Can You Outsource PR to AI?

With many corporate leadership teams issuing a strict mandate to integrate artificial intelligence across every business function, marketing and PR leaders are facing a tough operational question: Where does AI actually drive efficiency, and where does it create massive strategic risk? On this episode of FiredUp!, we take a practical look at how to rank your public relations workflow from low stakes to high stakes. We address the risks of over-automating your communication channels,  how to properly apply human judgment and workflow automation, and more. This week, episode 139 of the FiredUp! podcast is about if you can outsource PR to AI! 

Download the Multiplier Marketing Megapack today. Exclusive offer for all our listeners — get all our Startup Guides in one go! Over 50 pages of advanced tips and advice that dive deep into content marketing, search advertising, and marketing attribution.

In this episode of the FiredUp! podcast, the Firebrand team shares where AI makes sense in a  public relations workflow and actionable steps you can take right now to avoid losing credibility by using AI for too many PR functions. 

Morgan and Chris discuss:

• Automate the Low-Stakes Research: Use custom AI agents to handle time-intensive, list-making tasks like tracking industry awards, identifying speaker opportunities, and logging application deadlines. This high-efficiency research automation saves hours of manual labor while carrying almost zero strategic risk.
• The Peril of Automated Copywriting: While an AI tool can quickly structure text to look exactly like a press release or contributed article, it naturally defaults to recognizable, lowest-common-denominator language. Discerning editors and journalists increasingly reject these submissions, making human composition essential for serious media outreach.
• Protect Your Media Pitching: Keep automated text generation completely out of your direct reporter communication. Journalists intensely dislike receiving cookie-cutter, bot-generated pitches; maintaining a strict one-to-one human approach is the only way to build durable media relationships.
• Supercharge Media Profiling: Shift your AI usage away from writing and toward intelligence gathering. Use advanced prompt structures to analyze months of a target publication's coverage, identifying dominant themes, data preferences, and writer beats to manually craft a highly tailored pitch.
• Maintain the Thought Leadership Standard: True thought leadership requires a distinctive, original point of view that cannot be replicated by probabilistic sentence completion. Use human judgment and authentic storytelling to ensure your corporate messaging stands out from the sea of automated slop. 

Your Action Item This Week: Identify one highly manual, low-stakes research task on your plate, such as building a target event or speaker list, and build a structured AI prompt to compile the data for you. Use the hours you save to call a key partner or write an original, human-generated piece of thought leadership. 

Thank you for listening! Tune in to all the episodes for practical tips on crushing your startup marketing goals. Don’t forget to follow, rate, and review the podcast, and tell us your key takeaways!

CONNECT WITH FIREBRAND:

Firebrand is a startup marketing agency. We help tech startups secure outsized marketing outcomes on their path to growth.

Youtube

LinkedIn

X (Formerly Twitter)

Work with Firebrand

Apply to Be a Podcast Guest

5.5.2026

Transcript

Morgan McLintic: 0:00

Many companies have a mandate to use AI as much as possible, wherever possible. So today we take a look at where you can get the most value from using AI. Hello, everyone. Welcome to Fired Up, the podcast for marketers working in early and late stage startups. Hello, welcome to Fired Up. My name is Morgan McClintic. And today I am joined by a head of media, Chris Albrecht. Hey, Chris, how are you?

Chris Ulbrich: 0:31

Hey there, doing great.

Morgan McLintic: 0:33

So we're going to look at where you can get the best value from AI in PR, like where can you really sort of outsource PR to AI, and obviously, as a PR agency, we have a bit of an insight and a bit of a bias here, right at the top of the show, because I think there's a lot of parts where you know the state of the art of modern AI would not deliver, but there are areas where it really can, and so let's just break up the PR process, some of the common activities of PR teams, and then stack rank them from low stakes to high stakes about where the best use cases are, and where we feel okay AI could probably work really well here, and then others where we think, look, this is a really high stakes activity approach with caution, so let's start with the low stakes, and I think, as a general theme, you know, when it comes to research, AI is really good, and so one activity within PR, which is fairly common, is researching awards and the deadlines and speaker opportunities, and this is a fairly low-stakes activity, and fairly common for AI already. I'd say,

