Share this
by The Notified Team on Oct 21, 2025 10:06:40 PM
AI is reshaping how companies get discovered online, and press releases are at the center of it.
That was the focus of our recent webinar, Rethinking Press Releases for the Age of AI Search.
This 45-minute session brought together industry experts to unpack how press releases directly influence AI search rankings, what’s changing in journalism and how public relations teams can adapt for maximum visibility.
In this blog, you’ll learn:
- Why AI search is reshaping how press releases are discovered
- What journalists expect in the AI era
- How to optimize press releases for both humans and AI crawlers
Why Does AI Search Change the Game?
Ella Darlington, Associate Director at ThoughtLDR, set the stage:
- 40% of PR tasks in the UK are now AI-supported, and its use across business is only set to rise.
- With opportunity comes responsibility. Communicators need to balance trust, transparency and credibility by making their content easy to find in AI-driven platforms.
The takeaway? AI is raising the stakes. If your content isn’t credible and structured, it risks being overlooked by both journalists and generative engines.
What Do Journalists Expect in the AI Era?
Journalist and AI strategist Hina Pandya shared both opportunities and risks.
AI helps journalists check facts and gather data, but it can also bring in bias. At the same time, many are getting flooded with AI-written press releases and when those releases are sloppy or hard to verify, it quickly erodes trust.
Her advice was refreshingly practical:
- Keep titles direct and simple.
- Put the most important information at the top of your release.
- Be ready to back up claims with sources and case studies.
- Above all, invest in human relationships with journalists.
As Hina put it: “If you do send a press release, be prepared that you might have 10 journalists calling you that morning. Make time for them and they’ll make time for you.”
What is Generative Engine Optimization (GEO)?
Paul Stollery, co-founder and Creative Director at Hard Numbers, explained how PR is entering a new discipline: generative engine optimization (GEO).
Why does it matter? Because platforms like ChatGPT, Gemini and Perplexity increasingly shape how people discover brands.
Paul’s key takeaways:
- Earned media is still king. Media coverage is the top driver of brand visibility in AI tools.
- On-site content matters. Pages like “About Us” and FAQs are frequently pulled into these large-language model (LLM) answers.
- Bad content is riskier than ever. Poor-quality, AI-generated releases can hurt reputations or even get PR teams blocked.
How Can Press Releases Be More Strategic?
Philip Hodge, Sales Engineer at Notified, emphasized one theme: structure matters.
Press releases that mix human readability with AI optimization perform best.
Think:
- Short paragraphs and bullet points
- Direct, quotable soundbites
- Writing that answers specific questions journalists and AI might ask
He also underlined the value of feedback loops: combining human relationship-building with data-driven monitoring to see how press releases perform in search, in syndication and increasingly in generative platforms.
What Should PR Teams Do Next?
The discussion made one thing clear: The fundamentals haven’t changed: credibility, clarity and relationships remain at the heart of PR.
What’s new is the wrapper - how content is packaged for both people and AI.
Done right, your press release doesn’t just reach the media. It makes your story discoverable in AI search, trusted by journalists and credible to investors and audiences everywhere.
So, the question is: are your press releases ready for the age of AI search?
Subscribe to our blog to stay in the know on upcoming webinars!
Notified Webinar Transcript
Hina Pandya: Hello, I'm Hina Pandya. I'm a journalist, and I work on AI and strategy. At the moment, I'm focusing a lot on AI safety. So, I've worked in digital data journalism for at least 16 years now. So that's a lot of changes, a lot of shenanigans, and a lot of development. So, I think that's me in a nutshell for this.
Ella Darlington: Ideally suited for this conversation, of course. Paul, go next on my screen.
Paul Stollery: Hello, I'll go next. Thank you ever so much for having me here today. My name's Paul. I'm the creative director and co-founder of a tech PR agency called Hard Numbers. I won't give you my full bio, but I started my career in the very early days of SEO and then moved into PR where I've been for most of the time since. And AI in PR feels very reminiscent of that, in that we all know it's important. There are loads of data, but there’s often lots of confusion as well. And sometimes the data makes it even more confusing. So, I spent the last year or so trying to figure out what's going on and everything from generative engine optimization to AI optimization of content. And that's why I'm super excited to be here today.
