really interesting that you didn’t mention. And that’s that men are more likely to use AI tools than women across a number of sectors. And I think that should concern leaders because AI fluency is becoming part of professional competence. And if men, along with higher earners and more experienced workers, are building fluency faster,

What’s going to happen? And you know, the performance evaluations and promotion decisions and the visibility of the employees who are getting that kind of attention and informal influence may start reflecting AI access rather than raw ability. And here again, there’s a role for communicators by pushing AI enablement into say, manager toolkits, into your onboarding processes, into your training and team level norms is important.

as opposed to just letting it sit as an informal advantage for people who are already competent.

Neville: Yeah, like I said, work to do here.

Shel: Thanks, Dan. I am looking forward to seeing this WordPress release. I have to say, I really like the idea of collaborative editing. As you know, the FIR podcast network website is on WordPress and Neville, you and I both use it and the ability for both of us to go in and deal with that in more of a Google Docs setting than logging in and just pulling up the post.

makes sense to me. I definitely do see the issues with this as well, though, but it’ll be interesting to see this and the other changes as well. So thanks for the report, Dan. Really, really interesting. Well, we’re to stick with the AI theme again, probably not surprising given the impact that it’s having. And by the way, I have to say that when I scroll through LinkedIn, it’s got to be 80 % of the posts I see now are

AI related and that’s not hyperbole. It is a guess. I haven’t measured, but man, it is all AI all the time on LinkedIn. That’s what people are talking about. And it’s changing the relationship between PR professionals and journalists, just not the way a lot of people expected it would. The fear was that AI would automate the work. We’d have a lot of AI written press releases and AI written pitches and articles.

And yeah, there’s definitely a lot of that happening and people are calling it out. But the more interesting shift is not that AI makes it easier to produce more content, it’s that AI makes bad media relations more obvious and more damaging. Pete Pachal, who was a guest on FIR interviews, what was that Neville, about a year ago? Yeah, he makes this point in an article in Fast Company, AI is becoming a new interface.

Neville: A year ago,

Shel: For how information is found, prioritized, and interpreted. Journalists and PR people are both affected because AI systems more and more shape which stories surface, which ones get cited, and which narratives get visibility. Pete’s argument is that the advantage doesn’t go to the people who can generate the most material. It goes to the people who produce original reporting, useful expertise, clear narratives, and trusted relationships. That’s an important distinction for people who…

operate in the media relations world. AI can help you write faster, but speed was already part of the problem. Journalists were already drowning in irrelevant pitches before generative AI showed up. AI just gives every mediocre PR practitioner a way to send even more mediocre pitches even faster. The results not greater efficiency, it’s more noise. And journalists are noticing.

PR Daily reported on a global results communication survey of nearly 1,700 reporters across print, digital, and broadcast. 81 % said pitches and relationships with PR professionals are vital to their work. So journalists aren’t saying we don’t need PR, but 43 % expressed negative views about AI-generated pitches, saying they read like a bot wrote them and that they lack perspective and erode editorial trust.

So here’s the conflict. Journalists still need PR. They need access and sources and data and context and story ideas, but they’re getting a lot less tolerant of anything that feels mass produced, poorly targeted or synthetic. Medianet’s 2026 Media Landscape Report, based on feedback from 800 journalists, makes the same point more sharply. The report says three quarters of journalists have received pitches that appeared to be AI generated.

and about half said they could always detect machine written copy. I would argue with that, but let’s not go down that rabbit hole. The same report says 86 % of journalists now cite press releases as a key news source, which means the press release isn’t dead, but the stakes for credibility are higher. There’s also a widely circulated LinkedIn post citing the media net research saying 78 % of journalists report that receiving an AI written pitch

decreases their trust in the PR person who sent it. That’s consistent with the other findings. Journalists aren’t rejecting AI assistance, they’re rejecting lazy use of AI. So what should PR practitioners be doing differently? I’ve got five things. First, stop using AI as a pitch factory. This is the most obvious trap. If the output is a generic email with a personalized opening line,

and a weak story angle, AI hasn’t made you better, it’s made you faster at getting ignored. Second, use AI before the pitch, not as a replacement for your judgment. Use it to analyze everything the journalist has written recently, summarize themes, identify gaps, pressure test whether the angle is timely, and prepare sharper source material.

P.R. Daley’s piece makes this point well. AI can help with research, angle testing, drafting, editing, personalization, and follow-up prep, but the human edit is where you add that credibility. Third, bring journalists something they can’t get from a model. That means original data, direct access to informed sources, a useful articulate expert, a local angle, a contrarian but defensible point of view, or a story that fits the reporter’s audience.

Fourth, be transparent internally about what AI can and can’t do. PR leaders should have rules. AI can help research, structure, brainstorm, and edit, but it should not invent relevance, fake familiarity, fabricate personalization, or send anything without human review. And fifth, think beyond the pitch. In an AI-mediated media environment, you’re not just trying to get a reporter to open an email.

you should be trying to build a public record of expertise and credibility. That includes owned content, executive visibility, contributed thinking, data assets, analyst material, podcasts, newsletters, earned media, anything that reinforces a coherent narrative that AI systems will recognize and retrieve. So the future of media relations isn’t more automated pitching.

The future is more precise, more evidence-based, more relationship-driven, and more strategic. AI will handle more of the mechanics, but judgment, relevance, trust, and access become more valuable, not less. In other words, AI doesn’t eliminate the relationship between PR and journalism. It raises the penalty for abusing it.

Neville: Yeah, it’s a it’s it’s an interesting topic without doubt. I was actually pretty impressed with the five points mentioned by Courtney Blackan in the PR Daily report. And it mirrors, frankly, almost everything we’ve talked about in this episode so far. And indeed, in recent episodes that we have to keep repeating this really, Shel, and you’ve done a good job, I think, at outlining this is what you’ve got to do.

And it’s about this, but it also relates to these other 10 things we’ve talked about. But a couple of things that struck me here that really, really do resonate well. mean, research smarter, that makes complete sense. I mean, that’s got to be your starting point. But things like draft faster and edit harder. I like that one, I must admit. So you use AI to an AI tool to organize your ideas.

into a structured draft or just simply improve the overall language that you’ve done and rewrite some of it. To kind of anticipate criticism from those who don’t think I should be involved in any of this, I liken it to that’s what you’d be asking a colleague to do or that freelancer that you’ve hired to help you work on this. You’d be giving them the same request as you would to the AI tool.

So what’s the difference? One’s not a human. That’s probably the biggest difference. But I don’t get swayed by any of those arguments about you can’t use AI to do this. Now, of course, you’ve got to use it. The caveat is, for God’s sake, don’t just copy and paste that into your document in the center. Now, this is your assistant, not your creator. You’re the creator. And this helps you create very well, typically, all other things being equal. But I like that draft faster, edit harder.

And it’s kind of like your A-B testing or A-B-C testing possibly with the AI assistant to help you do this quite quickly. And it’s great. Personalize with precision is another one she mentions. Don’t blast out the same email to 50 journalists, which is what many people do it seems to me. You’ve got the ability to personalize those emails. And again, you know,

Your AI can help you with drafting that. So it will need to know quite a bit about the journalists and your relationship with them if you’re the PR person. So quite a lot of prep work you’d need to do here first. But the output from the AI will be pretty good if you do it right. So these are things that take your method that you might currently be using, which is prompt the AI to a totally different level.