Chris Ulbrich: 1:54

yeah, and we've built our own agent to do that internally, and as it happens, I just used it for the first time to spin up an awards list for a new client worked great. It saved hours of time, and the nature of the task is a sort of a combination of stored experience and discovery. So, if you are coming in to a new field and you don't have any experience with it, and say you're at an agency. If you've never worked with a client of that type, and you're just coming in fresh, probably the AI is going to do about as good a job as you are going to do in finding relevant awards and speaker opportunities, provided you give it enough context to evaluate the opportunities that it finds online. Now, in practice, if you are an agency like Firebrand, with all the combined years of experience involved, it's unlikely that you're going to be coming to just about anything completely fresh. When we start a an awards list, we have a decent idea what ought to be in there, but it does take time for someone on the team. If you haven't, say, work with a client in a particular space in a year, you have to refresh all the deadlines and submission dates. And when does the event actually happen? What are the criteria for entry, and whether it's an award or a speaker up, and you have to look all that up, and the agent can do that pretty darn reliably. I found with hours that I had to give it some direction to get it to just have it avoid some common mistakes, but once it had that context to refer to, it was pretty darn reliable, and this is the lowest stakes use case, really, that you can imagine for AI. So, with awards and speaker ops in the big picture, the downside of missing a relevant op is relatively low in the context of a robust program. If it's a really important op, it's unlikely you're not going to know about it, and or find out about it as the campaign progresses. So, the chance that the AI doing research is going to yield much worse results than a human doing the research seems to me relatively low, like they're probably going to come up with roughly the same results or better, and if they don't, the downside is relatively small, and the time saved is significant. This is any kind of list making, it is a very time-intensive task for PR teams, then it can be a useful task. It can be a rewarding task. It can force you, as a PR pro, to engage with the material, but if you're not working, if you are not a PR pro, and you're a marketing person that's trying to get across any number of tasks, this is a good one to outsource to the AI.

Morgan McLintic: 4:59

Yeah, absolutely. Absolutely, any human is going to miss a few, and this might too, but it'll really compile it in a nice format quite quickly. It can update and check, hey, just go and check all the deadlines, nothing's changed. Make sure, update all these prices on year two, just take another assessment on that, and it'll do that in minutes, and often that means you don't miss a deadline, and the work gets done right.

Chris Ulbrich: 5:21

Obviously, you're going to need to make sure you check any metrics that it inserts, because that's where the rubber can meet the road with the award up, but also if it's where the AI might be most likely to make it like a material error. Now, when it comes to speaker ops, I'm less worried about the pro's quality than I am with just the AI's ability to come up with a proposal that will meet the needs of whoever is putting on the event. Typically, with a proposal, you need to be thinking hard about the takeaway for the audience, and again, given sufficient context, an AI can probably do that job, but I have found working with clients that they, the context you have as a marketer or a PR person in coming up with a proposal for a client is pretty thin, and you have to do a lot of thinking and question asking to come up with a proposal that works. You could come up with a credible speaker proposal with AI, and it looks credible. I would be skeptical that you would be successful with it.

Morgan McLintic: 6:34

I would leave AI to say, "Hey, we're going to put together a panel on this topic with these people, because you might not be able to get those people, and it doesn't have any of that context, and know that you can speak to that topic, so that kind of big lift suggestion, no. But the kind of, here is the bio of the person, and I'm just going to submit them, I think, and it's maybe a low tier speaking up, and you just throw it in. I think it's going to be good at that. Let's move on to copywriting, a topic that we have covered lots of times, and as it evolves as state of the art, and as the norms evolve here, because I think it's really important, but we still got this under low stakes when it comes to press releases, anyway, like an AI press release that you give it the information, and you say,'Here's product marketing, here's information about our new launch for a product, here's the deck, the release notes, give me the press release, go, and you look at the outcome, it's going to look like a press release, right, and it's going to read blah blah blah blah blah, it's going to have all of its fingers and toes and quotes, and it's going to be pretty convincing, and I would say a lot of companies probably just are already at the stage of away we go press publish with minimal changes.

