How agency owners can use AI as an always-on thought partner

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 we need to talk to the robots today.

Gini Dietrich: Yes, I love talking to the robots. Let’s do that.

Chip Griffin: It’s a robot future, and we just, we need to, we need to figure out, I don’t know, some new ideas or something like that.

Maybe we should have a conversation with our friendly neighborhood robot.

Gini Dietrich: I like it. Let’s do that. You-

Chip Griffin: Actually, that would be a, that would be a good episode at some point to actually-

Gini Dietrich: It would be a, yeah …

Chip Griffin: to, you know, we could have our first guest. We could have, like, Claude as our first guest on the show.

Gini Dietrich: I love it. We should do that. That’s a good idea.

Chip Griffin: And, and, and see how that goes.

Gini Dietrich: Yeah, let’s.

Chip Griffin: I don’t know. Maybe that’s not a good idea. Who knows? Anyway, we are going to talk about AI today, ’cause we haven’t talked about AI at all lately. Nope. And we won’t be talking about it soon when we have the Saga AI survey results to talk about, hopefully on the next episode. But we thought we would talk today about AI in more practical terms because we’ve done a lot of talking about AI in, in sort of high-level strategic ways and how important it is to agencies and how we need to be thinking about it and integrating it and thinking about the costs of it and all of that kind of stuff.

But I think it’s helpful for us to have some conversations with and for our listeners about some practical uses of AI that, that we’ve used, that we’ve come across, that we use ourselves- Yep … in order to, to get the maximum value out of this new technology. And, a good place to start is how do you use whatever platform of choice you have, or maybe multiple platforms of choice, to help you as a thought partner to not just, you know, write things and spew stuff out more quickly or something like that, but really to hone ideas, to get advice, to have someone to bounce things off of.

And I mean someone in quotes here because, yes, I know it’s not human, okay? So don’t- … don’t send me notes about how, “You know these things aren’t real.” I know that, okay?

Gini Dietrich: I do know that. I got it. Yes. I am aware.

Chip Griffin: So-

Gini Dietrich: Yeah, I think it’s, I think it’s such a… First of all, it’s a good topic, and I think it’s something that’s fun for the two of us to talk about because I think we’re both, like, full-on in.

For me, when I realized it could be a thought partner for me, it was two and a half years ago, and a client came to me, and they came to me specifically, and they said, “We would really like to understand what the PESO model maturity ladder looks like, especially inside an organization like ours.”

And, at the time, I had an idea of what the maturity model looked like just generally, right? But being able to apply it to a really specific situation and a really specific organization and really specific brands, I hadn’t thought through yet. And of course, I can’t put the client’s information in there, right?

But I can think about, I could… I, at the time I was like, “Okay, so how do I start to think about this with the constraints that I know they have and sort of how they run PESO now, which isn’t in a true operating system, but more sort of pick and choose tactics.” So I went into, at the time, ChatGPT, and I think Claude does a better work of this now, but I went into ChatGPT and I started asking questions.

“So if you had to create… First of all, if you had to create a PESO model maturity ladder based on these seven sort of levels that I had already thought through, how would you do that?” And we just went back and forth, and we asked each other questions. And it would, it, it came up with some things where I was like, “Huh, hadn’t considered that.”

So then I would sort of put those over to the side and we would continue. And then I said, “Okay, great. Now here are some some constraints, right? We know, we know it has these, the organization has these constraints. We know that it takes, you know, six, six to eight months to be able to do anything, like all of this stuff. How would you change it based on that?”

And so we went back and forth on that. Now, granted, it took probably six weeks for me to get something usable to be able to take to the client, but I wouldn’t have been able to come up with that on my own. And I don’t think that even conversations with my team, we would’ve been able to come up with the, all of that on our own.

And so it really helped me just kind of think through some things that I hadn’t considered, some things I probably wouldn’t have considered, and it also helped poke some holes. So then I said, “Okay, great. Here’s what I’m thinking. Poke holes in it.” And it was like, “This doesn’t work, this doesn’t work, this doesn’t work.”

And so it just helps you… It was at that point, which I think was two and a half years ago, it was, it’ll be three years in August, really think it, think through sort of beginning to end that I wouldn’t have been able to do on my own.

Chip Griffin: Yeah, I mean, I think that’s a great discussion of your evolution on that, and in my case, I was, I was later to the AI party in terms of In-depth use of it.

Gini Dietrich: Mm-hmm

Chip Griffin: I, I was very early on to kick the tires, which may have been to my detriment, right?

Because the very early incarnations of a lot of these tools was not the best.

Gini Dietrich: Not good.

Chip Griffin: And, so I, you know, I certainly was, you know, spending a fair bit of time using it for, you know, the basics. You know, some basic writing and editing, some basic image creation. Certainly, you know, transcription, speech-to-text, those kinds of things.

But a lot of that, but not really as much in the in-depth strategic areas-

Gini Dietrich: Sure, sure …

Chip Griffin: until probably a little over a year ago when I started to realize that there had been this shift and that it, it was, at least to me, a lot more usable. And, really just, you know, opening my mind up to what you could actually do with the tools beyond the simple use cases.