Ella Darlington: Fabulous. And without, and last but not least, Phil.
Philip Hodge: Perfect, it's lovely to be here. Thanks for doing this today. I'm here with Notified. I'm one of the solution specialists, really going into the technical side of how we can enhance and grow and engage with our audience through press releases and through monitoring tools we provide, and we work on a daily basis. So, this is really an area that I think is growing, as Paul said, very reminiscent of the SEO and the challenges that we see ahead being something we need to focus on, rather than something that we can just put in that hole and ignore until we disappear. So yeah, really excited for this one today.
Ella Darlington: Fantastic. Setting the scene here. AI is now apparently, according to CIPR research, assisting 40% of PR tasks in the UK, which is not insignificant. And the overall use within all businesses is set to rise. That's, you know, regardless of whichever research study you look at. There were some predictions at the beginning of the year. PR Week spoke to a lot of pros at the top of their game. One such individual, the lovely Anna Rashid from Tangerine Comms, spoke about what she was observing in terms of AI-assisted, audience data-driven insights this year. There'll be trend analysis, and we'll use it for enhancing evaluation of things, albeit more expediently. The real gains, this was kind of a consensus across that group, will possibly happen more into 2026, 27. So really interesting point in time that we're just before this precipice of it absolutely exploding, but this feels like such a fuzzy medium that we're catching ourselves in between. I think what's really important is to start off this conversation, and what I think so many people on this webinar will find huge value in, is if we start initially looking at it from the journalist perspective: what has changed? What expectations on your role have changed? CRPR studies also identify that we need humans more than ever, regardless of what the AI interaction or interface looks like. Do you feel that it still applies to you in journalism? Talk to us through some of the changes that you're seeing in your day-to-day and the demands that are being placed on you, Hina.
Hina Pandya: Yeah, I mean, obviously, AI has revolutionized, tech's revolutionized the media industry. And we've been hemorrhaging jobs in our industry for a long time. And AI makes no, you know, the same deal within our industry, in terms of how we're using AI. We're using it for fact checking, we're using it for amalgamating a lot more data. In that sense, that's not new. We've been doing that for a long time. We can harvest code, do code and investigative journalism. We can scrape that all together anyway. AI might make it easier, but then there are biases. And in journalism, we have to be really careful about those biases and how they affect our readers, our target audiences, and generally the ramifications it'll have in global society. So, fundamentally, it's changing. How does that change our relationships with PR? You might be aware, I don't know, of PRs being, well, journalists being flooded by PR AI releases. And we're getting so many of them, and we're sort of sharing that in our forums and talking to each other like, are you getting a lot of this? Because even if we pick one of those up and then explore it, because that's an interesting point or story. We want to research it. We want to know where your sources are from. We want to know your case studies. And I don't know if you know this either, but two stories have made it into mainstream press, one being the King's Cleaner, which apparently was an AI story. Wow. And the second one was about three or four different lottery winners who didn't claim after dramatic stories of "the dog ate it," that kind of thing. And they did make it into mainstream press. And then a really great journalist, Rob Wolf, has written about it, saying that when they challenged and asked the PR agencies that they'd come from, they were like, "Oh, we delete case studies after 30 days," and this kind of thing. So that's an impact on our reputation, of course, on the newspaper, of course, on the readership, because nobody really trusts journalists anyway right now. So how does that it's just even worse. So, I agree with you. Human relationships have to take precedence. In the old days, it would be a PR who would phone me, and I'd make it, there'd be a really good relationship. I trust them, they trust me. We get the right story out, the right angle for them, their client, possibly. If I couldn't do that, then there'd be a negotiation of what I could do for them, or perhaps we didn't run the story. But that's the nature of it. It's a human relationship.
Ella Darlington: You've covered loads. I think folks will be interested to understand what you're assessing for, right? So, when you get those initial pitches in your inbox, what are you screening for? What gets into the maybe or yes pile and what gets an immediate absolute no? For what reason?
