#30 Harnessing AI in Digital PR with Alex Cassidy

Stella:
The Digital PR Podcast is brought to you by Coveragebook, the tool that is helping thousands of PR teams around the world spend less time and effort on reporting so they can put their energy back into the more important parts of PR. Coveragebook replaces the manual tasks of collating screenshots, sourcing metrics and engagement rates, and then designing and presenting into a report. We cut out the need to use PowerPoint slides, Canva, or Docs. But it’s not just about glossy reports. Many people come for the metrics, but then they stay for our estimated coverage views figures. That’s right. We predict how many people may see your article or social post. We don’t just tell you the total amount of people that happen to go to a publisher site or an influencers feed. We really are about helping PR professionals be taken seriously at the important time of measuring success. So why don’t you give it a try? Test the metrics and see how much time your team can save. Go to coveragebook.com for your free trial.

Steve:
In the last year, you’ve barely been able to move for another AI hot take or think piece. Is it going to steal our jobs or make our lives exponentially better? The jury still seems to be out. What we do know is that you’d be mad to not be at least trying out AI in all its various forms to see how it can change or even improve your work. Someone who has done this recently with his team is Alex Cassidy from Distinctly. So we’ve invited him on to chat more about how it’s gone and what it means for our industry. Hi Alex, thank you for joining us. Thank you very much, appreciate it. It’s a first, a first for this podcast that we have someone in the studio with us, which is immensely exciting.

Alex:
Yeah, honoured to be the first person to do that.

Louise:
You made the train journey now. For the listeners, could you give us a little overview of you and your career?

Alex:
I’m Alex. I’ve worked in digital PR for around eight years now. Prior to that, I worked as a sports journalist. I wrote a book about American football and then very much was like, I’m not gonna make any money doing this ever. So moved over to marketing. Started off at Verve Search and then was the head of outreach there until about 2020. Briefly in-house at RVU and then I’ve been at Honcho and now currently at Distinctly for the last two and a bit years.

Steve:
We mentioned how recently you’ve been writing blog posts and white papers, which we’ve both read, like Lou was quite vocal about, like, you should read this, really, really interesting, share with the team, etc. All on using AI in digital PR, which is a fascinating topic, which is why we’re here. Could you give a little bit more detail as to sort of why you did that and how you’re using kind of AI in digital PR currently?

Alex:
Yeah, so sort of touches on what you were saying in the beginning, really, that The idea of being mad not to at least test some parts of it. And I think one of the things that spurned us to organise it a bit better within our team was the fact that I knew that I was testing it out myself as an individual. I knew that other people in the team were doing that sort of solo, just to see what it was about and then I read some stats that were saying that a lot of people were afraid and this was maybe a year and a half ago and now it’s a little bit more standard where everyone’s kind of using it or playing around with it but there was a stat that was showing how scared a lot of teams are, a lot of individuals are to tell their bosses or to tell their clients or to tell anybody that they’re utilising this for like work which makes sense because it could be seen as making your job redundant at some point. It’s not going to do that which is a spoiler for the white paper, but as a result, I said, well, if this is going to be, people are going to be playing around with it, we might as well try and organise it and get people to use it properly, and also to try to take away that stigma within our team to allow them to test it freely without having to resort to doing anything kind of cloak and dagger style. So that was what prompted us to initially start a task force, which we call it, within our team to understand, well, if this is here, and it doesn’t look like it’s going away, it’s going to be sticking around somewhat, how can we ensure that the team is using it effectively, that we’re testing it properly, and also that we’re not doing anything that could get us in trouble, right? Like, if someone’s using it and they don’t fully know how it works, and it is quite a wild tool, which people can also misunderstand as to what it can do, then we need to avoid that as much as possible. So that’s what prompted the initial, let’s set up a team to sort of explore. And when we talk about AI, what we’re really talking about is generative AI in the sense of tools like ChatGPT, Gemini, and also the image equivalents of like MidJourney and DALL-E and stuff. Because AI is such a broad church and is this big umbrella that can include very extremely complicated things that you wouldn’t have the keys to, and things like generative AI, which is a little bit more accessible. So that’s what we mean when we say AI and digital PR.

Steve:
Yeah. That makes perfect sense. When you started this, did you find that within your team or within the industry generally, there are people that are kind of early adopters. So there are people within our team that were very much like, I’m using it, I’m into it, I’m getting a lot of value from it. And then other people, and I would categorise myself in here, it was a little bit more like Luddite. A bit Luddite. I don’t know what I’m doing. And like the first time I used mid journey, I remember playing around with it and for some reason I asked it to create an image of Chris Tarrant playing tennis. Zero idea why, why that popped into my head and I was amazed by the images but you know it was one of those where that came up with like seven fingers and like really weird. He does, he does. That’s a world first reveal on the pod. I don’t know if we’ll get that past legal actually. Tarrant has seven fingers. But yeah, is that the case? You found this early adopters and then sort of the more kind of, I guess for want of a better phrase, the Luddite end of the scale, I don’t know what I’m doing. So you kind of got to work quite hard to bring everyone along on the journey, right?