Gini Dietrich: Yep.

Chip Griffin: And I think the more time you spend with it, the more you realize just how helpful it can be. And, and particularly in, you know, small owner-led agencies, the… You know, we’ve all sat there and we’ve had an idea at, you know, 8 o’clock on a Friday night or over the weekend, and we’re like, “God, we, I wish we could…You know, I can’t bother the team with this right now though.” Yep, yep. “But I kinda, I kinda wanna continue thinking this through.” Yep. And so you now have this- Yep … you know- Thought partner … always on thought partner- Yep … that, you know, when that idea comes to you, when you’re watching some Law & Order rerun or whatever, you’re like, “Hey, you know, I just had this idea and, you know, what do you think of it?”

And, it feels weird the first few times you do it, this, to sit there and say to, you know, the chatbot, “Hey, what, what do you think of this idea?” And, but you learn so much from it because it is, it is able to ask questions, to poke holes, to find gaps. And so if you really treat it as, I don’t wanna say an equal because it’s not quite an equal with you.

I mean, you still are the decision maker and you need to remember that you’re always retaining that. But someone who can hold their own in a conversation with you. Yep. Once you accept that, you can get so much from these tools. And it’s not like, you know, two years ago, three years ago when it said, you know, everything is, “Oh, we love you, Chip. This is… You’re brilliant.” I mean, there’s still, there’s still a little bit too much of that for my taste in there, although I’ve coached it into, you know, all of the tools have instructions for me to, you know, knock that off as much as possible. You know, you don’t want it to be fully contrarian where it just disagrees with everything, ’cause you can easily turn your chatbot into basically it just will say the opposite if you, if you go too hard with your instructions to, to not say nice things.

But you want to get that, to that point where you’re able to just bounce these things around and say, “Okay, here’s my plan. What am I missing?” And that gap analysis is something that the AI tools do exceptionally well.

And part of it is just, you know, making yourself vulnerable to it and allowing you… I mean, and you have to remember that, I mean, that is the best thing about these tools.

It’s not human. It’s not judging you ’cause it doesn’t care. Right. It does not care. You know, this is not like having a conversation with one of your team members and they just sit there and they’re like, “Oh, no. Oh, Chip just really doesn’t get it. Why is he, why is he asking me this question?”

Because the tools, they really don’t care. And so it is a great place to make yourself more vulnerable and throw some things out there and see what works or what doesn’t. And, and there’s really no limit to what you can ask of it. You obviously have to judge everything that comes back. Not everything is going to be useful or correct.

But then, but push back and then say, “Okay, well, you know, I hear what you’re saying, but here’s why that doesn’t sound like a good idea to me.” And sometimes, in my experience, the tools will say, “Yep, okay, you’ve got a good point there,” or it’ll say, “No, I, you know, here’s why you should, you know, reconsider your point of view on it.”

And if you’re not seeing that, then you probably need to adjust the instructions that you’re giving to make sure- Yes … that it is open to having a bit more of a substantive dialogue with you. But once you do that, you really have something here that can help you to navigate client strategy, business strategy.

It can help you with how to handle HR issues. It’s not a substitute for legal advice or proper HR consultants and all that kind of stuff at this point, but it can help get you a lot of the way there if you’re willing to turn to it and say, Hey, you know, this should be my first port of call in a lot of cases and not the last resort.

You know, the other, the other way that you can use it as a thought partner on some of those things too is say, “Okay, you know, here’s, here’s what I’m thinking about or here’s what I need to solve. Ask me questions.”

Gini Dietrich: Yep.

Chip Griffin: And, and you know, we all think of this as we’re asking questions of the AI tools to give us the answers, but it’s also very good at asking you questions and- Yes … and drawing things out. Now, it, it does often need to be encouraged to ask questions.

Gini Dietrich: It does, yeah.

Chip Griffin: The, they are…

The, most of the platforms are starting to get better at having the, their tools pause and ask questions for clarification, but that’s not quite as good as actually saying, “Hey, you know, ask me questions about this so that you can, you can help me think this through.” And it will do that for you and then do follow-up questions and that sort of thing if, again, as long as you’re telling it that’s what you want.

And that can help as well because, you know, we’re,.. Even with our own teams, a lot of them won’t ask us questions, right? If you’ve got an employee- ’cause they know you sign the paycheck, and we’ve talked about this many times on the show. It’s really hard to get candid – not just candid advice, but candid questions.

They’re not gonna push you in the same way- Yep … the tool will. ‘Cause again, the tool doesn’t care.

Gini Dietrich: Yep.

Chip Griffin: It doesn’t care- Yep … if it hurts your feelings.

Gini Dietrich: Yep. And it knows it, it can’t get fired. Like – right? So- Right … it doesn’t have the same experience. I mean, it’s not a sentient being, so it’s not it does, just doesn’t care, and it doesn’t judge you.

I think that’s the other thing too. So I think there are so many different opportunities here for you to use it as a thought partner. I mean, the way that I have mine all set up is to do those kinds of things, and it’s to say, Okay- Well, wait a sec, I was thinking about it this way, what do you think about that?