Chris Ulbrich: 7:55

We put this in the low stakes category on the theory that if your back is to the wall, and you had to put out a press release this way. What are the chances that you're going to hurt your strategic goals by issuing an AI written press release or a mostly AI written press release? Given that it is increasingly common, you're not going to stand out from the pack for better or worse by doing it this way. The voice might be recognizably AI, but people aren't going to think much less of you for it, and probably as long as you're checking the main points in, like the bullet points on product, and making sure that the essential facts in the press release are right, it's not going to be a strategic disaster for your company, right? That said, I really don't recommend it to our clients. I go back and forth, because, like, writing a press release is nobody's favorite task, so even PR people that you might think want to, like, hold on to that task, because it's something they can own. Truth is, they would probably be a lot of them would probably be happy to have a magic press release machine that did a credible job of it.

Morgan McLintic: 9:03

Oh, you'd be a multi millionaire, right? Hundreds of press releases, and I don't need to write another one.

Chris Ulbrich: 9:09

Well, right, exactly. And yet I've just got done saying it will do a credible job, and what I mean by that is it will do a credible job for someone who's not reading closely, and this gets at the question of what the press release is really for, insofar as you are trying to write a press release that communicates the essential facts about your news and connects it with language that is credibly press release e, but doesn't need to be absolutely optimal. A press release can do that, but if you expect this press release to be read by a discerning audience that might be getting into specific details and might be actually zeroing in on the choice of this word versus that word, you describe yourself as. Are being a company that specializes in this category of technology, but I don't know. The last time we talked, you said you were over here in this different, and that's the kind of attention to detail that a journalist would bring to it. So I think about this as a low stakes use of AI if you are not demanding a lot from your press release, and I think that if you're trying to put a press release in front of journalists, it's actually a bad idea to do it with AI, because the AI just doesn't, in from my use of it, it makes decisions about the choice of words and the organization of ideas that can be confusing to anyone who's reading closely and really interrogating what you mean, and if you have an AI written press release that's going to be interpreted for the media by a trained PR person who can turn it into a pitch and is aware of the reporter's previous experience with you and how language in the press release might sound wrong to them, or might confuse them. Then, sure, it's relatively low stakes, but I would say this is relatively high stakes if you are taking the tack, which we do not recommend. That your press release is basically the main vehicle you have for interacting with media, so in that case you're asking the press release to do a lot, and I would not let the AI do it. If you're just trying to push a press release out the door, and you want, like, a record out there of the basic facts of an announcement, and you're not really trying to do media outreach around it, you could do worse than use AI. It's not going to be a disaster for you. You still need to check it, but you could save some time that way. It's just in practice for a company that's really trying to do media outreach. It's pretty rare that you're in that situation. You do actually find that you need the PR person to be involved in composition, and as we've discussed before, trying to mix AI composition with human composition is a lot harder than it sounds, a lot more time consuming, and it doesn't work as well as you expect it

Morgan McLintic: 12:10

to. A press release is a public statement, particularly you're going to put it over the wire from the company, and it needs to stand behind those words, so those words need to be closely chosen. That I would say it needs to be handwritten almost in its entirety, maybe in its entirety, if I'm feeling particularly Puritan. So, an AI-written one is just not going to do it, but we totally understand that sometimes that's what companies do. That's a bit about press releases. The other part of copywriting is bylines, and here this is sort of a vehicle for thought leadership, and I want to come on to crafting the thought leadership position separate from just writing the byline. So let's think about a case where all right, we know what we want to say, and I've got my bullets from my spokesperson, and I've done the interview, and I'm just gonna.. I need to turn this interview, maybe even done the interview right, and I just need an 800 word byline, feed it into Claude, press the button, job done. How do we feel about that?