Alex:
Yeah certainly we had early adopters in our team for some of it and I know that I was sharing blogs when some of the really early really like shaky mid-journey stuff was coming out and just testing it on and just sharing kind of campaign sort of like half campaign ideas of how you do things. I had a lot of people message me sort of saying like how is this tool working etc. Yeah, I don’t know if there was necessarily a certain personality type that tested it earlier, necessarily. We sort of found out afterwards. So one of the things we did when we started task force was to like survey the team at the initial stages to see what they were au fait with and what they weren’t. And also what they thought it could do versus what it couldn’t do. And we found that there was a bit of difference between people who had tested it and also surprised that more people were using it than we necessarily anticipated, or at least had used it once or twice. But because I think the use cases of it are still quite hard to know, we asked the team, what would you like it to be able to solve as a problem? And this is a team that predominantly is digital PR, sending emails, sort of like tip of the spear of doing the work. And it was always like list building, which came back because it’s the most time intensive thing, it’s so long. And it doesn’t solve that problem, really, although there are some tests that have been done with it. But for our purposes, it didn’t solve that problem. So when we asked them, what would you rather do? It was like, well, it doesn’t do the thing we want it to do that saves us most time. So actually, a lot of the other issues we had were learning how to do things is a very time intensive thing. And the people who have the least amount of time often are the more junior members of your team, who are doing like the work, the billable work a lot of the time in terms of sending emails, which is very time intensive. And it should be up to more of the seniors, theoretically, who have a bit more experience as well to work out how something can be applied to the work. You have to work together because I find as a head of digital PR that the further I’ve got away from doing the actual work in my career, how more often I have to check in with the team to understand what tactics still are relevant. The amount of times I would be sort of saying, well, why don’t you do it this way? And I just sound like a hundred years old.

Louise:
Do you remember though, probably when we were all starting out, when people were like, why aren’t you picking up the phone? And it’s like, have you tried calling anyone recently? I do worry. I’m like, am I going to be saying the equivalent of that? So you have to really kind of ground yourself in some work or some real experience sometimes.

Alex:
I want to think that there are some things that should be evergreen in it and there probably are some but ultimately it’s also I think a little bit of you forget how long it is to be sending emails and to be sort of on that grind of sort of chasing things. I know it’s not like a like a coalface job necessarily but it is it can be quite crushing like in terms of like just sending emails into the void and so I’m aware of their time as well so it was it was a difficult balance to be like you guys don’t have much time so we’re not asking you to test out loads of things so we had to sort of work out how can we ensure that we’re planning this properly so that we have the right people testing things but without adding because early on we found that asking them to utilise a tool to do something they could already do would take longer because you have to learn how to use the tool. Yeah. So maybe there’s a time save marginally in the future, but those marginal gains will come at the, in the short term expense of testing this out. And we are fully, like if we’re fully booked for like the next month or two months, can we afford to do that right now? Not really. And the team will be the first, I can basically be like, yeah, it’d be great if you could test this out on that. And then in the reality, they’re like, well, not really.

Louise:
So I was aware of that too.

Alex:
There’s a few things, concessions we had to make with that. And we found that a lot of what it can do well were things like subject lines. So again, when I first started out, one of the instructions to me was to like, you know, write 25 to 30 subject lines differently. Just like, just take them and work out how to say it best. And it is a skill which you then can learn how to make that short and snappy and appealing and appear in a journalist’s inbox better. But that is something which generative AI can do very well, to rewrite things and suggest other things. One of the first things we did was we created a tool called Headline Grabber, which enables you to type in sort of what your story is, have a drop down menu that would say, do this for tabloid, broadsheet, whatever. And then it would be, you know, I think it was like serious or kind of zany, things like this in terms of style, or clickbaity was one of them and it would generate like 25 headlines for you based off of that. It was just like a very simple tool that we made using the back end of ChatGPT. All that is, is just a skin on ChatGPT, because if you ask ChatGPT that same thing, it will do it. And all we did was just sort of like create a front end for it. But it was useful. I think the team use it in doses and probably still can think of the thing is you don’t necessarily need to do 25 to 30 once you understand how to do something quickly you can do five and we also encourage the teams like vote in slack on which one’s better and you can knowledge sharing is good in that way yeah so it’s not every time we use that but we did find it was a a use for it was there anything else that you found particularly good and then also something where you’re like this is rubbish for this yeah there’s I think there were probably more things that were left on the scrap heap where it just didn’t do the job properly. The key thing to remember with large language models is that they can’t think for themselves. They are just creating what they think should be next based off their training data, but they can’t really laterally think about how to answer the question. So when you’re saying to them, for example, coming up with ideas is a good one, because I think people do use it to sort of have a brainstorm and a back and forth with, which is fine. And it can be very critical of the idea that you send, which could be useful. But by design, it won’t necessarily create something original because the training data it’s used is based off of things which have already been produced, which digital PR arguably is doing that anyway. In terms of like, I might still be an idea repeating machine because I’m like a product of the industry and sometimes you just are like, can I borrow this idea and change the adjective and it’s a new thing. Not to say that it doesn’t happen, but I do think that it would just end up making even more of a loop in terms of the same type of thing in digital PR, which I think would overall end up being more of a problem because you’re just going to keep poisoning the well of what we’re doing and there’s not going to be anything fresh or new that we can come up with. I think it’s true that a lot of agencies, including ourselves in this, can sometimes do ideas formulaically, like that does happen. But this is like designed to do that. So it’s a bit different than just being like, because I can recognise that behaviour and try and change it and try and become, be more original. We can feed that back among our team. The tool won’t do that, because if you’re asking it to try and think of something that hasn’t existed yet, it would struggle to do that. Which we found to be the case, and as a result, I did a test where I wanted to see if I could use ChatGPT to come up with an idea, produce the idea by coding it and then also write the email. How much of it could it do with prompting? And I couldn’t get an idea out of it that I’d be able to produce that I thought could actually get links effectively. I probably could have but it would have been a bit roped and it wouldn’t have been a very interesting blog necessarily. So I was trying to get something which would be a bit more, I used it to sort of help out with me realising the idea as opposed to coming up with the premise and we’re now on the final stage of that blog where we’re outreaching it and we’ve had a bit of feedback from journalists so we should get a link which should be nice for the arc of it. But yeah, ideas are definitely one which I don’t recommend it to do unless you want something formulaic in which case you’d be brilliant at it. And then it can do a lot of the writing of the emails, it can certainly pass emails down and sort of cut them down, it can be a good editor in some ways, but I still would, I think the important thing which we miss with that, which is sort of a story of ChatGPT full stop, is that writing can be thinking in that sense, where the more the team understand the idea, I mean you might have seen before if you have an outreach person who’s come up with the idea, helped do the idea from a methodology perspective, or at least consulted on it, and then written the email and then sent it, there’s not just, I think it’s a lot more conducive to better results because that person not only feels more invested in the entire process, I think they’ve also had a chance to really understand it better, so could understand how to pivot it or understand how to email it differently. As opposed to, you know, it happens sometimes, but, you know, giving someone an idea and being like, go and get links for it. You could do that, but it isn’t the same type of thing. I think what we miss from having the team just, like, not writing the email as much is that they’re also not going to understand the campaign enough and probably not also going to care enough about the results, I think, in the end. It probably could do that, but we didn’t want it to because of what you would lose, I think, in the process of the job.