And that’s what you got to be thinking about now because this is where it’s all going. This is not way beyond just simply a chat bot. So it’s a really good topic. And these reports that you’ve highlighted, Shel, are great. Pete Pachal’s post is excellent. We’ve got to him back for back again from another interview, I think, because we interviewed him when he was just starting his business. And he’s gone places for that business now. So it’s worth reading.

Shel: Yeah, I think so.

Neville: That and the PR Daily report. do like those five points.

Shel: Yeah, remember, remember we interviewed Aaron Kwittken from PRophet, that’s PRophet with a capital PR. And one of the things that’s, yeah, it was. And one of the things that system did was identify reporters who have written about this topic. It reviews the content that they have written over the past recent period and crafts a personalized pitch for each of them, which then you can go in and edit. I don’t think you

Neville: Yeah, I do. Yeah.

Quite a while ago.

Shel: Sorry, Aaron, if you’re listening, I think you provide a great service, but people don’t need this anymore because you can create an agent that does that, identify the reporters who have written about this, review their most recent articles and craft a pitch for this press release. That can be done now internally with an agent that would probably take about an hour to create. mean, agents can go out and do amazing things now. Chris Penn.

just wrote a post, he found somebody’s wallet on the street and it had enough stuff in it that he could spend a few hours tracking down who owned it. There wasn’t a driver’s license with an address. There was some cash, there were a couple of debit cards, but he was able to give an agent all of that information and go do his work on something else. And after a couple of hours, it said, I’ve narrowed it down to these three people. And Chris was able to look at those three people and figure out which one it was.

lived really nearby and got the guy’s wallet back. We can do this kind of thing now in pursuit of PR objectives. The other thing I want to say is that I’ve gotten in the habit now of recording interviews and giving the transcript to AI and say, organize this into a first draft of a press release of an article of a change notice, whatever it might be. And I don’t copy and paste that in. That’s a first draft. It’s absolutely.

a case of draft quickly and edit hard. I hadn’t heard that framing before, but it’s absolutely what I do these days because it just saves a lot of time and gets me into the nuts and bolts of making this relevant without having to spend half that time just reviewing the transcript and organizing that into sort of a logical flow.

I think it’s a great use of AI and it’s one that I’ve been using for, geez, a couple of years now.

Neville: Is. I agree. So there are things that you’re accustomed to that work for you. But pay attention to this kind of thing, because this is taking it to another level that will benefit you. You just will not just but you need to clearly understand what this is. And Pete’s article, the PR Daily piece are two sources that will help you do that. But it’s definitely worth a look.

Our next story is a very different one. It’s not about something going wrong. It’s not about AI, but about something going exactly to plan. Apple announced that Tim Cook will step down as CEO later this year and become executive chairman with John Ternus, currently head of hardware engineering, taking over the role. On the face of it, this is a major moment. A CEO transition at a company of this scale. It’s what, revenues?

trillion, second wealth, the most valuable company in the world currently. It was number one not long ago, so it might retain that. It often creates uncertainty internally in the markets and across the wider ecosystem. What’s striking here is how little disruption there seems to be. There were no leaks. The announcement landed cleanly. The market reaction was muted and the tone throughout is calm, controlled and focused on continuity.

And that’s the real story, I think, because this isn’t just a leadership transition. It’s a masterclass on how to communicate one. If you look at the messaging, everything reinforces stability. isn’t disappearing. He’s staying involved as executive chairman. Ternus isn’t positioned as a bold new direction. He’s presented as a long standing insider, deeply embedded in Apple’s culture and products. There’s no sense of rupture, just a steady handoff.

The most important part of this story though is the announcement itself. It’s what happened before the announcement. Ternus didn’t appear overnight. He’s been gradually made visible over several years, fronting product launches, appearing in keynotes, becoming a familiar presence. So by the time this announcement arrives, it doesn’t feel like a surprise. It feels like confirmation. And that’s the key insight. This transition didn’t start with a press release. It started years ago.

What Apple has done is build familiarity, credibility and trust in the successor long before the moment of change. So when the change comes, the narrative is already understood. And that changes everything because most organizations treat moments like this as announcements, whereas Apple treats them as outcomes, the result of a story that has been deliberately shaped over time. That has practical implications because when transitions feel chaotic or disruptive, it’s often not because the change itself is unexpected.

is because the story hasn’t been prepared. The successor isn’t known, the narrative isn’t clear, the organization is reacting in real time. Apple avoids that entirely, not by communicating more in the moment, but by communicating earlier, by building trust before it’s needed. And that’s where this becomes relevant for leaders and for communicators advising them. The real question isn’t how do you announce a change, it’s how early you start preparing people to understand it.

Shel: Yeah, they didn’t treat this as a sudden disclosure. This was more continuity without pretending that nothing was changing. Right. It does a lot of reassuring work, not just about Cook’s remaining in the new roll through and actually staying in the current role through the summer and then staying as executive chairman, who’s going to work closely with Ternus during the transition. Also talked about

Ternus’s ties to Steve Jobs and Apple’s mission and its values. And that language isn’t an accident. I think the lesson for communicators is that the leadership transition needs facts and emotional reassurance, right? Employees don’t just wanna know who reports to whom. They wanna know whether the company they believe in is still the company they believe in. I do like in PR Daily’s report, the discussion of different audiences.

They didn’t send one announcement everywhere. They had public messaging and employee facing messaging and they both serve different purposes, right? The public version celebrated the legacy and confidence they had in this transition. The employee version was warmer. It was more grounded. And I mean, this is communication 101 in a lot of senses, but still something that we should emphasize. Consistency doesn’t mean identical language. Employees…

deserve a message written for employees, not a copy of the press release with Dear Team pasted on top.

Neville: I agree with you, Shel. This is an excellent example of how to do that. And yes, there wasn’t a single message. That’s very true. It was tailored messaging that showed clear understanding of those different audiences internally and externally. So a lot you can learn from that. And indeed, Ragan’s article by Allison Carter has some good insight in there that you can glean learning from. It’s worth reading that article too.

So I call it a masterclass. It probably is one of the best examples I’ve seen. Not so much the press release, but understanding about what led up to that and all the other communication that then occurred, the buildup. And I realized too, of course, that some organizations won’t know until nearer the time of announcement that there’s going to be a change. So this isn’t necessarily a blueprint you can apply to everything. But in the case of Apple,

It’s a very, it has a big effect on people news about what Apple’s doing. Steve Jobs was a magnetic personality or mercurial personality who famously coined this great phrase, I’d apply to Trump, often the reality distortion field that was his trademark in a sense that he was mercurial without doubt in leading. one thing that is notable, although it’s certainly not emphasized in any way that

that Ternus is a hardware guy, whereas Cook is a management guy. So Cook took over from Steve Jobs and transitioned Apple over that decade and more period to where we are now with the changes going on in the world generally and the tech industry in particular, that it probably does require more of a hard-nosed technological approach than purely business management to leadership.

And of course, if Cook’s going to be the executive chairman, he’ll be there to assist here and there. interesting time looking at a company like Apple to see this happen.