Chris Ulbrich: 13:17

So, for listeners who might not be familiar with the term byline, we're really talking about like a contributed article that would run in a third party publication. For purposes of this discussion, you could broaden the category to include your social media, like long form, so longer form social media, like LinkedIn posts, right? And that too, that sort of gets at the question of stakes. So low stakes would be, I'd say, a LinkedIn post, which is increasingly where people put content that would have, in the previous days, have been a blog post. AI composition is so rampant on LinkedIn that you're going to be fine at the same time people are so tired of it that you're probably not going to get the benefit you're hoping for if you are composing with AI for social, it's instantly there's a particular chirpy, I've seen someone describe it as like a golden retriever kind of energy about LinkedIn, about AI copy for LinkedIn, that's again instantly recognizable, it provokes eye rolling, so this is like death by 1000 cuts. It's low stakes in the sense that any one post is not going to sink your company or it's not going to sink your marketing initiative, but it does gradually, I think, erode trust in your messaging and in your company, and that it affects the sense that buyers are going to have of your professionalism and your capabilities, so it's not great, but low stakes in the sense that you're not going to have an immediate major downside if something goes sideways. Now, I would categorize contributed articles, though, as a medium stakes use case for AI. Because if we're talking about how likely is it that the use of AI is going to cause your project to fail, we are increasingly seeing editors who are just declining articles on the grounds that they can see AI influence in the writing, either because they are using commercial AI checkers, which different people have different opinions about their accuracy, but the fact remains that some editors use them. Some editors will reject your copy, or which you might have labored over one way or another. Even if you used AI to write it, sometimes it can still take you a long time to put all the pieces together for the AI, and all that can be wasted if your publication of choice declines the piece because you used AI to make it similarly, some editors, I feel like I'm pretty good now at just at spotting AI, and I have seen it. Look, I'm not saying that you will get caught or that editors will necessarily care. I have definitely seen it in the op-ed page of the Wall Street Journal, where the op-ed editor wrote a whole piece about a very indignant piece, because some detector had called out op-ed pages like his for allowing AI content to get through, and he wrote this indignant piece about impugning the accuracy of AI detectors and whatnot. Look, I guarantee you, I've seen AI. In fact, I've even looked at 800 word articles and I said, here are the 200 words that this person wrote themselves this op-ed, and this is the 600 words that the AI generated. So people who read these sorts of things regularly can get a feel for the AI voice, and I do feel like, as we've discussed on this program before, there's just a - there's a growing exasperation with AI-generated content, you could call it a backlash, what have you, but it's negative sentiment, and you do run the risk, not the certainty, but the risk that if you use AI, your work can go to waste. So, I would categorize that as a medium stakes use case, and I really wouldn't recommend it. I don't, I don't like using AI for writing articles, researching, yes, fine, but the more that I try to mix AI-generated content, aside from the occasional copy edit with human-generated copy, I am increasingly skeptical that the two can be reconciled cleanly.

Morgan McLintic: 17:31

So, medium stakes, as in you're not going to achieve your goal, i.e. the editor's not going to accept the byline, and even if you are successful, you're not going to achieve your goal because your readers are going to not read it or have a negative reaction to it, and not hear your message, so you're actually doing yourself a disservice. There's a risk

Chris Ulbrich: 17:50

that you'll generate a negative reaction.