Steve:
That’s such a good point because there’s quite a lot of, again, for want of a better phrase, quite a lot of churn in not just digital PR, but kind of marketing generally. It’s like onto the next campaign. If you lose that step of creating things yourself, fully understanding it, then it only contributes to that. Then it’s almost just like it’s done because I’ve got this help and it might make it a bit quicker in some ways, but then it’s, Yeah, you’re not as interested or invested. The way we talked about Propellernet is it’s a really good starting point, like ChatGPT is a really good starting point for like giving you info ahead of ideas sessions, sort of about an industry, about what’s topical, you know, it kind of gathers information quite quickly, quite readily rather than doing your own desk research, trawling through Google, etc. And it’s also quite a good finishing point, as you say, for like checking things or like, oh, I need to organise all of this data. If it’s a data story, can you tell me the best way to do that? So everyone’s going to think about it differently and it’s advancing all the time, right? You’ve mentioned a few use cases here, like what it’s good at, what it’s not so good at. For you personally, what do you use it for in your day to day? And the reason we ask this and me personally asking this is because I love it. I’m fascinated by it, but I am one of those people who I’m almost sort of struggling to cement it within my day to day. I think it will come, but like I almost forget sometimes that it’s there, like readily available. So it’s like, oh yeah, I could have just done it this way, but I’m so in the routine of stuff. How do you, how have you overcome that? How do you use it in your day to day?

Alex:
Yeah, I could look at my ChatGPT history to see some of the stuff that I’ve used it for, but I think it tends to be things that I am replacing some Google searches I think with it. The good example is we were doing a bit of ideation around travel and we were talking about expressions and idioms that exist in other languages that might not exist in English classics. space to play in for an idea’s perspective. So for that, that to me is a very easy thing to ask GPT to generate in lieu of a search, right? Because a search might return similar results, but it’s probably going to be a bit more disorganised just to find on the page, I don’t know what the website’s going to be. And is it going to be as thorough? I can ask it to keep generating them, it will hallucinate someone like make some up I don’t think like I would sense check it against somebody from those countries like if I say can you can you give me 50 Spanish idioms it may well be like well I’m just gonna start translating English ones in spark it could do things like that yeah the problem is I don’t trust it for a lot of that stuff but things like that I tend to do things which probably could be searched on Google and found in some resource because it’s it is taking them from there but it could also take it from its um original source too. I’ll be honest, a lot of it is just that. I also use it to code stuff. That’s probably the thing I use it the most for. Simply because, I don’t know, I don’t need it to do a lot of the writing stuff for me personally. But the coding has been useful. So I got it to code my personal website, for example. which is really simple and plain. It’s just HTML and CSS. But I got to do that and I didn’t have one before. People set up their own personal website, it’s like, I’m going WordPress and I’m working out which template to get, and I’m going through all of that. It’s really easy because it’s all front and it’s like, is it though? I think having something which I could just copy and paste straight in, and it is a bit like Year 8 IT make your own website level in some ways, but that’s okay because it’s live and done. It doesn’t cost me anything comparably as well in terms of like hosting, it’s very very easy. So I’ve got to do that and I’ve got to do to code a lot of the tests we’re doing with trying to see if it can create a campaign. So one of the things I would love for it to be able to do better is that when I was starting off in the industry in 2016-2017, Verve, we used to make campaigns that were a bit more, well they were very, very in depth in terms of like scale. Probably like sledgehammer to crack a nut in some situations, which we found, which is why we’re now, as times gone, and a lot of agencies did it that way, it was like bigger, interactive campaigns, we had like five designers, six devs working on this output. It was brilliant, because a lot of the process of making the campaign was also a lot of fun, and over time, I think the industry’s matured and realised you don’t need to make this big thing to get the headline which will still get the coverage. So you’ve sort of shortened the creation part of the job. And I get it as a result because budgets have also shrunk in that time because there’s more competition. You can’t separate all of those things from happening, they happen together. One of the criticisms of ChatGPT as a company and as a tool is that it’s massively overvalued, which is probably true. Like they say, a good journalist who covers it is a guy called Ed Zitron, and he talks about how it’s like a trillion dollar company and it doesn’t solve a trillion dollar problem. Which is, yeah, because it doesn’t really, right now from what I can understand. A lot of it is promises and it’s going to be able to do this and what it could do in the future and they’re constantly raising money for it. I’m not, that’s not like my, I’m not an expert in that, but I recommend reading him because he is very, very critical of the company and really knows his stuff. For me, and what we do, we don’t need it to solve a trillion dollar problem. Could it just solve a 15 grand one, and it could help split, narrow the gulf between what we used to be able to create on certain budgets versus what we can do now with what is the reality of the industry. And that would be great. I think it would mean that we can create things that are a bit more interesting without it just necessarily being like a digital PR turning into what can become like a headline generating industry where it’s just like can we just get the headline, sometimes people aren’t even necessarily putting it on blogs anymore, it’s just like exists in the press release and in the final article and it doesn’t exist anywhere else. Which, you know, is a consequence a lot of time of budget. But can we start making things again would be pretty cool, which is what some of the tests I’ve done have been. So we tested it, like getting it to code. I do also consult with a dev who helps me like ensure this thing would be correct. And none of this is for clients right now. It’s all just my own testing. And it would require, I think, an expertise to ensure that it’s secure, that it’s done properly, etc. And also, I think you still need human input and guidance. So, we wanted to see, for example, a classic PR story, which is which first names are the most common in songs, right? So, loads of songs like Georgia, you know, is a classic one which gets talked about a lot. There’s loads. Jenny is another. As I said that, I was like, it’s a classic. I was like, I don’t know any. So that does exist as a campaign and people have sort of totted up and counted it so I said okay what if we use it Spotify API to generate those results but also we used it to create a tool which you could search your name and then have it appear with all of them and then you can suddenly create a playlist which would be an insane playlist to listen to because it would just be like all disparate types of like music. It’s just literally, it’s every song with Alex as a name. So it did all of that. It made the tool, I put it online, it’s called Bortify, which is based off of, this is Simpson’s reference, where there’s a kid called Bort.