Hina Pandya: That's a really good question because it's not changed, right? I'm still looking for the story I can sell to my editor. I'm still looking for the story that's going to turn the story up a notch. So that's not changed. What really works for me, and is kind of an SEO sort of tool, is don't be funny in your title. Give it to me straight, three or four words, cut it down. And as soon as I open that email, I want to see the first three paragraphs, the first three lines, not paragraphs, sorry, didn't mean to excite anyone, but put that in there and I'll know straight away if that's something I can use. Things that turn me off, things that block me at the moment, are certain press release companies. We all talk to each other, and we know what to avoid, especially because, as freelancers, we don't want our reputation tarnished. So, either you can call that person and say, "Hey, you sent me this press release. Can you talk to me about it?" Or if that doesn't happen, if that can't happen, I'm not going to pitch it. So that's done, you know. So, if you do send a press release, be prepared that you might have 10 journalists calling you that morning. Make time for them, and they'll make time for you. I think we can get on really well when it works really well.
Ella Darlington: Thank goodness. And you know, we've got things like tech JPR being reinvigorated, so all power to those groups and things that connect us and unite us. Paul, I want to hear more about this from your angle because you mentioned GEO right at the top of your self-descriptor. That's intrinsic to a lot of the work and the studying that you've done. Initially, help us define that and how you guys are looking at that at Hard Numbers. Also, as part of this broader conversation, what are the expectations on how we draft things, just thinking about content as a whole and GEO as a filter. Has this changed the nature of how we write, the length, the style, query-driven subheads? Help us understand it.
Paul Stollery: Yeah, of course. So generative engine optimization is easiest to find when you compare it to other things, right? Earned media is getting your message out in the news or in the trade press. SEO is ensuring you have visibility in search engine rankings. Generative engine optimization is ensuring that your message is getting carried through various LLM-driven chat platforms. Now, there's a huge overlap between all of this. I mentioned at the top of the call that there’s loads of confusion, loads of data out there. We actually, at the end of last year, did a big research piece into what drives answers on these LLM-driven chat platforms. There are lots of nuanced findings, but the key thing is: the more things change, the more things stay the same. So, the biggest driver, which is obviously good news, is earned media was the number one source of answers about brands on these LLM-driven platforms like ChatGPT and Perplexity. The second most important one is on-site content, about 60%. The report is on our website, hardnumbers.co.uk, and it’s on the homepage under “Reputation in the Age of AI.” The number one piece was media coverage, and the second piece was on-site content. If you look back at when I started, everyone was talking about off-site and on-site, which is in media and on-site. Everyone in PR has been talking about on-site and off-site or owned and earned content; the labels change slightly. With generative engine optimization, 80% of it is overlapping. I actually believe that the SEO team should never have been separate from the PR team. It was just because most PRs were a bit afraid of all of that sort of like meta tanks. That sounds a bit scary. But I think GEO represents an opportunity to bring those three things back together. Where it differs, from more traditional PR, from maybe SEO is just some more technical aspects. So, for instance, these LLMs love structured data. One of the top pages, actually the most popular page with on-site content, wasn't actually the homepage. Everyone in SEO thinks about home pages, landing pages. With GEO, if you're asking about a brand, the most popular page is the about page, which is often underloved. It's normally “chuck the bio up, we'll look at that again in five years.” Actually, it's one of the first areas LLMs tend to go through search. So, there's all sorts of nuances. They love FAQ pages, that are more popular than they would be on other channels.
Ella Darlington: Exactly, because FAQ pages are specifically engineered to answer queries, which search is increasingly structured around.
Paul Stollery: Yeah, it’s question and answer. There’s 20% nuance, and that’s important. Brands can get an edge here or be drowned out. But centrally, make sure we're doing the fundamentals right: getting the right coverage in the right publications and measuring it. The more things change, the more they stay the same.
Ella Darlington: Exactly. And what does that mean for your team when thinking about the touchpoint with journalists, which tends to be a press release? Does that alter the approach? Does it change what you get clients to sign off on or what you put in front of clients? Does that look different?
Paul Stollery: Yes. So, I think there's a wider point. When we did that research, we asked questions around brand framework: “Is Apple a trustworthy company? Is it innovative?” The answer was either yes or, as an LLM, “I can't form an opinion.” That's great news for PRs because we want them to say nice things about brands.