Shel: Yeah, and the press release sends some messages without explicitly saying anything. First of all, the fact that they did pick a hardware hardware guy says a lot about where Apple is heading. They faced a lot of criticism for their failures around artificial intelligence, which isn’t even mentioned in the press release in that sense.

Neville: Yeah, does.

not mentioned.

Shel: Message. What have been Apple’s wins under Cook’s leadership? I mean, the Apple Watch was a big one, and a lot of people thought it wasn’t going to be. They kind of laughed when it was introduced. But there are a lot of people wearing Apple Watches out there now. Big success. But mainly he consolidated manufacturing in China, which may not end up being a great thing. But it’s he I mean, it’s made them a ton of money. What did he do? Triple?

their revenues, as you said, they’re the number two most valuable company in the world. Now they’re gonna refocus on hardware, on product, the stuff that has made Apple from the get-go. Software, mean, you can talk about iOS and the computer software platforms they produce, but you never hear a lot of…

Discussion of those at their big events. It’s, you know, we’re coming out with a watch. We’re coming out with a Vision Pro, which has been something of a failure. So this is a reemphasis on hardware. They’ve made that point. Casey Newton came right out on hard fork and said that Ternus’s first act should be just doing the deal with Google to integrate Gemini into Siri and be done with that whole thing because Siri was supposed to get that AI update.

It’s been a couple of years now and it just hasn’t happened. So they have said a lot in this press release and not all of it was necessarily explicit.

Neville: No, you’re absolutely right there. So interesting time in the tech industry generally and seeing where what happens with Apple in the coming year or so.

Shel: Well, my favorite new word is slopaganda, and this refers to AI-generated propaganda, that cheap, fast, emotionally loaded, and designed less to strategically persuade anybody about anything than it is to just flood the zone with images, memes, fake scenes, shareable outrage. The most visible example of slopaganda right now is Iran’s use of AI-generated, Lego-style videos aimed at Donald Trump, Israel, and the U.S.

They’re far from subtle. show caricatured Lego versions of Trump, Benjamin Netanyahu, missiles, burning ships, collapsing American power, and they use rap tracks, absurdist humor, conspiracy references, and the visual grammar of social media and not the language of state diplomacy. The New Yorker reported on Explosive Media, which is an Iranian digital media enterprise, which got started posting

pretty routine anti-Western content that didn’t get a lot of uptake. Then they discovered that these AI-generated Lego-style propaganda cartoons was its breakout format. The clips accumulated millions of views. They were reshared by Iranian government accounts. They were promoted by Russian state media and even picked up by anti-Trump protesters because the imagery was so flamboyantly anti-Trump.

The group told the New Yorker that it could produce a two-minute video in about 24 hours. LeMond adds an interesting scale point. According to Cyabra, a company that analyzes content to distinguish authentic activity from coordinated manipulation, that’s right off their website, according to them, pro-regime videos received more than 145 million views across X, Facebook, Instagram, and TikTok during the second half of March.

Explosive media eventually acknowledged to the BBC that the Iranian state was one of its clients. It had initially claimed that it was all independent. And this has captured a lot of attention, first because it’s visually disarming. Lego’s familiar, playful, global. It turns geopolitical violence into something that looks like entertainment. Analysts say the Lego format serves as a kind of Trojan horse.

reaching people who wouldn’t otherwise engage with war-related content. It also works because it’s emotionally true to people who’ve always wanted to believe the underlying message. Viewers may not literally believe Iran is winning the war in the way videos depict, but they can choose to believe the emotional premise that the U.S. is weak, Trump is ridiculous, and Iran is standing up to a global oppressor.

And it works because it speaks the language of the target audience. This isn’t old school propaganda. It’s fast, caustic, meme-literate, and platform-native. In information warfare terms, this gives Iran something it used to lack, cultural reach into Western audiences. It lets Iran fight asymmetrically using ridicule and narrative disruption where it can’t match the U.S. militarily. But this is not only a geopolitical story.

The same tactics are going to show up in business. Maybe not tomorrow in Lego form, but the pattern’s just too useful to stay confined to politics. An activist shareholder can use an AI-generated video to ridicule a CEO, to dramatize a company’s alleged mismanagement, or turn a dry governance dispute into a viral morality play.

A disgruntled customer could generate convincing scenes of product failure, employee misconduct or customer mistreatment. A labor dispute could be amplified with synthetic stories that blur the line between real worker grievances and invented incidents. An unscrupulous competitor could see just asking questions content that implies safety failures, financial instability, executive hypocrisy or environmental misconduct. An example from Canada.

matters here. The Canadian Digital Media Research Network identified a coordinated network of 20 inauthentic YouTube channels targeting Albertans with nearly 40 million views. The channels exploited real grievances and pushed narratives normalizing a move for secession and even U.S. annexation of the province. The report says the accounts pushed an Albertan perspective

that researchers found absolutely no evidence that the account owners were actually Albertan. That’s the bridge to business. Slopaganda doesn’t have to invent grievances. It can exploit real ones. A company with a safety incident, a layoff, a product recall, a labor dispute, or an unpopular executive decision is already vulnerable. AI just makes it easier for hostile actors to package that grievance

into emotionally potent, shareable content. So what should communicators do about this? Well, first, obviously, build monitoring capability for synthetic narratives, not just mentions. The risk isn’t one fake video. The risk is a pattern. Repeated themes, recycled scripts, coordinated accounts, sudden spikes, and emotionally consistent attacks. Second, prepare your verification protocols now.

If a video appears showing something damaging, who determines whether it’s real? Legal? Security? Coms? IT? Outside forensic consultants? You know, that first hour is really important. So knowing who to go to to find out whether this is real or not is really critical. Next, strengthen your owned record. If AI systems and social audiences are going to interpret your organization through fragments,

Make sure there’s a clear, accessible, credible body of truth. Your policies, your timelines, FAQs, source documents, leader statements, and plain English explanations. And finally, scenario plan for synthetic outrage. Not just misinformation, but ridicule. Means move differently than allegations. A dry correction rarely defeats a funny attack. Communicators need response options that are fast, human, factual, and proportionate.

And, you know, one last question to address here, should communicators use Slopaganda themselves? No, they shouldn’t. Not if we’re talking about deceptive, synthetic, emotionally manipulative content designed to obscure truth. That’s not communication, man. That’s reputational arson. But communicators absolutely should learn from the format. AI-generated creative can be used ethically.

if it’s clearly labeled truthful, brand safe, and grounded in real information. But understand attention has moved toward visual, fast, emotionally resonant storytelling, and we should move along with it.

Neville: Yeah, it’s an interesting topic, isn’t it, Shell? I think you’re kind of no communicator should not do this. That’s a message clearly the US government’s ignoring, judging by what they have been doing, or the White House, should say, that’s reflected back in what the Iranians are doing and their proxies and indeed in individuals by the thousands doing the same. So misinformation, disinformation, fakery, it’s everywhere.