Steve:
I remember Bort.

Alex:
My son is also named Bort. So, because it’s a first name thing, we did that. It’s live and it works. What we’ve done is we’ve used it to also get the data, which you can use for the story, because the story there is here are the most popular names, male and female. You could also use it, I guess, reactively in terms of, you could use pet names probably and all sorts of different things, or famous celebrities. new baby, etc. But the key thing is, I think if we were doing that for a client, typically now, it’s a budget, you wouldn’t make something, you just would get the data. And then you would publish it on a blog with like here on it. But I think making something there can add something could increase the chance of getting links. And it’s also just a bit more fun, I think, as an actual tool. Yeah, that’s the thing we’re sending out now to see if we can get links for it. It was part of the test to see if ChatGPT could create everything for that as a process, and it did, and it works pretty well, which is cool. So that would be the thing I’d like it to be able to eventually enable us to do.

Louise:
So would you say that actually you enjoy using it more to broaden what you can do with your job and things like that rather than taking parts of your job and doing it for you? So you said like writing, for example, you don’t necessarily need or want that. You just hear a lot of negativity around like AI and about how it will take jobs or how, you know, people are using it for writing and the writing’s awful and things like that. But actually it’s a nice way of framing it to think like there’s a part of this industry which for many reasons has gone and I would like to bring it back in a way that’s going to work for the now. And that’s nice to be able to use AI. I hadn’t really thought about it like that. And it’s actually a much more positive story and opens up a lot of different avenues that you could go down with it.

Alex:
I think importantly though, there were devs that used to do that work at Verve, right? And I think that’s why it’s important still to have it as an enhancer. So one of the things that we developed off the back of the white paper was can we essentially create a, not manifesto, but like a set of rules for us at Distinctly, like a statement of this is how we’re going to use generative AI it came down to we want to use it to enhance what we can already do as humans, as opposed to have it replace things, like you say, because I don’t think it actually is good enough to do that anyway. And a lot of what they said is they need more training data for that to happen, which they can’t get. So one of the theories is it will start training on itself, which will just become a kind of, it will make it even worse, and it will never really be good enough in that sense. And one of the other things we found with it was that it also can’t be trusted with like client data or things like that. So it has to, like you say, have a human element to all of it, it has to be quite strict of how we use it. And we do want it to be an enhancer of what we can already do versus using it wholesale to just create facsimiles of what our work used to be essentially. Another thing we found is that some people have used it, I think part of the reason they were saying like Harrow recently shut down or has moved or changed was because people were using it to create comment wholesale and send that out.

Steve:
Cheat the system basically.

Alex:
Yeah, and I mean, it’s an interesting one because it’s almost a further step. You could argue that there’s a lot of PR agencies that have essentially been writing comment on behalf of the experts for a long time anyway. And that’s just sort of a more acceptable like brown paper bag for the thing that was still happening. And somebody’s just taking a step further and there’s not even a, I guess they could say it was signed off subsequently by an expert, so that’s fine. I get it. It’s not exactly the same. It isn’t like we’ve gone from only experts providing the advice to now an AI doing it. There are a few murky steps in between which has sort of enabled it to be a slipperier slope, I think. It certainly was a problem. I think it was happening at such a scale and could happen so rapidly because it doesn’t require much effort to kind of just basically get it to generate a comment on anything that you have. Which is why I think reactive and sort of personality or utilising personalities in reactive is going to become so much more important. Because you need, I think journalists are worried about them using experts that don’t exist. So they’re going to need to be like, is this person real? Can I actually rely on them for comment?

Louise:
Check them out a bit better. Exactly, yeah.