Paul Stollery: Yes. So, I think there's a wider point. When we did that research, one of the things we asked is, we basically developed the brand framework and we said, is Apple a trustworthy company? Is it an innovative company? Is it right? And we asked loads of brands across different questions and then looked at the sources. When you asked, "Is it a trustworthy platform?" and I was looking back at the source data the other day, "Is it a trustworthy brand?" the answer was always either yes or, as an LLM, "I can't form an opinion." So, it wasn't ready to criticize. And that's actually great news for PRs because we want them to say nice things about brands. But on a wider societal point, and obviously, yes, with PRs, but not to get too deep, we're kind of people first, and we do need journalists to hold brands to account. It's quite worrying. More broadly, on how our staff engage with journalists, the thing that I say to all of our staff is: it's riskier than ever to put poor content in front of journalists. Five years ago, if you put bad-quality content in front of journalists, that wasn’t good. You risk souring the relationship, failing to get coverage, failing to meet targets and that's really important. Now, if you put bad content in front of journalists, especially if it's AI-written, very often, if it's poor quality, they'll just say, "Is that written by AI?" You can get blacklisted. It's not just a one-to-one thing, an entire agency, as Hina mentioned, could get blacklisted from a publication. There's also the reputational risk. Not a week goes by where I don't see an angry tech journalist with a reputation for calling PRs out. They screenshot emails and say, "I've just got this," sometimes tagging the person, sometimes just sharing the screenshot. I saw one from a product tech journalist a couple of weeks ago, and the brand was still included. They blurred out the name, which I think is the right thing to do, because often it’s a junior journalist and there’s privacy to consider. But equally, the brand often ends up online, and there’s a reputational risk. The thing that I always say to my team is: it's more important than ever to consider our own reputation as a business and as individuals with journalists. Because the worst this gets and it is going to get worse and worse - the more important that reputation and that relationship will be.
Ella Darlington: Agree more. Yeah. There are thousands of AI tools now. You mentioned a couple of them, you mentioned Perplexity, you mentioned ChatGPT, but there are multiples of thousands. Around 40% of US VC funding went just towards AI-focused companies in the last year. So, they're just going to keep proliferating, and regulation might not be able to keep up. We really have to be custodians of what we put out and make sure that it’s of a good standard and that it maintains our integrity on both sides. Phil, I want to bring you in to talk about…
Hina Pandya: Can I just comment on Paul's point? Because it's such a good point. the relationship that we have with PRs nowadays, because of AI, because of our reputations, because we're worried for ourselves too. I mean, it's not just an agency worrying about themselves and their brand. Staff jobs are hemorrhaging, so we're all pitching for work. And if our reputation is sullied… But equally, what Paul was saying about putting something out there, we talk to each other. I don't know if you are following many Sun journalists, but there's a great travel editor who has put out the names of these PR companies. What we're finding is that the AI PR companies are changing their email addresses so they can bombard us again without being blocked. So, then what do we do? Do we make our emails private? Do we only? And then it just makes our jobs harder as well. So, it's a symbiotic relationship. And Paul mentioned before the AI-generated headline, that's all anybody reads. Journalists can spend hours on an article, and Google will condense it down. And is that what you need? Is that what we need, you know?
Ella Darlington: That's it. Sorry, my video paused there for some reason. I think I was just hooked on what you were saying, Hina. And it's so true. I think we're always keeping our reputation front of mind, but we're also responsible for the next generation coming up, aren't we both within our practices and in educating them around integrity and being a filter. Phil, I want to bring you in to examine and pick apart what you’ve identified as some of the things that might be broken, could be fixed, or where there are opportunities for us to improve in PR and journalist relations. If we think about it, press releases are the jumping-off point, but it's a broader discussion. What are some of the things you’ve dug into and discovered are potentially broken or could be done better?