I read a post about this at the end of March that looked deeply into what AI generated, how it’s being used by both sides. And there’s a number of reports, notably Deutsche Welle, the English language news service from the German broadcaster and France 24 as well, had some really, really good, well researched articles with examples of

what’s happening in this area. There’s a great one someone posted showing a Lego box of, you can visualize it from the description that we see on the TV news all the time. Residential buildings, apartment blocks in ruins blown to bits. And this is made out of Lego bricks. And it’s, you know, Lego logo, and it looks exactly like a Lego product. So brand

a brand is being, you know, brought into this unwittingly. But the reality is that communicators are in between a devil and deep blue sea here, I think, because if you’re in a business, you’re not in the defense industry, you’re not involved in anything with a war going on. Yet some of your clients are kind of on the fringes of all this by the nature of their business. So

they’re dragged into in the case of Lego, good example, what do do about it? Do you respond in kind with some kind of jokey thing about the, you know, this, you know, whatever it might be in Iran, this example. It’s a it’s, it’s a call, I would say, there seems to be a movement, if you like, to this kind of thing is a matter of normality.

And I think it’s very dangerous. Philip Boramantz had a really good piece in the middle of March on what this war is teaching us about communications generally, not specifically crisis, although it’s mentioned in there. The BBC had a report early March about AI-generated Iran war videos, surge of those as people have the tools to create these things. So that is part of our landscape.

So it’s something that communicators, it’s a question with no easy answer. The question you asked is not an easy answer, but it may be the one that we have to find an answer to. I mean, that’s easy to say that. I mean, I don’t know what that answer is. So it is that the most striking thing occurred to me is that the, not the sophistication of these tools are not slickly produced. They are produced.

I could you say it’s for those who are savvy with social media and social networks and what works in terms of spreading what is spreadable? What is memeable? And we are not part of that. And if you’re not, you’re people talking about your brand and you and you’re not there. So, you know, it’s a big question.

But it’s something we have to try and understand what’s happening and somehow come up with an answer.

Shel: Yeah, you raise an interesting point about brand safety. Is Lego going to issue a takedown notice to Explosive Media, which is a digital media company in Iran, probably not knowing that they’re not going to respect a takedown notice or there’s no court that you can go to necessarily. So basically you kind of have to live with it if you’re Lego. And I suspect that’s what they’re doing. But from a communication standpoint,

It’s really important to understand that the Iranian produced stuff is getting far more traction than the US produced stuff. And the reason for that is that it leverages grievances the Iranians already had and other people around the world already have with the US. The US stuff is just showing the attacks on Iran. And if you think about the average American or perhaps even the average Brit, what grievances do they have against Iran?

I mean, the grievances here are within the government, not within the broad population. And that’s why these are so effective, is that the Iranians and other populations around the world do have grievances, justified or not. So as you look at this stuff moving into the business world, consider what kind of grievances people might have with your organization. That’s where they’re going to attack you. That’s where you have to build up your defenses now before they do.

Neville: Yeah. So it’d be interesting to see where are we at nearly halfway through the year, not quite actually. So quite a bit less than halfway. Yeah, it is. But it makes me think I wonder what, you know, the big picture of trust and the reporting we see on that notably the Edelman Trust Barometer. What changes are we to see as this year plays out as it were? We have a war in the Middle East that

Shel: It’s still going pretty fast.

Neville: Anyone who even has a fleeting interest what’s going on in the Middle East knows that this is situation that has been the case for millennia, frankly. But in modern times, this has been since 1948 in the creation of the State of Israel. This has been happening in the Middle East. A war one way or another between tribal factions and then states have gotten involved in this. Iran

from what I can understand, has long been a thorn in the side of US governments over different presidents over the decades. It doesn’t resonate that way here, notwithstanding some things that have happened, but there were decades ago in the UK that, you know, the notion of the fact the US really was the only country that could do what they did to to bomb Iran and start a war to undeclared.

not asking anyone to help them and then complaining when no one came to their help. So it’s a dreadful situation, the war itself, obviously, but also the murkiness of what it has created in the context of what we’re talking about. That, you know, we’ve talked about this element before, which is that, you you do not control the message anymore, even if it’s about you. That is no

more true than what we’re seeing right now. The Iranian government doesn’t control any of the messages, not really. It’s anyone who’s got an internet connection and a tool to create an AI generated video or whatever it might be, and then share it online. That’s who’s got control, but only in a limited way because it then goes out there and anyone can do anything with it. It’s making it onto some

traditional media, not just social. So it’s who knows where it’s all going to go, Shell. And as this war continues without any sign that it’s going to suddenly stop, this is the new normal.

Shel: Yeah, and keep in mind, we’re not talking about a single piece of content. We’re talking about flooding the zone with multiple pieces of content that reflect the same grievance and make the same points and have the same punchline that get people to watch it and have it appear wherever you might be getting your content. So you got to look at it that way and take steps to deal with because.

Yeah, I don’t have a good business example yet, but I’ll happily make a bet with somebody that within two years, we’re going to see this kind of content aimed at business from that disgruntled investor or unhappy customer or whoever it is. So this is so easy to do. And that’ll wrap up this episode of For Immediate Release. Our next monthly long form episode is scheduled to drop on Monday, May 25th.

Neville: Great fun.

Shel: So we’ll be recording on Saturday, May 23rd. In the meantime, we hope you’ll comment. As always these days, all of the comments that we shared in this episode came from LinkedIn. And you’re welcome to look for our announcements of new episodes on LinkedIn or Facebook or threads or blue sky. And we’ll check for comments there, but you can also send them to [email protected].

I’m going to come up with a contest and probably announce it in the May episode for an audio comment. Anybody who submits an audio comment will put your name in a hat and draw a winner and you’ll get something. I’ll have to figure out what. We don’t have FIR merch anymore. Maybe you want to start that again. Now we’ll come up with something. We’ll come up with something. But you can leave an audio comment by attaching an MP3 file to an email.

Neville: Ha ha ha ha.

No, we don’t. Maybe we should, we’ll come up with something.

Shel: Or clicking the record voicemail tab on the right-hand side of the FIR Podcast Network website. You can comment on the show notes that we leave on the FIR Podcast Network. So many ways to leave a comment. And we also have a community on Facebook and an FIR page on Facebook. Any of those places will do. And we also hope that you will leave your ratings and reviews of FIR.

wherever you get your podcasts. And we will be resuming our short midweek episodes next week. Look out for those. Best way to have those is to subscribe to For Immediate Release. And that will be a 30 for this episode of For Immediate Release.

The post FIR #511: Doing AI Governance Right and Still Getting It Wrong appeared first on FIR Podcast Network.

Employees have long found ways to use software tools to get the job done, even when those tools are not approved. It’s called Shadow IT, but ever since generative Artificial Intelligence hit the scene in 2022, employees have adopted a new version: Shadow AI. The company approves Microsoft Co-Pilot, but employees opt to use their smartphones or personal laptops, along with their personal accounts with ChatGPT, Gemini, Claude, Midjourney, or whatever best suits their needs.

For most companies, this is a problem that needs to be addressed through repeated policy announcements and vigorous crackdowns. One company, though, took a different approach. In this short, midweek FIR episode, Neville and Shel outline what the company did and how communicators might advocate for a version of this approach to aiding in AI adoption and speeding up productivity gains.

Links from this episode:

• The Hidden Demand for AI Inside Your Company
• Shadow AI Threat Grows Inside Enterprises as BlackFog Research Finds 60% of Employees Would Take Risks to Meet Deadlines
• FIR #419: Is Shadow AI an Evil Lurking in the Heart of Your Company?
• The Rise of Shadow AI is a Double-Edged Sword for Corporate Innovation

The next monthly, long-form episode of FIR will drop on Monday, April 27.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Shel Holtz: Hi everybody, and welcome to episode number 510 of For Immediate Release. I’m Shel Holtz.