Morgan McLintic: 17:52

Yeah, even though, yeah, I mean, 100% Even on LinkedIn, people who I know, who I know are very great writers, although I then read their AI stuff they post up there, I'm like, why are you doing this? I'm silently judging you, and I don't want to, but you just can't help it. You're like, okay, you didn't even write this, and now you expect me to read it. All right, now we want to carry on. We're into our medium stakes here, so the pressure is ratcheted up, and I want to start with one of the things, which is pretty common for emerging brands, is rapid response to breaking news or a call for sources from editors looking for commentary? Hey, does anyone can comment on agentic orchestration, and how you do that? And you think, oh boy, have I got a lot of things to say about that, and it helps you get your thoughts across, and it helps you break into new publications and build new relationships, really important. Also, ultra competitive. We did a whole episode about this. Speed is really important, accuracy is important. Take a look at that if you want to know the real breakdown on that. But you know, very tempting to use AI here, Chris, and I would say monitoring the news, or even monitoring the tools like Prof Net, you know, I have some Claude Code things, I, you know, I press the button and Claude Code takes a quick look at Prof Net for a few things I'm looking at, and says, hey, you know, is there anything there, because their search system is terrible, and it does it really cleanly, and it can find some ops, so like sort of finding the opportunities and calls for sources, monitoring what's going on in the news, you know. Here are the publications I care about. Here are the topics I'm looking at. Give me a summary of the news. Probably most people have got some kind of little clawed script on that. I'm going to say that part of this is fairly low stakes, like finding the ops is a research task, like finding the awards, and if you haven't already automated that through AI, then you probably should, because yes, there's value to reading the news. I'm not saying don't read the news, but there's so much to read, filtering things out, so that you can focus your time. This is. This is a great use of AI, and I think the ship has sailed on this because you should be using it, but having found this opportunity, what's AI's role in the next step, which is crafting a response?

Chris Ulbrich: 20:12

I would say ideally none, and I'm increasingly an absolutist on this point, similar to what we've seen with bylines, and maybe even more so, editors are rejecting quotes, comments that are clearly AI-generated, and at the same time, our clients, who are increasingly accustomed to doing everything with AI, to answering their emails with AI, to writing speeches with AI, or drafting memos, internal comms with AI. They think nothing fair enough of writing their own quote with AI. So we'll send over an opportunity with some suggested avenues, ways that the spokesperson could go about answering the question and trying to get a few thoughts that we can turn into a comment, and five minutes later they'll come back with half a page of commentary. We thought, "Wow, this person could really crank it up. Then you look at it, and it's just full of the AI isms - it's not this, it's that, or what we're seeing here is a quiet shift. All those tells, and you just think, "Okay, there's no way that this not only is this not going to pass muster, but we're actually failing in the task we've set ourselves, which is to establish ourselves or establish our client as having a distinctive voice that is valuable in the conversation and is actually standing out as a thought leader on the topic we're commenting on, we always call these thought leadership programs. I think I would go so far as to say it is impossible to be a thought leader if AI is generating your thoughts. It simply can't be done, because you're going to get a lowest common denominator version of whatever idea you might have had, and yes, it does take time to wrangle your thoughts into sentences, and we can help with that, because we've worked with you long enough to understand your shorthand and how you would say things, and the longer we work with you, the better we get at that, but AI, while in theory it can be trained on your voice, I haven't seen it be successful at that, and it regularly reverts to its bad habits of composition, which are really recognizable. It reverts to its own voice more than you would hope, even when it's been theoretically trained on yours. I suspect I don't know that much about AI in the big picture, but I get the sense that whatever training, whatever amount of training you give it, whatever amount of like fine tuning you can do for it pales in significance to the tremendous amount of training it's had to develop the habits that it has out of the box, and it takes a lot of, if you're going to steer it away from those habits, so that it's not recognizable. It takes a long time, and I'm not even sure that you can do it, and most of our clients haven't done that. So, just the more I see clients try to draft quotes with AI, the more I counsel against it. Yeah,

Morgan McLintic: 23:25

I've tried to do this, by the way. I've spent several days building an agent that takes all the quotes from this spokesperson before, here's their bio, here's everything they've ever written, here's all the articles they've done, every podcast transcript they've ever been in, like trying to sort of replicate their brain, if you like, and everything that they've said to then say okay, and then vectorized all that, and then you put in your query and say, here's your new call for sources. What would this spokesperson say if they've said something about this before? I want to know that verbatim, so give me that, so that I can, because we can't remember everything for every spokesperson, right? But just give me what they've said before, and then give me what you think that they would say based on this received body of knowledge, the sort of custom thing for this particular spokesperson, and the topics that they want to talk about as their platform, and it works, and here's the things that they've said before, and maybe you can reuse those quotes verbatim, but you still have to rewrite the quote, because if it then says, "Hey, I think they should say this, you then have to go, "Okay, but now I have to say this, but in different words,

Chris Ulbrich: 24:34

but still sound like the spokesperson,

Morgan McLintic: 24:35

but still sound like the spokesperson, and that's doable, and maybe that saves you some time, but you've got to maintain this thing with all the latest quotes, and we've got to keep it up to date with their thoughts. So, I've tried to crack this with an agent, and I'm not there yet, because you still have to rewrite. Maybe we'll get that.