Philip Hodge: I think when we're looking at this, we're all talking about the right side of things that transparency, the trust that we are building is paramount. We need to maintain that process as we go through. However, AI tools are there, but they are just a tool. They are there to enhance or enrich the way that we are working and bring greater reach and greater progress. But if we take away that humanity and creativity from a lot of the elements that we work through, we lose the side that we all enjoy and the interesting parts that could be picked up by someone doing their job as an investigative journalist. When we are working with clients, we do see a wide range of metrics that people class as successful. Some see this as just a matter of pushing us, getting as much publication and syndication as possible, getting it onto as many different sites as they can, that’s their big tick. Others only want to get it into one outlet; that’s all they care about. And I think, obviously, utilizing the tools in the right way can allow for those success metrics to go through. Is one right, is one wrong? It obviously depends on the way the PR is looking at their structure and what they’re trying to attain. Ultimately, this means the journalist is stuck in the middle. They’re getting bombarded because we want to go into that outlet, or they’re getting bombarded with information that is not always targeted or appreciated by their structure and what they’re looking for. And I think this brings into focus the challenge PRs have today in utilizing a single press release of 400–500 words while trying to target multiple different assets as we go through. We need to make sure that we’re hitting journalists, that it’s engaging, and that there’s an element of humanism. But we also need to hit the syndication, make sure it’s readable, and make sure it works for SEO, which traditionally has been a lot about the metadata, not so much about the content. But with the GEO element that Paul was talking about, that is conclusive to the text that we are writing and what we’re adding in. I think this is an area where people start to get a little scared that we are trying to write to an AI bot first and the human second, when in reality, the way the human mind works and the way bots are picking up this text is very similar. As you said, having easy-to-read and clear statements of answers, even the term AEO, answer engine optimization helps. Having that in your first sentence, making clear what we’re going to talk about today and what the answers are, makes it easy for me as a reader and easy for the bots. Bullet points, quotes, graphs, tables, these are all elements that our eyes get drawn to as humans, and they identify them as clear outlets and points that we need to work with and understand as the key message the press release is trying to push. They are also picked up from a GEO perspective, which is really interesting. As we start to build these approaches, we optimize the press releases for this purpose. It can go the other way, where it becomes a little formal and templated, but we can also build that relationship through personalized messages to journalists. So, we have our press release, but we also reach out to them. If I were a journalist receiving a message from someone I knew, and it was just spam, then yes, I might read it but I want to understand why I would potentially be interested. By adding personalized pitches and messages to preface those press releases, just to say, “I think you would be interested in this because I saw your piece last week, and it was really good” floating a journalist’s ego never hurts. So, there's ways that we can put in those multiple use cases and target effectively journalists, the code, the GEO, and obviously that human readability. And I think that side of things, if we can bring those best practices together, then we're going to have a successful return on the press release we push.
Ella Darlington: Absolutely. We're going to throw across to questions in just one second. So if you've been transfixed and you haven't noted down what your burning question is to the group, do bear in mind that we've got journalism experts, we've got search experts, we've got research experts, but with communication being our uniting force here in the room, please put those questions down. We might not be able to get through all of them, but I've got a fabulous team in the background who's funneling those through to me. But I did want us to leave everyone with a piece of advice that everyone can take away from some of this discussion. We've covered lots of ground. We've spoken about the realities of how the fundamentals haven't changed, the functions that we're fulfilling, and that integrity should remain intact. But sometimes the wrapper has changed, or the nuances have changed, or the structure. Perhaps one thing, Hina, I'll start with you. One thing is your bottom-line tip for the audience who are on this call? They're senior PR professionals, some might be in-house, some might be agency side, thinking about lots of different clients, time-poor, but earnest and really wanting to get it right. What's the one thing you would advise them when thinking about AI being in the back of their mind but maintaining a good relationship and being at the forefront?
Hina Pandya: I'd have to start with: your industry's changing, so is ours. Because we're symbiotic, keep that at the forefront of your mind. Follow how we're revolutionizing our own publications and our own media, whether or not you pitch to or want your content to end up in them. Like Philip said, multiple publications are syndicated, or if you're just targeting one particular magazine or newspaper, you have to understand that we are also changing our methodology, how we're aggregating algorithms and putting that out, one news section, or like Apple, which has its own news platform, an aggregated news set. We have to consider what we're using, which will impact how you work as PRs as well. So, I would say, as it changes, keep abreast of those changes. Keep talking to your journalists, keep your relationships going, reach out to them, we love cups of coffee, do invite us. And if you do events like this, invite us to that, because it's just as interesting for us at this time. Everything's changing so fast, and that will help your symbiotic relationships as you move forward, as we move forward. That's how media starts to… I mean, Rob Wolf, the journalist I talked to you about with the AI press releases, he says it's already going up in smoke. We're already apocalyptic in our coverage. You have journalists with very different perspectives, but absolutely, you know, it is changing. So keep an eye on what we're doing as you do what you do.
