Ensuring AI is an asset — not a liability — for your agency

Agency Leadership Podcast

In this episode, Chip and Gini highlight the challenges and potential pitfalls of over-relying on AI for content creation in PR and marketing.

They discuss instances of AI-generated content gone wrong, such as the fake book list published by the Chicago Sun-Times and poorly crafted AI-generated pitches. The hosts emphasize the importance of human oversight, individuality, and storytelling in maintaining quality and building relationships with the audience. They also delve into Google’s EEAT guidelines and how PR professionals can leverage their expertise to stand out in search rankings.

Finally, they discuss practical ways to efficiently use AI while ensuring the content remains authentic and relatable.

Key takeaways

• Chip Griffin: “As we start to try to create mountains of content using AI, we start losing individual perspective, personality, those kinds of things that can help set us apart as agencies and can help set our clients apart.”
• Gini Dietrich: “Let AI help you draft it and then add in your experience and expertise. That’s where you’ll win every time.”
• Chip Griffin: “If we’re just churning stuff out because we think that’s the only way to win the future battle of AI search, we’re mistaken.”
• Gini Dietrich: “AI is absolutely a huge, huge help. It still needs to be fact checked, it still needs oversight.”

Related

• What does ChatGPT and generative AI mean for PR agencies?
• Is AI writing an agency’s friend or foe?
• Setting AI policies for your agency

View Transcript

The following is a computer-generated transcript. Please listen to the audio to confirm accuracy.
Chip Griffin: Hello and welcome to another episode of the Agency Leadership Podcast. I’m Chip Griffin.

Gini Dietrich: And I’m Gini Dietrich.

Chip Griffin: And Gini, I think I’m just gonna use some AI to spew out all, all sorts of spam for these episodes so that we just, we’re not even part of it anymore. We’re just gonna contribute to this atmosphere of junk content that’s proliferating these days.

Gini Dietrich: Yeah, I read. As we’re recording this, the Chicago Sun Times got in trouble for publishing a summer reading list made up of real authors, but fake books. And their excuse was, we hired a freelancer to create this. So I was like, oh, that’s akin to saying the intern messed up. You still, especially as a newspaper, have to verify that it’s accurate. So yeah, the list was of books, was all made up. Books that AI had generated for them and they didn’t check it. And then other newspapers picked it up and ran it too. So there’s something to be said for AI’s great for some things, but it still needs human oversight.

Chip Griffin: It, it not only needs human oversight, but I think that, you know, one of the things that, that we lose as we start to try to create mountains of content using AI is we start losing individual perspective, personality, those kinds of things that can help set us apart as agencies and can help set our clients apart.

Because if, if we’re totally dependent upon AI, obviously hallucinations continue to be an issue, right? And, and yes, and can absolutely cause problems. There are, you know, legitimate fears over inadvertent plagiarism on the part of AI and those sorts of things. And those are all important.

But I, I think, I think so much more important is that you start to lose the voice of the agency, the voice of the client if you rely too heavily on AI. And, and I think everybody listens to us, knows that we are fans of the use of AI, right? Yes. In the work that we do, and we should find ways to leverage it in order to be more efficient and all of those kinds of things.

And, and, but that’s not saying that you should just contribute to this mountain of content that people are committed to putting out right now, whether it’s on LinkedIn, right? And we see all this AI generated garbage on LinkedIn. Although, yes, just for the record, all these people who post these dumb things about if you have an em dash, it’s AI generated.

That is hooey.

Gini Dietrich: I agree because I use them.

Chip Griffin: I have used em dashes forever

Gini Dietrich: Forever, right? I use it too.

Chip Griffin: I have not once ever used AI to produce even a portion of my LinkedIn content because it’s not… I use AI for all sorts of stuff. My LinkedIn content is not a place where I feel like I need to do, I feel like I can write, you know, a five or six paragraph post all on my own, thanks.

Gini Dietrich: All by yourself?

Chip Griffin: All by myself.

Gini Dietrich: Well, at the same time that this is happening, and I completely agree with you, that AI can make us more efficient, but I think that we’re overusing it. Is the Wall Street Journal published an articles say that called Will AI Empower the PR industry or Create Endless Use of Spam, which is always awesome to read.