Chris Ulbrich: 24:53

Maybe we'll get there. It's funny, I was just listening to a radio piece today on NPR about the opening of the new Teddy Rosen. Museum in North Dakota, and it sounds really elaborate, and one of the exhibits is done in partnership with Microsoft, and they basically let you talk with, in real time, with Teddy Roosevelt's ghost. I mean, it sounds like it's a hologram of Teddy Roosevelt or something, but you ask it questions, Teddy Roosevelt answers, and you know, I only heard one exchange, and it sounded fine. The interviewer asked why he likes to say bully, and Hollow Roosevelt gave a reasonable answer, but you know, maybe at the cutting edge what Microsoft is doing there with this big budget project, it is possible to capture someone's voice credibly now, but I'm not seeing that yet with the kind of AI setups that our clients have, and

Morgan McLintic: 25:50

then that instance they don't have to then have the output not be AI, the output can be AI because you're just consuming it, you're just listening to the person, it doesn't have to then go through another AI filter, I don't know.

Chris Ulbrich: 26:01

Oh, it's an AI generating the answer based on the training it got from all of Roosevelt's writings, which

Morgan McLintic: 26:08

they've got a lot, because you know, yeah, exactly. Okay, so rapid response, do the research, you should, but craft your own thoughts about it, you know, there's not really a shortcut to that. And then we're going to talk about media pitching with AI, and I just feel like we got this under medium stakes, but you know, this makes my skin crawl. Just talking about it,

Chris Ulbrich: 26:31

I feel like it's high stakes. It depends on the circumstance, or like on your media program, but I think you could make a good argument. This is, I think you could say it's medium stakes from the standpoint of your total marketing program. If you pitch media this way, you will fail, and so then your marketing program can only be as good as it will be without media. And even though media is my job, I'm not going to pretend that it is the be all and all of marketing, and that a company without a media program is absolutely hopeless, has no chance of making it with marketing. You do, but if you care about your media program, you absolutely cannot have AI pitch the media. It's again instantly recognizable. Reporters hate it, absolutely hate it. You will never build a strong relationship with any kind of reporter that is going to be meaningful to you, and they're just going to receive it as spam. There are these agencies out there who are pitching. Sometimes we come up against them in new business, not often, but once in a while we get the sense that the companies we're talking to are also talking to agencies that are promising a fully AI-enabled workflow. No way. I mean, like, Godspeed if you give it a try, but I am confident this is gonna a disastrous approach to AI. I mean, if you just listen to reporters talking about how they feel about AI-generated pitches, the loathing is palpable, so just don't do it.

Morgan McLintic: 28:06

Media relations has to have a relationship, and they don't want a relationship with a bot. But I will say that doing the research into, okay, who are the reporters that might be interested in this topic

Chris Ulbrich: 28:18

could

Morgan McLintic: 28:18

be quite good. Read all the things that they've just written to find out what they typically care about, and give me a synthesis of that, so that I can refine my pitch a little bit. If you don't have the time to do it yourself, or you don't know it out the box, probably quite useful to do it with AI than not to do it at all, and finding the people, researching them, having it inform the pitch that you do, I mean, just to be clear, we never advise just one to many media outreach, like almost email marketing to reporters on your list, that everyone gets the same thing. There are schools of thought that sort of spray and pray might work, and I'm sure that there are agencies that still do that, which, if you're a PR person, you just wish they'd stop, but there's always a new generation of people who do it. So, we're talking about one to one emails, which is the best practice. Don't have AI write that pitch, but maybe have AI help you with it, so that you can craft it, and you're a bit more conversant with it. And, like, I think you're missing a trick compared to not doing that at all. Surely,