Neville Hobson: And I’m Neville Hobson. There’s a quiet tension playing out inside many organizations right now. On one side you have leadership teams, IT, legal, and compliance, all trying to put structure, governance, and control around how artificial intelligence is used at work. On the other side you have employees who’ve already moved on. They’re not waiting for official tools. They’re not sitting through pilot programs. They’re not asking permission. They’re opening ChatGPT on their phones. They’re using Claude in a browser tab. They’re experimenting quietly, often invisibly, finding ways to make their work faster, easier, and sometimes better. And in many organizations, this shadow AI behavior is still being treated as a problem — something to restrict, monitor, or shut down. It’s a topic Shel and I discussed on this very podcast in episode 419 nearly two years ago, and it hasn’t gone away.

Neville Hobson: In fact, recent data suggests it’s accelerating. A study last November by Blackfog and Sapio Research found that nearly half of employees surveyed in the UK and US are using unsanctioned AI tools. Even more striking, 60% said they would take security risks with those tools if it meant meeting a deadline. So this isn’t fringe behavior — it’s become normal. An article in the Harvard Business Review this month argues that instead of treating unauthorized AI use as a compliance issue, organizations should see it as a signal — a sign that people are already finding value in these tools, even if the organization hasn’t caught up. We’ll explore that idea in just a moment.

Neville Hobson: The article calls this the hidden demand for AI inside your company. And when you look at it through that lens, the picture changes quite dramatically. Because instead of asking, “How do we stop this?” you start asking, “What are we missing?” The piece goes further than theory. It looks at what one organization actually did when it recognized this dynamic: BBVA, a Spanish multinational financial services company with more than 125,000 employees. Rather than clamping down on shadow AI use, they moved quickly to provide a secure enterprise environment. But more importantly, they didn’t try to control everything from the center. They took a different approach. They identified and empowered what they call “champions” and “wizards” — the people already experimenting, already curious, already building things. They created a network, a community of practice, a way for ideas, use cases, and practical solutions to spread peer to peer across the organization.

Neville Hobson: And the results, at least as reported, are striking: thousands of employees actively using AI tools, thousands of internally created applications, and measurable time savings of hours per person every week. But perhaps the most interesting part isn’t the numbers — it’s the philosophy behind it. The idea that successful AI adoption doesn’t start with a perfectly designed top-down strategy. It starts by recognizing that innovation is already happening, just not where leadership expects it. So the question becomes: do you try to control that energy, or do you find a way to harness it? And that opens up a much broader conversation, one that goes well beyond technology. It touches on leadership, trust, and culture — on how change actually happens inside organizations. And, importantly for communicators, on how you surface, legitimize, and guide behavior that may already be happening under the radar.

Neville Hobson: Because if employees are already using these tools — and most evidence suggests they are — then silence or restriction alone isn’t really a strategy; it’s a gap. So in this conversation, we want to explore that gap. What shadow AI really tells us about organizations today, whether the BBVA approach is something others can realistically replicate, and where the risks still sit, because they have not disappeared. And we should be clear: BBVA may be an outlier. It’s a highly data-mature organization with strong leadership alignment. Many organizations don’t have that foundation. So the question isn’t just whether this works — it’s whether it can work anywhere else. And what that means for the future of work, and for the role communicators play in shaping that future. Shel?

Shel Holtz: Well, a few thoughts, starting with the fact that BBVA has the financial resources to provide a secure environment for those tools that employees are using. There are many organizations whose IT budgets are razor thin and don’t have those resources, so they would need to figure something else out. But I think there’s a caution here worth raising. The numbers from Blackfog are real, even if the framing from the Harvard Business Review is optimistic: 34% of employees using free versions of tools when paid, approved versions exist; 58% of unsanctioned users on free tiers with no enterprise protections. The reframing from threat to signal doesn’t eliminate the exfiltration risk — it reframes how we need to respond to it.

Shel Holtz: Communicators should be careful not to let the BBVA-style narrative become an excuse to ignore governance. The right frame is: harness the demand, don’t suppress it, and build the governance at the same time. Employees using unsanctioned tools and putting secure data and company information into them — that’s a governance risk, and I don’t think we can ignore it. I mean, I think what BBVA did is great, and I think they baked it into some governance while looking at a new approach they could afford to take. But for many organizations, governance is still a requirement.

Neville Hobson: Well, I agree. It’s important and it’s not to ignore by any means. I think, Shel, you fleshed out a little bit the survey that I mentioned, which is actually useful to have that level of detail. But the big question for me is: if this is the picture in many organizations, according to that survey — compared to data previously — this is getting worse, or rather, it’s happening more frequently. People are just going ahead and using what works for them as opposed to what’s the official thing. What is that a symptom of? Maybe a lack of trust? It’s probably a mix of things. And to me, the communicator’s role here seems to be to try and help people on the one hand understand what the tools can do for them, and on the other hand to help the organization understand that we need to address this issue. People aren’t using the approved ones. They’re doing stuff on their own, and that isn’t good.

Neville Hobson: You mentioned security risks. The Harvard article goes into some detail about that, as indeed do the people who conducted that survey. You can just picture the severe risk. We’ve seen examples in recent months of organizations that have suffered from unauthorized use of unapproved software tools — not necessarily generative AI tools, but software certainly. And it’s a big deal. So the question — do you try to control all this and look at ways to stop it? — we asked this very question two years ago in our conversation, and we could probably just insert the recording from then and replay the answer. But let’s talk about it. I don’t think they should try and stop it personally. That’s a fail. There’s no win in that at all, certainly not for the organization. So how would communicators go about that, do you think?

Shel Holtz: Well, I’m not suggesting that organizations crack down on this and become Big Brother, looking at the tools that people are using, especially when they’re using them on their personal phones or personal laptops. But there are definitely things communicators can do. The first is to surface and amplify the internal use cases — not just the fact that people are using these tools, but what they’re using them for. When the security people and the legal people find out that this is actually driving effective work product from these employees, I think there might be more appetite for figuring out a way to bake this into the governance documents and policies the organization has established.

Shel Holtz: And I think giving employees permission narratives — telling them it’s safe to experiment, letting them know how to do it, and suggesting where the guardrails are — matters. So if you are using shadow AI, here are the things to be careful about. Let them know what the risks are and how to avoid falling into those traps. Communicators can also translate the IT and legal guardrails into plain language that doesn’t read as prohibition, because prohibition just leads to negative thoughts from employees about the organization, and then they’ll just continue using what they’re using. And then there’s collecting and routing the demand signal back to leadership. Why are employees using these when there are approved tools around? What are the advantages? So that leadership can make investment decisions that match the patterns of usage employees are actually engaged in. There’s a lot of work here for communicators that goes beyond simply saying, “Don’t do this.”