Alex:
And it just seems really short-sighted, which I think it is by the people doing that, because one of the things that we rely on as digital PRs in our industry is a press which is strong and robust and sites that don’t go under because of like any kind of reasons even if it’s content churn and content being stolen or and sort of losing positions in the SERPs or being dominated by a few smaller publishers like it affects us as well if that is not robust and isn’t strong and large the larger the press the better are the easier our job is yeah in that sense and I think anything that can be seen to jeopardise that or cause it to be weaker, just eventually will roll downhill to us and we won’t be able to secure enough links or there won’t be enough quality links to get through. If online publications are just using AI to produce content, then who’s the person we pitch to in the first place? Why are we then needed as a part of that process? It’s very interesting dynamics if we don’t understand how this can be segmented and how it can be organised between us and AI tool separate and journalists and the companies that they work for. Like it’s a tricky balance I think ecosystem wise.

Steve:
So very tricky balance we’re touching on the a lot there on the risk that AI poses to digital PR and what I find is the the kind of life cycle of these things is people tend to see something and certainly on social media etc they panic and go like well this is This is going to end like these jobs. And, you know, I think the Telegraph had a think piece where it was like, these are the jobs that are under no threat from AI. And these are all the ones that are. There is, as we touched upon, there is a very real threat. As an example, one of my best friends who lives out in Argentina is a copywriter and has been for many, many years. And he’s seeing a lot of his work erode because of AI. But sticking with digital PR, and we have touched on it already, but do you think, it doesn’t sound like you do, but do you think there is a risk that AI is going to pose to digital PR and those sort of traditional digital PR roles.

Alex:
It’s very, very hard to be able to predict all of that. And I think by nature of my position, I’m not necessarily going to be like fully, I guess I have to be a little bit optimistic about it. I think that the biggest risk for it is that AI affects the overall ecosystem around it. Like I said, like that could be the biggest risk to me. I don’t think it is in a position. It isn’t good enough to be able to do what we do and have the human element of it. And it’s not automated enough yet in terms of the sending out of emails. I think you still need to be more bespoke to get journalists attention that way. Plus the general awareness of it being a thing means that people love to flag and flag it and say like, oh, this is Gen AI. This is ChatGPT. This is clearly like a bad thing. We shouldn’t be doing that. People love to flag it on LinkedIn, flag it on when they see a blog, all of that stuff. So I think there will be more of a demand for can you verify your sources? Is this actually a good piece of work that’s being produced by journalists? That’s my hope. But it could just end up as being like that we just have to like abandon certain areas because like Gen AI slop has just become so so embedded within what it is. I don’t know when this is going to be going out but like today and yesterday has been an even bigger Twitter exodus in terms of because people are A lot of complaints I’ve seen about it being just like Gen AI just like fully taking over the timelines, which is interesting, not just all of the other issues. It seems to be that’s one which people are particularly irritated by because it’s annoying.

Steve:
Well, yeah, one of the comments I saw, we were actually talking about it over lunch earlier, because I think you have, well, it’s defected the right word, it’s not, is it? But you’ve gone to BlueSky. I set up an account today, actually. I’m going to do one today as well, so we should give out our handles later. But yeah, one of the reasons that I would want to move is that I used to find the comments quite interesting like on certain stories and particularly in our industry, but it does feel like when you click on the comments that almost everything is just not related at all to that post and that really bugs me because it’s like it’s pointless.

Louise:
That is just slop.

Steve:
Yeah, it’s like a Temu ad or something random like, oh, you’ll never believe what happened here. Number three is wild and you’re like, I kind of want to look. It’s annoying. I would agree with you, Alex. It’s difficult to predict, but I’m optimistic. I genuinely don’t think it’s going to take over our jobs in the way that people get scared about, but it’s definitely already and will continue to affect the ecosystem in a way. It’s going to be intriguing to see how that develops, but we try and see it as like a an enhancer of what we do. How can we enhance what we’re doing in the same way as you do? And I think for anyone listening, that’s probably the best way to think about it and be positive and try and use it in that way.

Alex:
So much of what we do at agency level involves a human element in terms of just negotiating with clients. I’m not talking about like the work going to journalists, which does as well, but also just like the work required to run an agency or run teams is so human centric. that it’d be very difficult for me to see that if you could automate some of the more fundamental parts of this job in terms of that. And I think, again, also the importance in digital PR of having somebody to blame, I think, when something isn’t going right, and also having the trust from somebody that they’re going to pivot it and be able to do that proactively, is 99% of what we deal with in terms of, like, the communication, I think, of an ongoing campaign. Yes, sure, sometimes you’re doing it, it absolutely smashes every time. And you know, a lot of the campaigns we do distinctly do that exact thing. But I think that realistically, there are always going to be situations where you’re going to have to speak to somebody and say, yeah, we’re on it, we’re trying to fix it. And so that’s where I think the agencies that utilise generative AI too much and rely on it too much, they probably weren’t doing those type of things effectively before anyway. So I imagine it’s taking a sort of section of the market that a lot of agencies might not even be concerned with necessarily, because it would be kind of an edge case, I think.

Louise:
We mentioned about AI slop and people getting dissatisfied with traditional platforms. People might be tempted to start using ChatGPT because it has its slightly more newer search function, starting to challenge Google perhaps if people aren’t satisfied with that anymore. There are already lots of questions around, OK, well, how do I appear in ChatGPT search in the same way that I’ve been paying you, SEO agency, to help me appear in Google search? How does digital PR affect the search results in ChatGPT? Is there a link there with how you can be appearing as a brand more so in that platform?