Ella Darlington: Absolutely. We're starting to even see some writing styles shift and change subheads, particularly with new media journalists first. We're really noticing a shift. I'm going to flip the order because Phil, we came to you earlier on. I really want to hear your parting piece of advice before we jump to questions.
Philip Hodge: Sureness, I think, actually, following this is perfect. I think that feedback loop that's coming through and the ability to check in and understand what is working and what isn't working is something, again, we can take from a human perspective. This could be obviously being published, getting direct feedback, or picking up the phone and talking to journalists. But we also have the ability to look at this from a data perspective as well. It could just be that we're not even getting through to the journalists, they're no longer opening our emails. We can track that to see whether that is taking place or not. At that point, that's when you pick up the phone and ask them, why? What is going on? Are we actually not relevant? And have that human side - don't just ignore it. But we also can go through to the generative engine elements that are coming through as well. I think being able to track how we're being used, the brand recognition, and the elements that are coming on AI tools is a huge part of where we need to ensure that we're keeping numbers, seeing where we're mentioned and how often, the same way we used to, we started to track or continue to track SEO. GEO is also where we need to understand that data, and obviously that grey hole that we currently have at the moment, something that we are fixing with AI visibility. Ensuring oversight of that metric helps that feedback loop to bring the targeting back together.
Ella Darlington: That's it. I've seen a few people asking what GEO is, and maybe some people started a little bit later, generative engine optimization, right? But Paul, what's your one piece of salient advice?
Paul Stollery: Well, I wasn't going to plug the white paper, but as that's the question, numbers.co.uk to see our garden generative engine optimization. No, seriously, it is kind of linked to that. So, the tip is: do the research, right? So, if you're working in-house or even at an agency, there's nothing stopping you from sitting down and thinking, what types of questions might people be asking about my brands on these platforms? You can log into like six to eight of them and have a sense of what's driving the answers on them. You can even go one step further and do that systematically. It sounds really technical, but it's just a plug-in on Google Sheets. You can set up a spreadsheet and have 100 prompts and get them all out of various different platforms. That's what we did. A year ago, all of our clients were asking us, "Why is ChatGPT saying that about our thing?" And we're like, oh, we don't know; it's probably similar to search, but we don't really know. So, we developed that brand framework, we used that plugin called GBT for Sheets, and then we looked at it across a multitude of answers. We were able to get a sense of the answers. These things are black boxes, but that doesn't mean you can't do a little bit of research to figure out what the general trends are. Not to paraphrase Fox Mulder, but the answers are out there if you just sit down with a spreadsheet and nerd out for a little bit.
Ella Darlington: Exactly. Right, we're going to go straight to questions. We've got some really interesting ones, really meaty ones, especially about industry shifts. One is around newswires. So has AI got into newswires in the US, the UK? I think people aren't really seeing much. How important is that? I think someone's being really specific here, talking about investor relations. But has AI got into the newswires? I don't know if people have got a perspective on that. And then another question we had was more around local news and publishers: how AI is being used. Is it being used in journalism? Is it being used to help sift through press releases? And are there any considerations behind that? I'll throw those questions out to the group, whoever wants to go first.
Philip Hodge: I can pick up a little bit on the financial side of things. Regulatory news that's produced is critical because obviously a press release stating someone's financial results can have huge impacts on the way the market comes through and on the buying and selling of shares. So that element is critical to how that content is being provided. Now, is AI being used to write some of those press releases? Ultimately, at the moment, no. The way I'm seeing the conversations with investor relations officers and across financial teams, they are still very close to that human approach of making sure their information and content is written by themselves. Are they using AI to help structure some of those things? Yes, that is taking place. But the critical nature of AI being able to affect share price is still not being taken on by a lot of the IROs. The summarization of some of those releases, because some of those documents can be 20–30 pages long with multiple tables and forms, is something we’re seeing across multiple platforms. As Paul was saying, same information, same content, but the summary is being picked up in different ways by different agents - Bloomberg versus Reuters versus Investigate, they're being picked up by different applications differently. So being able to structure your release and utilize that, GEO is hugely important because the summarization of your content, that first read by the AI bots, can be more critical than you really think. So yeah, that financial side is a huge part of what we're looking at.