But it goes on to say that a pitch for March stood out from the flood of usual emails from PR agencies that had an unusual subject line. It was, you care more about Tesla than a cancer killing thousands. The email went on to scold US-based reporters for giving priority to Tesla earnings news and celebrity gossip over the work of a German firm that develops tests for cancer.

And then, and then. It rattled off statistics about increasing cancer rates and stated quote, this isn’t a future problem. It’s happening right now, but you wouldn’t know because you’ve been covering crap that doesn’t matter. Do better. All written by AI.

Now, if you are using AI to craft your pitches and make you more efficient, that’s fine. If you’re using AI to help you personalize to the person you’re pitching, that’s fine, but for the love of God, like have some oversight. Do you really think do better at, in an email scolding reporter is a good idea?

Chip Griffin: I mean, there may be a time and place for that kind of messaging, but I would want that to be incredibly deliberate and, and not put that in the hands of computers to figure out when to do that.

Right. And, and in general, I would, I would advise against it, but I mean, I could probably see if, if you had a relationship with a reporter, maybe you would take a similar tone. Maybe not identical, but you know, you might. You know, there, there are certainly reporters I’ve known over the years that I could be pretty blunt with, and, and I have in the past said to them, I don’t know why you would write about that particular story that you

Gini Dietrich: that’s, that’s like a one-on-one thing, right?

Correct. That’s not a pitch that you send to all reporters.

Chip Griffin: No. Yes. Well, I, I also don’t believe in sending.

Gini Dietrich: Fair,

Chip Griffin: broad-based pitches I believe in,

Gini Dietrich: yes,

Chip Griffin: one-on-one pitching. And I know that that is not a popular thing amongst many in the PR community,

Gini Dietrich: but it’s far more effective.

Chip Griffin: It is. It is far more effective.

It does require more effort.

Gini Dietrich: It does.

Chip Griffin: But I, I think that, that you owe it to yourselves and to your clients. To do it that way instead of building these large lists and just blasting stuff out. I think that was certainly true 20 years ago. I believe it’s even more true. Absolutely. Today.

Gini Dietrich: Absolutely. Yes.

Absolutely.

Chip Griffin: And I would rather send a targeted pitch to three people. Yep. Than an untargeted pitch to 3000.

Gini Dietrich: Because if you send a targeted pitch to three people, you’ll probably get three responses and maybe even two or three placements. If you send it to 3000, guess what? The Wall Street Journal does an article about how much the PR industry sucks and we’re contributing to the sea of spam.

That’s awesome. Awesome.

Chip Griffin: Well, and I, I think, you know, part of the, the challenge is that agencies and clients are hearing from the gurus that, you know, in order to be relevant to AI, we need to be churning out more and more content, right? To help train the algorithms and all of that sort of thing.

And so, so the solution that people see is, okay, well if we’re gonna have to create this much content, we’re gonna have to use AI to do it. Which first of all, creates this ridiculous cycle where AI is essentially feeding itself,

Gini Dietrich: right? Yes.

Chip Griffin: I mean, you know. Mm-hmm. And, and then, I mean, basically that’s sort of like intellectual inbreeding.

And you know, who knows what you end up with as a result of those genetic mutations,

Gini Dietrich: right?

Chip Griffin: But you know, more importantly, I think that

Gini Dietrich: intellectual inbreeding.

Chip Griffin: I think that, that we, if nobody else has used that, I’m taking that one. That’s amazing. You know, so, so we need to be thinking about how we’re putting things out there that AI can consume that really represents who we are, how we think, and, and do the same for our clients, because that’s where we’re gonna have the most power and influence. It’s not by putting out stuff that’s the same as everybody else. It’s just a different byline. And, and I think that this is getting lost. And, and if we, if we’re just churning stuff out because we think that’s the only way to win the, the future battle of, of AI search.

We’re mistaken.

Gini Dietrich: Yeah, 100%. So two-ish, two and a half years ago, Google came out and, and elaborated on their EEAT, which is is experience. Experience, expertise, authority, and trust. And what it essentially said was, if you can demonstrate experience and expertise in your content, that builds authority and demonstrates trust, we’re gonna rank you in search, right?