Alex:
My understanding of this, and I think they haven’t been very transparent in terms of how it works, is that it’s just searching a version of the internet anyway. So if you are trying to affect that right now, the best thing to do is just create content or try to get links to increase the rankings of that content for the queries, it would be. Because it’s just reproducing it, I think, via Bing rather than via Google. In that sense, a lot of what we’re doing does not change because it’s still the same type of outcome. If you’re talking about purely from like digital PR to affect their search perspective, it shouldn’t change any of the tactics. And I think we’ve been looking at how you can actually track whether or not you’re getting referrals from ChatGPT as well. I think that they’re only going to improve that ability, I think, as they try to, like you say, take more of that market. But most things we’re doing at the moment is like, okay, there probably be more of an on page considerations in terms of that, like trying to create content that can be both on page and off page, like effective, which a lot of what we do anyway, in terms of that, I think the type of searches will start to understand more as we understand user behaviour, in terms of what people might search in ChatGPT versus what they might search in Google and how those two things work together, which I think is some of the stuff which people talk about now, with TikTok, for example, which isn’t I’m not 100% up on, but it’s that two different types of searches happen, and they sometimes are related. Someone will check something on TikTok and then subsequently finish the search in Google. I think that’s probably some of what you’ll see in chat GPT. The question is whether or not they can, the trustworthy element of chat GPT, or if people understand that it can hallucinate and can reproduce things that don’t exist. You can test that yourself by asking it. It will never say it can’t do something because it’s just trying to give you a result that will be trying to answer the question every time, which is where hallucinations can come. I don’t think people necessarily ever understand that, which will cause an issue, I think, in time, because people are like, well, why is it saying something which isn’t the case? Like, I’ve asked it for a list of books on things which I’ve written books about, and it’s given, like, ten results which are just not none of it is me. I’m the only person who has written a book on a specific subject and it’s in like here’s five different books by five different authors, none of which are real. It’s just made them up fully.

Steve:
It blows my mind that how it can hallucinate because as you say there’s a purpose behind it because it is trying to answer that question but it it does kind of, I think that’s quite scary. Then you get into the idea of deep fakes and things like that. The sort of the idea you can just make stuff up that people could without the correct training or knowledge just thinks real and start quoting that like oh there’s been this book or this thing I saw someone say this and that’s the bit that really scares me actually if I’m honest.

Alex:
Yeah a lot of people who’ve used it for dissertations and things like that thinking it’s giving you Like, even just, if it’s giving you a real book name, at a minimum, would be like, I’ll reference this in it, it just does not exist.

Louise:
Is that why where they’re using links more, to like, where they kind of use that kind of link to their source, is that to try and foster a bit more trust and a source? I am assuming it can’t hallucinate if it’s then also giving a link to a source.

Alex:
To be fair, a lot of this stuff was prior to it being more connected to the internet where it was doing that. It’s still possible for it to hallucinate in the same way because it’s the same tool, but now that it can access the internet, it would be able to say, here’s actually who wrote this book on this subject and give you that. So it is less frequent with a search. functionality but then you say well then what’s the benefit of using this over using say Google to get that same information?

Louise:
And also when Google famously brought in their AI overview it’s like how many rocks should I eat a day? It’s like seven and it’s like that was from a joke article or something like that to understand that nuance wasn’t really there.

Alex:
Which I think is something we’ve looked at as well, and I think there’s a lot of talk about like AI optimisation being the next acronym, which everyone’s going to sort of use as a part of that. And so it will be a combination of AI overviews and sort of search GPT, but it’s still so new. If you think about how SEO is still seen as a very new industry comparatively, in the sense that it hasn’t been going for much longer than Google has, right? which is very much a 21st, really a 21st century job and industry. And digital PR is a spinoff of that, which is even younger in that sense, which I’m sure that there’s like an official age, but let’s say it’s like 15 to 20 years old. That’s so young in terms of maturity for an industry and understanding how things work. And as a result, this is even less than that. And this would detect that no one really understands from like a consumer level and is changing all the time and is probably overblown a bit. So like, it’s so, none of it’s settled at all. So as much as there’s a real leap to try to get to a point where you’re selling it as a service, I think we have to at least wait a bit before we can effectively be able to say what it’s doing or how, and certainly then how we can affect it, right? Because even if you look at digital PR, there’s still how much links can be effective is still something which is a closely guarded secret, or is, some people believe, is more than others. And so, if we’re dealing with that, after the amount of time that I’ve been building links, which is nearly a decade, that isn’t solved. I can imagine that in the next 10 years, we’re having the exact same debate about how to get to the top of an answer in ChatGPT would be my guess, and no one will probably ever know. I know it’s like a rusty answer.

Steve:
It’s true though like I endlessly fascinating because it’s a really really good way of looking at it there’s sort of various ages of these disciplines and the industry we work in but in terms of the future for AI in PR you touched on certain things there but is there anything else that you could predict that you think might happen in the next few years like it’s changed so much in the last year, two years and it’s obviously going to continue to develop, improve, change, etc. But with your crystal ball Alex, what do you predict might happen?

Alex:
In the general digital PR sphere?

Steve:
Yeah.