Neville Hobson: Agreed. And in fact, you can learn from much of what BBVA did, even if you’re not an organization with that established foundation and 125,000 employees. They did things most companies aren’t going to be able to do. For instance, they reached an agreement with OpenAI and deployed a customized version of ChatGPT Enterprise in a secured, exclusive cloud just for the company. The reasoning is interesting. What the Harvard Business Review report says is that the strategic decision was clear: it was more dangerous to have unmanaged, hidden AI usage than to rapidly deploy a managed, secure solution that aligns with existing needs. Most companies aren’t going to be able to do that. So it comes back to perhaps what you’ve just proposed — explaining it to people, the pros and cons, the risks, and so forth.

Neville Hobson: But I think you need more than that, too. Otherwise, you’re going to have significant numbers of people who will ignore it and just go ahead anyway with what they’ve been doing. So maybe elements of what BBVA did — for instance, the network of internal champions and expert wizards to spread knowledge, rather than the formal top-down communication you might expect. You’d have people within the organization who are knowledgeable, who have a history of responsible use themselves, who can help explain to others and help them replicate that. You end up, I think, with steps toward broad compliance that everyone can buy into. That would be helpful, because I can see that the idea of anyone in an organization of whatever size just doing their own thing with whatever tool they like is not a good idea at all.

Neville Hobson: And that isn’t unique to this. We’ve had that kind of conversation in decades past about software. I remember when Hotmail first came out, and when Microsoft Network first came out, the arguments in organizations — and indeed the one I worked for at the time — was, “You’re not allowed to use this on your company laptop, so use it on your own,” stuff like that. That’s definitely not a good thing. So you need to act to address issues like that so that people trust you and respect you and are willing to follow a restriction — or a behavior change, if you like — that would help. It’s interesting, the learning you can get from BBVA’s example, even though you’re not an organization that size with a budget to match. It’s a lot about education. It’s trusting employees, absolutely, as you pointed out, Shel. But I think that’s a two-way street. You need to have a quid pro quo: if you have these freedoms to use whatever you want, you need to do it responsibly. Share your learnings with others in the organization. Things like that. To me, that seems like a really good place to communicate.

Shel Holtz: Yeah, there’s communication happening at BBVA. They have 11,000 active users and 4,800 custom tools being used by those folks. That didn’t happen because the communications department posted an article about them. This was peers talking to their peers about what was working. It validates something you and I have been talking about for years, which is that authentic, lateral, employee-to-employee storytelling beats top-down cascades every time.

Neville Hobson: Precisely.

Shel Holtz: But it is communication. And why wouldn’t that be something the internal communications department jumped on and helped to facilitate — providing the channels for that, rather than the sneakernet that’s probably happening now? And also, because they’re engaged and trying to keep this from happening below the surface, they’re in a position to identify the use cases worth taking to leadership. The Blackfog survey you referenced found that almost 70% of C-suite executives believe speed is more important than privacy or security. So if people are getting things done faster — if you can demonstrate that there actually is productivity improvement happening, and it’s because of the tools employees are using that aren’t approved — I think that’s motivation for leadership to look at either approving those tools or finding ways to allow people to use their own accounts while protecting the integrity of their data.

Neville Hobson: Yeah. The results the Harvard Business Review reports from BBVA are worth noting, even though the scale isn’t what many companies would experience. They talk about 80% of usage of the system they set up coming through direct chat prompting, and the remaining 20% through employee-created GPTs. Now, this is not shadow AI — it was part of the rollout of what they did. But these numbers are quite impressive. Over 83% of employees now use the system every week, averaging 50 prompts per week. That’s above comparable enterprise deployments, says the review, quoting OpenAI. Users report average time savings of two to five hours per week — a number worth noting. More than 4,800 custom GPTs have been created internally, and they’re used three times more frequently than the enterprise average. So they’re ahead of the game in that regard. The article goes into more detail about which departments are more active than others, and so forth.

Neville Hobson: It also prompted a thought in my mind: the other surveys I’ve seen and other reporting on the resistance from leadership in organizations — that isn’t minor. It’s not a little thing. It happens, unfortunately, too frequently. I’m thinking of keystroke logging on employee usage, auditing computers surreptitiously and covertly without telling them, watching which apps they’ve installed — and indeed, probably more common, your company laptop refusing to install things that aren’t on an approved list, or reporting to IT that you tried to install stuff. This is a dreadful situation in organizations. It’s common, but we’re going to see more of it, I think, because that seems to be the way of the world these days on distrust. This is a diminished-trust environment we’re talking about. So in all of that, where do we sit in terms of enabling stuff like this? We can see the advantages of allowing employees to use tools like this. I think the better way is to try to do something within the framework of the organization — not, “Oh sure, go ahead and use ChatGPT whenever you want on any device, no big deal.” I wouldn’t be keen on that. I wouldn’t stop it, but I would look at ways of weaning people off that approach. We have to help them and encourage them to do this. And that, I suspect, is a hard task for communicators — to persuade leadership to do that if the climate in an organization is resistant to it anyway.

Shel Holtz: Well, I think it is a hard sell to leadership, but we have data. We’re supposed to be engaging in two-way communication and facilitating two-way communication. One of the roles of internal comms is listening. And it doesn’t have to be through direct information that you get from people through focus groups or surveys — it could be this Blackfog survey. When 49% of employees are using unsanctioned tools, and 63% think that’s fine as long as there’s no approved option for what they want to do, you may look at that as rogue behavior, but you can also look at it as market research. And communicators are the people in the best position to translate that data into something actionable for leadership. You take that to leadership and say, “Look, this is what’s happening. We’re the ones who can interpret what the behavior means and pass that along to leadership.”

Shel Holtz: I think part of our role is that listening through the data that’s already out there — and maybe what we can determine is going on in our own organizations — and taking that to leadership and saying, “Look, this isn’t going to go away if you crack down on it. It’s not going to go away if you block installation on company laptops. People have their own phones. People have their own laptops and tablets. This is going to continue.” And this isn’t new. I mean, this goes back to the earliest days of computers. I think I’ve mentioned this once or twice on the show, but I needed to produce charts and graphs in the mid-‘80s, and I wanted to use Harvard Graphics because somebody had shown it to me and it was what worked, and the company had a different program that was terrible. So I just used Harvard Graphics. I bought my own copy and installed it. There were no blocks back then — you put the floppy disks in the drives and it installed. People are going to do what they need to do to get the job done. Maybe some will pay attention to what the official rules are, but I think the governance needs to be flexible enough to adapt to this. I applaud BBVA for what they did. Again, I don’t think every organization is in a position to replicate it, but I think you can take lessons from what they did.

Neville Hobson: You can. Not everyone can roll out what they rolled out — enterprise licenses and so forth — but some of the things they went about, and how they went about them, definitely. One thing the review article points out quite strongly — a very, very good thing — is that they say, toward the end of their conclusions, that in whatever you do, there must be a hard human-in-the-loop rule. Human employees should always own the work. There should not be direct writes to core systems. Internal GPTs need quality scores and guardrails. They specify scope and context, include samples, and so on. This is simple, scalable, and non-bureaucratic.

Neville Hobson: So that’s something that kind of ties back into this emerging phrase — if it’s even emerging — of human-centered AI. Let’s look carefully at this. It’s about people first, technology second, and the human needs to be in the loop. The “hard human,” as the review calls it — I interpret that as meaning someone who’s actually cognizant, aware, and able to act upon things that matter, to keep humans in the loop, to own the work, not the technology. You’ve got to think about things like that. And I think for communicators, that’s an important aspect of what they do — having in mind that element that is about the people first. So when you’re trying to persuade leaders to take a course of action you’re recommending, this needs to be in your mind too: that the humans need to be in control.