So you’re gonna go up higher in search results. Then on May 21st, they published an article or a blog post that essentially said, this is how we’re training our generative AI. So if you are not demonstrating your organization’s experience and expertise, you are not going to rank. You’re not gonna show up in generative AI search, which is where, what’s now at the top of search rankings, right?

Or on the search page. So first of all, from a communications perspective, this makes me happy because this is what comms pros should be doing any way, regardless of the tools that are available or anything, it’s, we’re building reputation, we’re building trust, we’re building authority, we’re building credibility, and we’re doing through that, through our storytelling.

So it gives us a leg up, it gives us, it gives pr pros a leg up for sure. But you can use AI to generate a draft of, of copy for sure. It’s where you add your experience and expertise that’s going to help. It’s the stories that you can tell about your subject matter experts, your employees, your organization.

It’s the case studies, it’s the testimonials. It’s the pieces that you can, that you should pull in that make it, make the content different than anybody else. Because now that signals not only to the human beings that are consuming it, but to the Google algorithms, to the search algorithms, and to generative search as well.

That this, you have something different than anybody else. And experience and expertise will always be different than your competition. So it gives you a competitive advantage every single time. Let AI help you draft it and then add in your experience and expertise. That’s where you’ll win every time.

Chip Griffin: Yeah, and, I think we also need to remember that for most of us in the content that we’re creating, either for ourselves or for clients, we would be better off not focusing on search algorithms and keywords.

Gini Dietrich: Yes. Right. Correct.

Chip Griffin: And, and how to, to convince the AI to incorporate our stuff and instead think about are we putting out quality content?

That’s right. That really speaks to our target audience. I think even more importantly, are we creating content that helps us build a relationship with our target audience. Because I, I think that one of the things that we are learning here is that as, as good as we might be at tricking AI into capturing our content or in the past, you know, convincing the algorithms or whatever.

At the end of the day, it’s having that tight network of people that we real, that are fans of our content and what we’re putting out that can help us to grow our businesses and can help our clients to grow. Now there are exceptions, and if you’re a huge global brand, you know a lot of that advice doesn’t really apply in the same way because you’re just trying to reach so many people. But the, the vast majority of us, for ourselves and for the clients we work with have a much tighter target. And so if we’re thinking about how we can build those relationships with our content, I mean, think about this.

If, if you are listening to us, we are building a relationship with you, even if this is the first time you’ve listened to this podcast. Have we used AI to help us with topics in the past? Sure. Absolutely. Yeah, it absolutely helps inform what we do. We have, we’ve used AI to generate a whole list of potential topics.

Today’s was not one of them, but it certainly could be, and you know that it makes sense. But we’re creating the content, we’re offering our own opinions. We’re doing it in our own voices, and it allows us to build a relationship with you, the listener. You need to be thinking about how you can do the same with your written content, your audio content, your video content.

If you do that, you will win no matter what happens. No matter how AI changes, no matter how the search engines change. And, and I’ve always said, if you focus on the algorithms, you’re gonna have to constantly change everything you’re doing. It’s like playing whack-a-mole. But if you focus on creating that good content for your target audience in a relationship building way, you will win no matter what.

Gini Dietrich: So this reminds me of when I started speaking, probably in 2008 ish. I had a really good friend who’s who, who spent most of his career doing keynote speeches. So that that was his career, speaking. And he said to me, when you speak, you want to tell stories in a way that will allow the listener or the attendee to put themselves in the client’s shoes.

So when you talk about like the PESO model, for instance, which didn’t exist back then, but let’s just use that as an example. If you talk about the PESO model, for instance, you talk about it from the perspective of when we worked with a client, they, they were, they thought they were using the PESO model, but what we found is they just were using tactics from each media type.

And when we worked with them, we demonstrated the value of integration and creating the multiplier effect. And you demonstrate your expertise through those client stories because now the person listening goes, oh gosh, we have that same challenge. I bet I, I should call her and see if she, her agency can help.