Chris Ulbrich: 29:20

I'm of two minds, because this is just anecdotal, but I never found Chat GPT to be especially good at understanding what sorts of reporters might be interested in what sorts of story. I found that when we switched over as an agency, Claude, I found that Claude actually has much more sensible things to say about why a reporter would or would not be suitable for a particular story, and that might just reflect the sort of uniform progression of AI. Maybe Chat GPT would have been just as good if we were still using it, but I still find that even though I've been pleasantly surprised by its ability to judge the suitability of a story. It's not very good at at figuring out what makes a pitch good, so if you're developing a pitch with it, that is, then a lot of times research into reporters will bleed into pitch writing, and I think you just need to be careful to not, and it's not even just that the AI is writing the pitch, that's a bit of a problem, but I think that AI is its training. It's been so traumatized by its training on endless reported complaints on places like Reddit, or I don't know, LinkedIn, or wherever, whatever it was trained on. It's been so traumatized by complaints about needing to make your points early in the pitch or keep your subject line to a certain number of words, all of which are good advice, but it won't even acknowledge really that you've made progress. It gets caught in a loop, but I've seen reporters say this is true about when they use AI to edit their pieces or give them advice on their pieces. The AI is never satisfied, it never knows, it'll never admit that you've got a good pitch together, or that you've got a good story together. If you're a reporter, it's always finding nits to pick, and you could end up spiraling on that rather than just realizing I've got something pretty good here, we can go out with it.

Morgan McLintic: 31:17

That's fair. And I also find the opposite when I'm brainstorming with it, when it goes, oh yeah, this is the idea. It's locked. Go with that, and you're like, wait, no, this is clearly half baked. What are you talking about? I have used it. Look, just look, we're talking about media pitching. I've also built research as part of an agent. Look at this publication, gather up the last 90 days of every article that they've ever done, and analyze it across. Is it news, is it opinion, is it commentary, is it feature stuff, does it mention a vendor of any kind? What are the main themes, and literally break down what is this publication, it's HR Brew, whatever it is, what is it structured to write about, and what most often gets covered? What are the main companies that get covered, the main themes, and then who are the main writers about this topic, and then as I'm writing, as I'm thinking about my ideas. Okay, these guys normally cover news, and 87% of their coverage has data in it. So, do I have data, because otherwise I'm punching in that 13% and it helps you go, okay, I need to sort of narrow down my pitch. It needs these components. I personally find that quite useful, particularly if you're not familiar with the publication out the gate.

Chris Ulbrich: 32:27

Yeah, like in general, AI excels at summarizing, distilling, characterizing, so if you're asking the AI to look at a publication and call out what are the elements that typically appear in their news coverage, which is an incredibly useful exercise. If you don't have a PR team to do it for you, AI can really help you profile your target publications and help you think about the kind of stories that might appeal. That's a great way to use AI.

Morgan McLintic: 32:54

Don't do the outreach, but do the research. Yeah. All right, Chris, that was great. You know, I was.. it sounds like we said at the beginning of the show, we're going to be a bit biased about human judgment and human creativity and building relationships, which is fundamental to PR, but I do want to say that we find that the potential for PR in a lot of these areas to accelerate the workflow on the research, the profiling of research opportunities, reporters, et cetera, just understanding what's going on is really there, and it's automating, it's freeing up some time to be more creative, and to build those relationships, and to spend time on other things. We're certainly trying a lot of these things, but I think the vast majority of activities here are not quite ready with the current state of the tools to be fully outsourced to AI. All right. Thank you, everyone, for listening. If you enjoyed that, we would love it if you would follow the show, subscribe, and, of course, drop us a comment. Tell us where you are finding great use of AI, and if you disagree with us, then that's fine, and we'd love to hear, and we will see you all again next time.

Unknown: 34:02

Bye.