Alex:
Yeah, that’s interesting. I think, and this is something which anecdotally maybe you’ve seen as well in terms of your new business generally, that over the last few years there’s been a lot more competition in the industry where more people are doing it. I think speaking about the maturity of the industry in the sense that we now have people who have been doing the job long enough. Like when I was getting promoted and started working in the industry, there just weren’t many people that had done it long enough to be able to be qualified to be like a head of digital PR or a digital PR manager. Like it was just like there’s hasn’t been around long enough or hasn’t been like really called that for long enough so people could could do that job. So whereas now a lot of people have been doing it for five or six years and like can be head of digital PRs, and there’s a lot more competition for those roles. So as a result, more agencies can start doing it. More in-house teams can start to create their own teams as well. And I think that’s what we’ve experienced over the last year and a half, two years, is why there’s so many more agencies sort of starting it, because there are people to fill those positions. And as a result, that has affected competition, which means budgets have gone down, KPIs have gone up. And those things have created quite a difficult environment sometimes to kind of create lasting relationships, I think, with teams because there’s so many options and then you have freelancers in the mix too where similarly they can take work which sort of might not be enough for an agency or performer they might have allocated more budget to do that now we don’t need to we can just use a freelancer which is brilliant I think competition can be fantastic for innovation and can be really good for also like the individuals. But because it’s all sort of in the UK, there’s just not much room, I think, overall. And I don’t know what the endgame is for that, because it can be very hard to create teams that sustain and that develop and grow, I think, as a result. Whether or not that will mean that we’ll get more of a spread of it outside of the UK, it’s kind of interesting how it is such a UK-centric industry. And I think that’s a consequence of there being a lot of PR and comms roles generally here historically and the way our press is sort of quite easy to get links in typically historically. So I think that will either continue to proliferate or completely implode on itself where it will end up with fewer agencies and maybe even I think probably fewer freelancers and things like that because I just I don’t know how sustainable it is in terms of in terms of that. Hopefully not like I think it would be great if it continues to grow or if it could spread out more across Europe or spread across the US but I don’t think it’s going away but I think we’re going to almost sort of hitting a bit of a tipping point I think in terms of size of the industry perhaps.

Steve:
Yeah, I think there’ll be a lot of agencies that will probably fall by the wayside, especially like with AI adoption as well, because it’s time intensive. You’ve got to get the training right. Yeah, it’s competitive out there, but I think AI is going to help us in lots of ways to, well, a lot of the ways we talked about refine what we do, speed up what we do perhaps, but some agencies aren’t going to be able to make it work. I think that’s just by nature and that’s often the case with any like technical innovation, but it’s certainly interesting out there.

Louise:
On the flip side, there is no PR without journalism. What are your feelings towards the future of journalism, particularly with AI? You mentioned already that there are AI journalists out there. what that is, I’m not entirely sure, that kind of like hybrid of a journalist with a login to chat GPT perhaps, I’m not sure. But where do you think things are going, particularly in the UK?

Alex:
That probably worries me more, I think, but it’s felt like it’s always been a bit of a, that the journalism industry’s probably been writing about its own demise for a long time, I feel. I do think a lot of the stuff that’s happened with big publishers dominating SERPs and big companies dominating SERPs from more of a B2C perspective, with how some of the smaller sites have been, like some of the stuff that Giselle’s been talking about with like House Fresh and things like that, probably is, whilst it might be seen as something completely separate from what we do, obviously a lot of that are publishers and what used to be news sites, et cetera, that have just pivoted into selling affiliates, basically doing affiliate work. And I think that that’s a sign of how their models as journalists and publishers have been affected so much from like, and they’re having to find ways to raise money probably. It’s happening at a bit of a like a sort of silly scale, to say the least, and it’s obviously crowding out and punishing smaller sites as a result. But I think that’s also linked to how to sort of what could happen with journalism, where it’s just going to get weaker, or it’s going to get more churn based. And it’s just so difficult. I do think that the stuff that’s really good is the subscription stuff that you see on Substack. So you see quite a lot of these, I can’t remember the name of the company, but they have a bunch of regional sites. So I think it’s like the Manchester Mill and they have a Sheffield version of it too. And it’s all centred on Substack. They can actually make money because it’s a paid for model and they use Substack sort of basis for that. That, to me, is such a good indicator. I think newsletter journalism, anyway, it’s not a revolution to say that now. It’s a very established part of the media diet for a lot of people, but newsletters are almost certainly going to be more and more important because I think it is a person, and a lot of them are journalists who have been let go from their jobs. It’s personality driven, which is massive. You can be kind of consumed at your own pace as well, which is always big. And it’s got a very clear model for making money in that. There’s a really good blog, Creators Spotlight, which is run by, I think it’s run by Substack or Beehive. It’s run by one of the platforms. And they talk about some of their newsletter sort of creators and how they make money and why they decided to do their blog. And it’s just a really good insight as to what is worth sort of following and what probably the future of, certainly the short term future, is for a lot of journalists to make money but also I subscribe to loads of newsletters because it is just a nice a nice way to sort of get my news and sort of in niche areas too and sort of hobbies.

Louise:
It’s interesting like I think in the last six months I’ve seen so much more reference by people outside the digital PR and journalism industry about how crap the experience is on a load of different like news sites so like particularly the regionals, like the fact that you have to dodge through all the adverts, the fact that you can barely read it, the fact that maybe the story isn’t even about their local area. I just hear more and more people say, like it’s just become a kind of in-joke about how awful that experience is. And whereas we have been used to it, like because we’ve been on those sites every day, like for years, it’s started to get so bad it seems that it’s filtered through to people just seem kind of sick of it. And hopefully in a way that forces publishers to improve the experience and But you can understand they pepper it full of ads because they need to get money.

Alex:
I’m trying to open the BuzzStream buzzmarker on one of those news sites which just had full ads and it would just feel like your computer is going to explode. It sounds like an aircraft warming up. So I think the fact that news sites are so much cleaner because it’s all just a pay-for-direct model is better. It’s just like, it seems logical to me that’s how it would go. But people, I think, psychologically have such an aversion to paying for anything on the internet. It’s something, but I do think that could be just, that could be something that can change, that we can age out of more and more. I’m more than happy to pay for like Patreon, I pay for like YouTube premium because I don’t like ads, for example, which it probably isn’t typical. I’m willing to pay for stuff and I think it just requires a bit of time for that to become normalised again for like paying for journalism and paying for those things. Local journalism is a perfect example because I’m quite a parochial man in terms of caring about my local area and it’s very hard to read the newspapers there. And like you say, more often than not they’re covering like national stories, so I don’t need them to cover. I get they do it for clicks and they need to compete in the SERPs and that’s why they do it, but I just don’t want to read about a major news story that I can read about on BBC News on like the Surrey Comet, like it’s just not what I’m trying to do. So, it’s frustrating, but I’m sure they are desperately trying to find ways to fix that themselves internally, like their audience, you know, growth directors and stuff are probably saying like, how can we replicate what is a newsletter model, paid for model, and fix it to it. I think unfortunately for them, they’re owned by big conglomerates who are very, very slow moving, and that can just affect change a lot of the time, and it’s quite tricky.