Ella Darlington: And then any take on AI being involved in filtering or journalism?
Hina Pandya: I just wanted to address the bit about the newswires, because while the UK doesn't specifically target newswires, I mean, everything's global, right? So, like Philip talking about Bloomberg, we have our UK office here, right? So, we're still going to see that North American newswire, and we still follow up on all those stories. So, we still get them, we're still looking at all those things, even though they're targeted at a different part of the globe. But, you know, it's kind of ubiquitous. And I just feel like investors will probably put money in those kinds of things because they know it's going to hit everyone.
Ella Darlington: Yeah, absolutely.
Paul Stollery: I've been working on something on the filtering front. So we've been trying to build a little thing where you put the press release in and you put the master media list in, and then AI will check, based on the last 12 months of stories written by each journalist, how similar the content is to the journalist, not predicting whether they're likely to cover it. But the initial finding was it's rubbish at doing that. I thought to myself, that seems like something that's well within its remit, right? Because, you know, in all seriousness, we were using APIs that had access to search. So, in theory, you could do a little search, put all the content in, and on a scale of 1 to 10, this is how similar it is to the last 12 months. But it was really bad. And for a really funny reason: the LLM platforms are really eager to please. And I hate talking about it this way, but as you talk about like a person, if you ask for feedback on your writing, it'll probably tell you it's good. If you say be critical, it'll probably tell you it's bad. And what we found in doing this was it either gave all of the scores like 8 or 10 out of 10. And this was like we were testing on an acquisition for a legal firm. I had BBC journalists there, and it was like 8 out of 10, it will cover that. Or it was overly critical. And if you apply a critical eye, it made all of them threes and fours. Now that's fixable with a bit of smart thinking, and I'm sure people have done a better job of solving that, and I'm sure we'll be able to figure that out. But in filtering a lot, there's a lot of stuff which is brilliant at, and a lot of stuff it's like, I really thought you'd be good at that, but it's surprisingly disappointing.
Ella Darlington: Absolutely. And just, I think somebody asked the question around AI being involved in journalism. I was speaking on a briefing call to some of our panelists in recent days, and sadly, when I was pitching a very local story, I spotted the use of AI journalists, which I don't think is a widespread thing, but I think it's become something that folks have relied on. And that's largely due to capacity, really. So, it is sad, but it's not a reality everywhere. But I think journalists have to do more with less. Two questions together: one, should there be a disclaimer on press releases stating that AI has been used to either write it or research it? And second, how can we as PR practitioners use GEO and the way AI search looks for credible sources of information to show the value of PR in an industry that's always struggled to talk about and quantify ROI? So that's talking about using the functionality of some of these platforms to help with enhancing perhaps written content. We were talking about earned and online, right? So, this is probably more of the online, but also maybe to help report on quantifying our ROI, easy for me to say and value. But the first question there was around disclaimers.
Paul Stollery: Yeah.
Ella Darlington: There's so much to be said, right?
Hina Pandya: There's just a lot to be said. Yeah, I think I always put a disclaimer, or I say no AI was used in the writing of this post because, like, I'm just going to be really honest here: if you couldn't be bothered to write it, why should I be bothered to read it? You open a magazine, or you buy a magazine or a newspaper, and you're expecting that human beings went and did the research, or did the race, or did whatever, you know, you expect them to show up at a war zone, they wrote about it. You expect that. If AI is, somebody's asked what screams AI-generated, it’s when people use, when you see long words in there, or really strange syntax, or you're just like, this makes absolutely no sense. This is really great word soup, but it makes no sense. That kind of thing is rife. And so I think ethically, if you can claim to say, look, this part is AI, don’t make it all AI, because I don’t want to read all AI, neither do you. So, if you can like Philip was saying about the images or whatever, if that's come up from AI, say so. Then I’ll know what I’m using is AI-generated, and I can tell my reader that as well. Then there’s a trust relationship.