Or how do you think about this? Or, you know. And again, ask me questions because trying to figure this out. And I put all sorts of things in there, all sorts of things. I’ll say, “There’s something about this document that bothers me, and I can’t quite put my finger on it.” And it’ll say, “It bothers you because this, this, and this.”

And I’m like, “Yeah, that’s right.” Or, you know, one of the things I think we’re starting to find is that the design work you can get from the AI is pretty darn good. But you have to be… With a human, you don’t have to be ultra specific about what you’re looking for because they’re creative. But with the AI, you have to be ultra specific about what you’re looking for.

So I will say, “Here’s what I’m trying to accomplish. Here’s, here are our brand colors,” blah, blah, blah. “This is the vision. How do I describe that in a creative brief?” And I almost have it write its own creative brief, and then I feed that back to it and say, and make some changes, and then say, “Here’s what I’m trying to do.”

And I get far better creative design from it when I do it that way, and it’s created its own creative brief. So I think there are lots of ways that you can use it in interesting ways that you’re probably not considering right now.

Chip Griffin: Absolutely. And, and I think, you know, maybe a future episode’s a good time to talk about, you know, the specifics of how you use planning and context and providing information to get better results, because I think that’s, that’s an area where a lot of people haven’t invested enough and don’t invest enough, and it’s really just kinda fire and forget.

And, and that tends- Yeah … to be where you get the worst results.

Gini Dietrich: Yeah. Yeah, yeah.

Chip Griffin: And, to your point, if you’re, if you’re sitting there and providing it with the information that it needs, I mean, simple things like what are my brand colors? What’s my brand philosophy? Yep. You know, what, what’s my own personal risk tolerance?

I mean, you- Yep … you wanna make sure that you’re stockpiling that kind of information because otherwise you’re not gonna get useful feedback or useful questions from it if it’s assuming that you are one way and you are not. It’s the same as anybody else. If you… I mean, we’ve talked about this before, you really have to sort of start to act like these tools are human- Yep

in that you can’t just expect them to run from a standing start and have a good result. Yeah. They really do need good guidance from you. Yes.

I think the other thing we do, we do always need to be careful about with these tools is that, that they are, they do somewhat, you know, uh, float with the winds, if you will.

And so they are not always consistent- Mm-hmm, mm-hmm … in what they might recommend. And so- Yep … so you do have to accept a little bit of that, that it’s, it’s kind of like having that employee who, you know, one day they’re very cheery and thinks this is the right direction, and the next day they’re kind of like, “Eh, nah.”

So, you always need to keep yourself in the driver’s seat when you’re using it as a thought partner. And I know I’ve certainly had situations where, you know, I’ve fed it some things, and it’s come back with a recommended strategy, and then a couple of weeks later I, you know, I ask something about the strategy as if I’ve adopted it, and they say, “Well, I don’t think that’s a good strategy.”

Ah. But it was, it was crafted by you.

Gini Dietrich: It was your idea. Yes.

Chip Griffin: But, so then when you dig into it, you can sit there and say, “Okay, well, I, you know, I see why it said it the first time because it backed it up, and I can see why it’s saying it this time. And so now my job is to reconcile that and decide which one I want to lean into- Yep … more than the other. Yep. I, I mean, you… And, and whenever you’re using it as a thought partner, it’s always a good idea to ask for supporting information for whatever it’s suggesting. Not to… And, usually they’re pretty good. I mean, actually my issue used to be that it provided too much, that particularly ChatGPT just loved to just spit out like a whole research paper for it.

Right, right. Like, ” I think that your brand color should be green,” and then it’s like- Right … 200,000 words on why green is the right choice. Yes. I’m like- I don’t- … “Dude, just tell me green.” Yeah, yes. Okay? Like- Yes … if I want that detail- 100% … I’ll ask for it. Yes. So they are certainly more balanced than they used to be.

But they will still generally provide backup. But if not, ask why, and that’s when it will start to provide, you know, either citations or facts or at least something around the reasoning of it so that you can then judge, “Okay, that does really make sense to me,” or, “No, that, that doesn’t really fit for what I want because…”

Yep. And then always tell it back to it. Like, you can’t just, you can’t just get something back and say, “Oh, I don’t like this,” and then move on, close the chat, and go to another one. You’ve got to have that back and forth- Yeah … so that the tools can learn from what you are thinking and, and what you are feeling from the responses that they’re giving.

Gini Dietrich: Yeah, I mean, there have been times where it’s late at night, and I’m still going, and finally I’m like, “I can’t…” I will literally say to it, “I’m about to die. My eyes hurt because I’m so tired. I have to go to bed.” It’ll be like, “Okay, good night, Gini. Sleep tight.”

Chip Griffin: I can’t claim to be having conversations quite like that with the tools. That’s not quite my style. But to each their own. All right, well, I think with that we probably should say good night to our listeners now, and we will draw this episode of the Agency Leadership Podcast to a close.

I’m Chip Griffin.

Gini Dietrich: I’m Gini Dietrich.

Chip Griffin: And it depends.