And so I started doing this and he was right, of course, because you know, now people go, oh, we have that same challenge. I bet she can help. Same thing with your content. If you’re doing that and you’re telling the stories and you’re helping people see how you’ve helped other organizations that are like the organization that you’re talking about, suddenly they see themselves in that same space and go, oh, they can help me with this. They’ve done it before.

And not only does it do that, but it also trains the LLMs and the algorithms to understand that this is your core expertise. And you can do this for clients too, but this is your core expertise. Now we’re gonna push them up in search rankings, so focus on the humans first.

Demonstrate that expertise and experience that you have through your storytelling, and you’ll win on search as well.

Chip Griffin: Yeah. And, and that idea of making your content relatable to your target audience, it’s really hard, at least for today, for the AI to do for you. It may well come in the future, and, and it is getting better, particularly if you use the same platform for a lot of your content because the memory is getting better.

And so if you provide it with, you know, it gives you a draft, you give it the final draft. It will remember that. And so it, it has gotten, in my experience, pretty good at recapturing some of the stuff that you said three or four months ago. Right. And bringing it back in. That will help over time. But it’s really hard for it to, you know, out of whole cloth to, to come up with relatable content that really speaks to your audience based on your experiences with clients or your own experiences yourself, because it has no way of knowing that.

Gini Dietrich: Yes. Yeah. So one of the,

Chip Griffin: I mean, unless you are already a public figure who has shared everything about your life. Its unlikely that it’s going to know all those details and even then there’s probably stuff, you know that, that you could share that that’s not out there, even if you are, you know, president of the United States or CEO of a Fortune 10 company or something like that.

Gini Dietrich: Absolutely. One of the things that we do is we, we will, we throw in case studies and testimonials and things into the project knowledge center so it it knows and I’ll say, Hey, can you find me a case study on X that will help me tell this, and then it just reminds me, you’ve got this or this, or you can talk to that.

Another thing that we’ve done is for a client, we did, we do a quarterly NPS survey for them. And part of that NPS survey, our open-ended comments, we export that into a CSV file, throw that into their GPT that we created for them and have it build the testimonials and case studies from that for us. And so it takes, you’re right, it, it starts to learn and it takes all of the information that you’ve done, but still it needs the human oversight.

Like you can’t just be like, it’s still gonna hallucinate, it’s still gonna make, I remember a couple weeks ago it was like, here’s a case study for you. And I’m like, that never happened, that sounds like a great case study, but we literally never did that. So you still have to review it. You still have to make sure that it’s telling, it’s not fibbing that it’s telling the truth.

Chip Griffin: Yeah. And, and I mean, there are ways that you can, as we’ve talked about in the past, where you can leverage the AI to make things easier for you. Examples, like you gave, I, I love using it for summarizing other things Yep. That, that I’ve done.

Gini Dietrich: Yep.

Chip Griffin: Even then, though, to your point, you gotta review it because,for example, if I ask AI to summarize this episode, I guarantee because we go off on different little tangents, right? It, it will not craft that summary in the, the exact right way. And, and I’ve seen it on some of the episodes we’ve done and the, the AI summary is just way off base because. Particularly ’cause we start with usually a tangent of some kind, because I’m trying to be witty

Gini Dietrich: Yeah.

Chip Griffin: And failing and, you know, it, it, it causes it to think that that’s actually a key point when in fact it has nothing to do with anything whatsoever. Just,

Gini Dietrich: yeah. I, I think it definitely needs oversight. The, I think let the Chicago Sun Times be your example. Let this pitch of where they sent it to thousands of reporters telling them to do better. Let that be your example of it absolutely can make you more efficient. It absolutely is a huge, huge help. It still needs to be fact checked, it still needs oversight. It still needs to be approved.

Chip Griffin: And if you’re gonna be in an arms race, at least build missiles that work.

Gini Dietrich: Yes.

Chip Griffin: There you go. What food for thought

Gini Dietrich: That is definitely food for thought.

Chip Griffin: On that note, I think we’ll draw this episode to a close before we really confuse the AI. I’m Chip Griffin.

Gini Dietrich: I’m Gini Dietrich,

Chip Griffin: and it depends.