Steve:
It’s like at the advent of the internet, we opened the gates without really thinking what next or how we monetise this or how we do it properly and now it’s kind of trying to close it off and sort of really think about it. I’m similar to you, I have Patreon and subscribe to stuff and pay for it gladly, stuff I really, really love, but I think also there’s a limit to that in a way because I’m a big sports fan and you could spend hundreds and hundreds of pounds a month just to get athletic. Yeah, exactly. And it’s just, it becomes too much. So then you get more selective. But so I think some publications will have smaller audiences, but that will pay. And that might be the model in the future for some of them. We’ll see. We’re going to ask our final question, which is a choice. We have a choice this season. So Lou, fire away with the final question.

Louise:
Would you either like to share with us a digital PR campaign that you’ve particularly loved recently or the brand that you wish that you could work with?

Alex:
Yeah, I’ll go brand, I think, because I can’t think of a digital PR campaign off the top of my head, which I think would be it. But one of my favourite brands is a brand called Lucky Saint, which is the alcohol-free beer. A couple of reasons I like it is that a lot of what happens with alcohol-free beer options Which kind of frustrates me as someone who cares quite a bit about the space is you get what’s sort of this Trojan horse marketing of the alcohol brands are sort of utilising their 0% as a way to just advertise the main thing. There’s not many brands in the alcohol free space, major brands. There’s a lot of smaller batch stuff which is really good and worth checking out. But a lot of the time it’s like Heineken Zero, Guinness Zero, Peroni Zero. They’re just very clever the way they do that marketing because they’re able to essentially get in spaces that have previously been restricted for alcohol and it’s like Guinness… Zero. And I think that will inevitably be cracked down on because it’s just a way of circumventing traditional blockers for alcohol advertising, which I think is important. Why I like Lucky Saint, I think, is that it has become the standout brand as an alcohol-free brand in and of itself, without it being sort of an asterisk against an existing sort of alcohol brand, which is probably owned by a major company. Big fan of it. Really fan of some of the way they do their marketing in terms of the out-of-home offerings that they do, which are very, very cool. Very good brand identity. in terms of literally what it looks like. It’s also the best one, like as someone who’s- Is it nice? I’ve actually not tried it to be honest. It’s worth trying on tap. The walrus just here, they have it on tap as well, which is worth it, I’m pretty sure. Also, they did a thing where they bought a old pub that I believe was shut down during the pandemic in somewhere in central London, I can’t remember, I think near Regents Park. They fitted it out as the Lucky Saint pub and so it’s a working pub which I think does serve alcohol as well. They have other brands but they have their own stuff there on tap. And their offices are above it too, which is very cool. I’ve been there for a drink and it’s just very, very lovely inside. It’s a normal pub. authentic way to market your product as an alcohol-free beer. And they do a lot of running stuff. And they do running clubs, which again is a really logical fit for them in terms of marketing. And running clubs are super popular the last few years anyway in terms of ways to socialise in the city. But there’s a lot of things you can do as an alcohol-free brand if you’re not attached to a bunch of a traditional alcohol brand, I think. And they are doing all of those things, which I think is very, very cool to see. They don’t do it in a preachy way also, which I think is a very important line to walk when you’re trying to attract people to a brand which could easily end up being a bit preachy because it’s a bit health conscious or it centres around something which is like, you shouldn’t be doing this. Which is like, I think that they don’t seem to have that attitude at all. It’s just like we make a product which we think is really good and we enjoy marketing it. So, they’re a very, very good brand. Whether or not I’m doing digital PR for them would be interesting or not, I don’t know, but I do really like the way they operate.

Steve:
Well, if Lucky Saint are listening, do get in touch with Alex and Distinctly. It’s been such a pleasure having you on, having you in the studio for the very first time we’ve had someone in the studio and it has been a genuine pleasure and we’ve loved listening to you talk about AI and all your experience there. If people wanted to get in touch with you, either to work with you at Distinctly or just to kind of pick your brains, I mean, you may not want them to, but we’re going to give them some contact details now. So, how should they get in touch with you, Alex?

Alex:
Yeah, brilliant. So, search me on LinkedIn, Alex Cassidy, Distinctly, but you can email me alex.distinctly.co.

Steve:
Lovely, lovely. We’ll do get in touch with Alex if you have any questions, follow-ups on AI or you want to work with him at Distinctly, brilliant agency. And to finish off, we’re also going to mention your new podcast at Distinctly, Linking Out Loud. We have used AI to work out how to quash podcast has come up with a brilliant strategy, so watch this space for that. No we haven’t, it looks great. I haven’t listened to it yet, but you have and you really loved it.

Louise:
I listened to one of the episodes and it was very good, enjoyed it very much. So yes, if you like this kind of podcast then please do watch it.

Alex:
I know Rachel listens to you guys a lot too

Louise:
What a podcast-loving this has become.

Steve:
We could do one of those crossovers. We’ll get them on. People may want that. People will be calling for that. It’ll be like a Marvel crossover.

Alex:
I’ll dress up.

Steve:
Alex, it’s been an absolute pleasure. Thank you so much for your time. Brilliant. Thanks very much.