Neville Hobson: I have to say, this is great. I love stories and examples like this. I love them more than the ones that talk about disasters, although those are useful to know about as well. Yet I feel, as communicators, we have a constant, constant task on our hands to explain this to people in organizations, to help others understand. I think this is a good example — the shadow AI element. For me, if I were actively involved in an organization as the communications person, I’d be looking at: how do I persuade people not to do that? How do I persuade people to use the approved stuff? But at the same time, how do I persuade the leaders to make sure they offer employees stuff that actually works, that’s in line with their expectations, all that kind of stuff? There’s a bit of a job on their hands. And if budgets get in the way, then you’ve got an even harder job. But hey, that’s what we’re here for. That’s part of what we have to do.

Neville Hobson: These are good examples you can learn from. There are elements you could start on. And I think, like most things, Shel, you need to say, “OK, fine — this idea has a dozen constituent elements, and let’s just start with two.” So you don’t try to think, “Oh my god, this is a massive project. How on earth can we do this?” You look at just a couple of things. I like another point the Harvard Business Review makes: ensure that managers know what they’re doing. You can’t expect managers to be persuasive in encouraging others to use AI if they’re not good at it themselves. So there’s another element — you need to train them well, says the Harvard Review. At a minimum, they should learn how to write staffing notes, sensitive communications, and KPI reviews with AI help. So there are some things you could do straight away as a communicator in an organization. I’d say: good luck and godspeed, and it’ll all work out in the end.

Shel Holtz: Yeah, a manager’s role in all of this is probably an episode in its own right. I would just reiterate the point you made about the human in the loop. This is a governance element that should be overarching — not applying just to shadow AI, but to all use of AI in the organization. It should be a primary consideration in governance, not to turn things over to AI. Otherwise, you end up with fake citations going out to clients that paid a million dollars for your work — another little slap on Deloitte’s wrists. And that will be a 30 for this episode of For Immediate Release.

 

The post FIR #510: Should Companies Embrace Shadow AI? appeared first on FIR Podcast Network.

When bad actors use AI tools to clone a musician’s voice and upload synthetic versions of their songs, they can then file copyright claims against the original artist’s content — and win, at least initially. That’s because the systems platforms used to validate copyright claims are automated and configured to treat whoever files first as the rightful holder. The result: musicians like Murphy Campbell, a folk artist from North Carolina, lose both revenue and control of their own creative identity.

The same mechanism works just as well against any organization that publishes audio or video content online. In this midweek episode, Shel Holtz and Neville Hobson break down how the scam works, why it matters to communicators, and what you should be doing right now — before an incident forces your hand.

Links from this episode:

• AI Cloned Her Voice, Then Claimed Her Songs
• ‘This Is Not Me’: Inside the AI Scams Driving Musicians Crazy
• A Folk Musician Became a Target for AI Fakes and a Copyright Troll
• A traditional musician became a victim of AI imitations and a copyright aggressor
• ‘AI slop’: Emily Portman and musicians on the mystery of fraudsters releasing songs in their name

The next monthly, long-form episode of FIR will drop on Monday, April 27.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email [email protected].

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Neville Hobson: Hi everyone and welcome to For Immediate Release, this is episode 509. I’m Neville Hobson.

Shel Holtz: And I’m Shel Holtz. And today we’re going to talk about something else that communicators need to worry about. I think we need to develop a worry list for communicators. This one starts with a tale about a folk singer from the mountains of Western North Carolina. She’s named Murphy Campbell. She plays banjo and dulcimer and records old Appalachian ballads, some of them written by her own distant relatives. And she posts videos of herself performing in the woods. She has about 7,800 monthly listeners on Spotify. And she is, as Shelly Palmer put it in a recent column, exactly the kind of artist the copyright system was designed to protect.

In January, some of her fans started messaging her about songs on her Spotify profile that she had never uploaded. Someone would have taken her YouTube performances, run them through AI voice cloning tools, and posted synthetic versions of her songs under her name on streaming platforms. These fake tracks, to put not too fine a point on it, were really bad. Her dulcimer sounded like — and these were her words — a warbled metallic mess. Her voice had been deepened and auto-tuned into what she called a bro country singer. But here’s where it gets interesting for those of us in communications, because that’s not the end of the story. It didn’t stop at impersonation.

Whoever uploaded the fakes through a legitimate music distributor called Vydia (V-Y-D-I-A) then filed copyright claims against Campbell’s original YouTube videos — the very videos the AI had been trained on. Because YouTube doesn’t use humans to review initial copyright claims, Campbell stopped earning revenue on her own content. That revenue started going to the person who had filed the copyright claims.

She described herself as being in a weird limbo where “I’m telling robots to take down music that robots made.” Shelly Palmer called this a reverse copyright scam, and he confirmed, speaking to other content creators off the record, that this is more common than he might have believed.

Now, I know what you’re thinking — music streaming platforms, artists, what does this have to do with me? And the answer is everything. Because the mechanism that elbowed Murphy Campbell out of earning royalties for her own music will work just as well against any organization that publishes content on platforms with automated enforcement systems. That is virtually every organization that has a YouTube channel, a podcast feed, or any kind of public video or audio presence.

So here’s the structural problem as Palmer frames it. The copyright system we have was built on a foundational assumption that the first entity to register a claim is the rightful owner. That assumption held when human creativity was the bottleneck. It breaks completely when AI can generate a synthetic version of any content in seconds using any voice. Think about what your organization puts out there publicly — executive speeches, earnings calls, thought leadership videos, branded audio, training content, podcasts, content marketing pieces. Every one of these is a potential training data set for someone who wants to clone your voice, your leaders’ voices, and then upload a synthetic version through a low-cost distributor. We’re talking about something that costs $25 to $90 a year. Then they file a claim against your legitimate content before a human ever reviews it.

Neville Hobson: (pause)

Shel Holtz: That means the system is going to see them as the first one to file that claim and assume they are the legitimate copyright holder. Now, Rolling Stone confirmed that this isn’t an isolated case. Paul Bender, Veronica Swift, Grace Mitchell — these are just a few of the artists who have faced the same attack. One musician even ran an experiment he called Operation Clown Dump, uploading fake content under his colleagues’ names across platforms. His success rate was 100%.

So what do communicators need to do? First, audit your public content footprint. Do it now, before an incident forces you to. Know what you’ve published, where it lives, and what revenue or visibility is attached to it. Second — and here’s something that’s new for a lot of communicators — register your copyrights. Formal registration is the prerequisite for meaningful legal recourse in the United States. Third, build a rapid response protocol for platform disputes. The organizations that survived these attacks quickest were the ones who knew who to call and knew what to say. And fourth, have this conversation with your legal team today, not after something goes wrong.

Murphy Campbell eventually got Vydia to withdraw its claims, but only after her story went viral. Most organizations won’t have that option. Your story won’t go viral. The bad actor doesn’t need to win permanently — they just need the automated system to act before you do. And that is the lesson, and it’s one we’d better learn from musicians before we have to learn it the hard way.