Paul Stollery: Yeah, I think there should be, but realistically, I don’t think there ever will be. And I think the comparison, the point of comparison I always go to, is if, like, let’s say NatWest gets hacked and you see a comment from the CEO in the FT, the CEO didn’t write that comment, right? Their PR manager wrote that comment, or their head of PR wrote that comment, and the CEO signed it off. Now, purely from a moral point of view, I think if it’s signed off by the CEO, I don’t think it’s that different to who originally wrote it. The bit where I think, but the thing is, the most obvious thing to do would just be to say this is representative, the CEO’s point of view drafted by AI. But if you did that, it’s never going to get covered, right? So, we’re in this really weird situation where I think Hina’s point of view is objectively correct and we should absolutely do it. But if we’re all being honest with ourselves, there will always be those forces of pressure to move faster, with lower budget, and achieve more. That will always exist in a private company.
Hina Pandya: But to counter that
Paul Stollery: So yeah, I think there should be, but I think we’re probably going to have this really weird couple of years.
Hina Pandya: I mean, you’re probably right. There’s going to be a couple of years where people are going to think they can cut costs, and their brands are going to suffer for it. Because if a CEO for a big bank like NatWest would sign off on an AI-generated quote, and they are willing to stand behind that because they’ve signed off on it, but AI has generated it, is what I think you said, then there are possible ramifications for that down the line. And I think what we’ll see, like you say, in the next two years, is how that blows up, and how people are like, let’s not do that. Let’s pay for a proper person to write this stuff.
Paul Stollery: That’s the big thing. And I’m not just saying this because I’m in front of a journalist—we don’t do that for the risk. Because, again, as I said at the top of the call, there’s a huge reputational risk for it. But genuinely, because I’m… There are a lot of companies out there who outwardly talk about their use of AI and efficiencies, and whenever I know a PR manager there, I’m like, what are you actually doing? And oftentimes they won’t, for that same reason. But I do know some who are not listed but are in that top-tier tech company space, who do it. And sometimes their CEOs have talked about the use of AI, just not specifically said that thing. So, there’s definitely so much to be said on this.
Ella Darlington: There's so much to be said on this. But I think one word that we'll all just take with us, regardless of whether there's a slight hint of AI involvement, is ultimately integrity. That being at the core of everything. We are exactly at time. I know for a fact that every single one of us could talk for more and answer so many more of these fabulous questions around this topic because we've only really just scratched the surface, and there's just brilliant expertise on this call. So, I just want to take a minute to say a huge, huge thank you to each and every one of you, Hina, Paul, Philip, for taking part. There's a very quick, easy poll that should just take a few seconds for you to click through to provide feedback so we can keep serving you with brilliant content and brilliant panelists. Also, thank you all of you for attending. Please look out for the notes following from the CIPR team. There'll be links to where you can access this and start and continue discussions with your colleagues. But just a huge thank you to my fellow panelists. You've been absolutely brilliant. Thanks everyone for being so vocal in the chat, interacting, and sharing tips with each other. I can't wait for the next one. Thank you all so much and hopefully speak with some of you soon and see some of you in person, I hope.
Hina Pandya: Yes, that'd be lovely.
Paul Stollery: Thanks ever so much. And dear Jonathan. All right, take care. Bye.
Hina Pandya: Bye.
About Notified
We are Notified, and your story goes here. As the only technology partner dedicated to both investor relations and public relations professionals, we help you control and amplify your corporate narrative. Our fully integrated PR and IR platforms streamline every step—whether it's reaching the right media, press release distribution, and measurement or designing new IR websites, managing investor days, earnings releases, and regulatory filings. Connecting both worlds, GlobeNewswire is one of the world's largest and most trusted newswire distribution networks, serving leading organizations for over 30 years. Together, we empower communicators to inform a better world.
Notified is an affiliate of Equiniti Group Limited (EQ)
A New Chapter for Capital Markets Infrastructure
What’s New in Content OS: Helping PR Teams Win in the Answer Engine Economy