Neville Hobson: Extraordinary, isn’t it, Shel? I guess you could call it a new phenomenon, only in the sense of the speed with which this can be done. I must admit, I’m astonished that the system is such that the first person to file the copyright claim is assigned ownership. Maybe that’s similar here in the UK — every jurisdiction is different, of course — but that’s rather unsettling. It obviously goes back to a time when people weren’t exploiting the system the way they are now. There are similar examples here in the UK of this kind of activity where people unwittingly find that their content is being misused and misrepresented. And although no major artists — though I may be wrong about that — I did see an article noting that YouTube allows some users to clone the voices of stars like Charli XCX and Sia, with their permission. But unauthorized AI covers of artists like Harry Styles — hundreds of thousands of copies — is a widespread phenomenon, and one that barely registers in mainstream news.

A number of artists, a bit like your example of Murphy Campbell — there’s one I’ve heard about, Greg Rutkowski, a Polish-born artist known for his work on Dungeons & Dragons, who found his style being used in over 400,000 AI prompts, raising serious concerns about the obsolescence of human artists. And to your point about what communicators should watch out for: your corporate communication messaging that’s in audio, your CEO on an earnings call that’s been recorded and distributed. So never mind video — audio alone, at that scale of 400,000 AI prompts, is not a good situation. If you project the thinking out, this is utterly relevant to anyone publishing audio or audiovisual content online.

I find it astonishing that some platforms, notably Spotify — which features prominently in a lot of reporting on this — are being used to literally steal someone’s intellectual property by replicating it. And I think it reinforces the point that registering copyright isn’t an idle exercise. It’s something that should be front of mind, and it does other things for you as well as the owner of the property.

Something as simple as displaying a current copyright notice on your website — it’s remarkable how many sites I come across that still show “Copyright 2016,” never updated. Displaying a current notice signals that the business is active and its information is up to date. There are also tools to protect against AI scraping, though how effective they are is still unclear. Creative Commons licensing is another option, setting out the terms under which people can use your content — though that requires everyone to play by the rules, which frankly isn’t always the case these days.

Nevertheless, you’ve got some protection — or at least the peace of mind that you’ve taken steps. But it really is quite extraordinary, isn’t it, Shel? When I looked into what’s happening in the UK, I came across a recent movement — over a thousand UK musicians, including Paul McCartney, Annie Lennox, and Damon Albarn — who released a silent album to protest proposed legislation that would allow AI companies to train on copyrighted material without consent. It struck me as a real head-scratcher: why would a government enable that to happen?

Shel Holtz: Probably very effective lobbying from the AI companies, I’m sure, is behind that.

Neville Hobson: No doubt, no doubt. But there are other things going on — organizations like the Musicians’ Union and Equity campaigning for better copyright protection, consent, and fair compensation for creators. It’s not getting much mainstream coverage, but activity is happening behind the scenes. Nevertheless, the example of Murphy Campbell and others represents a genuine threat that you need to be aware of if you’ve got content online that matters to you. Never mind the “they shouldn’t be doing this” argument — the point is, if it’s important to you, have you thought about this?

Shel Holtz: If you think about the days before the web, copyright wasn’t something most people had to worry about that much. Professional artists with record deals had people to handle it. Same with authors — someone like Stephen King never had to worry that somebody would be the first to file a copyright claim under his name and siphon off his revenue. But now you have artists who don’t get record deals — like Murphy Campbell — publishing on YouTube and Spotify, building small followings, and making a reasonable living. This is the working class musician concept we talked about, oh, it’s got to be 15 years ago now.

The fact is, you can use Spotify and YouTube to build a following, play some small clubs a few times a year, and make enough to pay the mortgage and put your kids through school. You’re not going to get the penthouse suite from playing to 100,000 people, but you can make a living. But this has also opened up the ability for bad actors to take advantage of that. And now with AI able to reproduce your voice and create new music at scale, all the pieces are in place for this kind of theft. Unless you’re able to get your story to go viral — as Murphy Campbell did — it’s not clear what you can do, because YouTube and Spotify have set up systems that automate this process with no human review. When you used to register with the copyright office yourself, a human was checking. So it’s not likely most organizations have revenue-generating content online — though I’m sure some do, and I’ve actually argued there are ways to use content to generate revenue.

For example, I’ve always loved the idea of a Webcor YouTube video series called “Building for Girls,” where our employee resource group, Women of Webcor, does a five-minute lesson every two weeks on construction to get young girls interested in STEM and engineering careers. Get enough views and YouTube starts paying you. If you don’t copyright-protect that content, someone can come along, produce similar videos, claim the rights, and suddenly your revenue is going to someone else. But even if you’re not producing revenue-generating content, there are other reasons to ensure nobody else can claim ownership of what you create — especially as content marketing demands more and more output. So yes, register that copyright.

Neville Hobson: Yeah, it made me think about watermarking for written content — though I’m not sure there’s something truly effective offering the same protection for audio and video yet. And even if there were, you’ve got situations like Murphy Campbell’s, where it’s her style and tone — the whole persona that defines her music — that’s being copied. And you don’t know about it until strange things start happening: your revenue drops, someone says “I love that new song you just published,” and you discover it wasn’t you. Or you read a review and think, wait — I didn’t write that.

Shel Holtz: Or “I hate that new song you published” — in Murphy Campbell’s case.

Neville Hobson: Exactly. I’m sure people are working on the technology. You’ve got digital rights management, which isn’t new, but I’m not sure it helps here because the issue isn’t copying your content outright — it’s imitating or repurposing it at scale. Hundreds of thousands, or millions of instances. I think the platforms need to do far more than they currently are. It’s a similar argument to what we’re hearing here in the UK about Meta and X doing nothing effective to protect children. This is in the same territory, and it needs a lot more from those platforms — who are making serious money throughout all of this. As to what exactly “more” looks like, I’m not entirely sure, but they need to do more.

Shel Holtz: Yeah, and they probably won’t until there are some high-profile, visible court cases that create real reputation issues for them — then they’ll take action. The easy thing to do right now is simply register the copyright. That’s your protection. When someone imitates you, or claims the content you produced is theirs, you have legal standing to act. That’s why you need to have this conversation with your legal team.

But I wouldn’t wait for either the platforms or the government to do anything. They’re both reticent to act. You have the ability to do something about this right now, and it’s just a matter of working with your legal team and filing those copyrights.

Neville Hobson: Yeah, exactly. And even using Creative Commons licensing — if you’re an individual without all the formal resources, but you have a niche following, even that’s a start. Keep a record of every iteration of everything you’ve created — “I did this in 2017, here’s proof, backed up here.” That gives you something to stand on, a way to demonstrate that you can act if someone uses your content. And if you don’t do this, there’s another consequence worth considering: your original content gets buried in search results because the AI-generated imitations have somehow accrued better signals to rank higher. That kind of pollution from AI slop is its own problem.

Shel Holtz: Yeah — and then people stop paying attention to your content altogether because they’re so fatigued by the AI slop that they tune everything out. But at least this one has a solution communicators can follow: something new to add to the copyright to-do list. And that will be a 30 for this episode of For Immediate Release.

The post FIR #509: Does Corporate Content Need Copyright Protection? appeared first on FIR Podcast Network.