Greg Kihlstrom, Principal of The Agile Brand, shares why starting small is the smartest way to begin your AI journey, how leaders can balance innovation with guardrails, and what happens when unchecked tech sprawl creates more problems than it solves.
Greg Kihlstrom, Principal of The Agile Brand, joins the show to talk about the practical side of AI adoption in MarTech.
Greg and Sara dig into the tension between innovation and operational guardrails, the risks of tool sprawl when “anyone can build an app,” and why proactive leadership is essential to keep experimentation from turning into chaos. Greg also offers clear advice for cautious adopters: start with a small, meaningful use case and be ready to fail a couple of times—because those early failures are how you build long-term success.
Greg:
“Pick a small [AI] use case and follow through with it. If you're the kind of person that's reluctant to do something that might fail, try to be okay with getting it wrong the first 2 or 3 times. Those are good muscles to build regardless.”
Links & Resources:
Timestamps:
00:00 - Introduction to Greg
00:32 - AI Hype and Realities
02:20 - Effective AI Implementations
05:01 - Challenges of Democratizing AI
09:21 - Advice for AI Beginners
11:08 - Final Thoughts
#martech #AIinMarketing #contentstrategy #datamanagement #progress
0:00:00.2 Sara Faatz: I'm Sara Faatz, and I lead Community and Awareness at Progress. This is 10-Minute MarTech.
0:00:05.2 Greg Kihlstrom: What happens when everybody built their own app to do their own thing and now that person quits, is fired, whatever? What do you do? I've given it a little bit of thought, but that's coming.
0:00:17.2 Sara Faatz: That's Greg Kihlstrom, principal at The Agile Brand. Let's get started. So Greg, let's get started. What's keeping you up at night right now?
0:00:34.0 Greg Kihlstrom: You know, with all the talk about AI and let's call it energy around it, which I think, you know, mostly positive. I mean, there's plenty of hype and I think there's been some empty promises and there's been research showing, you know, not as good return on investment as some may have hoped. But with all of that, what kind of keeps me up at night is making the most of those ideas. There's some amazing platforms out there. There's some amazing opportunities to make the most of it and to not even replace the humans in the mix, but to augment and actually elevate the role of people. But to do that, you've got to have good data. You've got to have connected data. You've got to have teams working together in ways that they're just not used to working together. So the connections are kind of what I turn over in my head.
0:01:26.0 Sara Faatz: What do you see, in your opinion, and what are some of the things that maybe AI has been hyped up more than it really is?
0:01:33.7 Greg Kihlstrom: There's been a lot of hope that it can do a lot more autonomously than it can. And, you know, even with agentic AI, it still requires a human in the loop and it benefits from a human in the loop. It's not even just that it needs it. We actually benefit when we're inserted at strategic points in the process, even if it's like semi-autonomous or whatever. And so, you know, I think this idea of like set it and forget it. Well, you know, you can do that with AI generated content, but I think we all know what that gives us because we probably read it all day long on Twitter, LinkedIn, whatever. It's at best stuff we've already seen before. And at worst, it's kind of garbage. So that's the outcome of that approach of just like, oh, okay, well, AI, I'll just outsource it to AI. That's not really how it works.
0:02:20.7 Sara Faatz: What are some of the best implementations of AI you've seen from a MarTech perspective?
0:02:25.4 Greg Kihlstrom: I think still the small use cases and just teach a team that, yes, they can really benefit from this, that it's not a threat to at least most of their jobs. When I say most of their jobs, I mean not their entire role, even it's like it's probably a threat to the stuff they didn't want to be doing anyway. I think just coming up with simple use cases, I've heard arguments to the contrary of like, don't just think about it as an efficiency play. And I don't think you should only think of it as an efficiency play. But I think you maybe start there because it's easy for everyone to see, okay, well, I can save 10 % of my day by just automating these things or by having AI do this thing. Again, took me away from things that a human actually brings value to the table. And so I think there's a lot of people that are reluctant because for whatever reason, they have anxiety about pushing the wrong button. So I think that's where it's also standardizing approaches.
0:03:26.7 Greg Kihlstrom: You've got to explore, you've got to go. You actually want some of those people that are going to go out there and find the new thing, even if it doesn't pass the company guidelines. You want some people to go out there and explore, even if it's on their own time, just for fun. But then you actually want a leader to come back and say, okay, well, it's great that there's all these things out there. Here's how we're going to do it and here's how we recommend doing it. Because both sides of that equation need some guardrails once it's actually operationalized.
0:03:55.0 Sara Faatz: Right. We talked to one person recently who said, I'm telling everybody, throw out the baby with the bathwater. Definitely throw, you know, do start from scratch. And I think that there is there's a lot to be said for that. There's also a lot to be said for what you've just suggested, which is more kind of meet in the middle, have some people who say, let's throw out the baby with the bathwater and other people who are saying, oh, hold on just a little bit. You know, and taking both sides of that can make for a really powerful AI movement.
0:04:20.1 Greg Kihlstrom: I mean, it always depends. I mean, I'm a good consultant, so I always say it depends. But whether it's the size of the org or the type of the team or just the use case that you're trying to solve, sometimes, again, a clean slate may be what's what's needed. But generally speaking, I'm working with enterprise organizations for the most part. Like, it's hard to throw any that that's the like tech debt or even culture or operational debt is kind of the thing that's that everyone's burdened with. So it's hard to throw things away. So it's, you know, small, incremental change, but directed by a strong leader that has a vision for it, I think, is what's what's needed.
0:05:01.2 Sara Faatz: You mentioned tech debt. And when you think about AI and the ability for people to create quickly, but there's also, you know, I think there's a potential, I think three to five years from now, we're going to be talking about this technical debt that we now have because we had people who were empowered to build but didn't necessarily have the understanding of, okay, when we need to upgrade, when we need to modernize, when we have complex integrations, how do we change and how do we move forward with that? So it'll be really interesting, in my opinion, to see where the industry goes and how do we how do we take what's really an interesting or compelling productivity gain and make sure that it's hate to use this phrase, but future proof, for lack of a better word. Right.
0:05:44.6 Greg Kihlstrom: Yeah. I mean, you bring up a great point because, you know, democratizing data democratization is kind of the where some of this stuff started. But you're talking about vibe coding. It's like, you know, I've played around with those things. Like, I've built like five apps in the last, you know, few weeks just like playing around with stuff. And so like what happens when everybody can just make a tool? Like right now, those tools are, I love them, but they're still a little, you know, they're a little clunky. But like give them another year and it'll be as easy as using ChatGPT or whatever. And so now anybody can write an app, just like now anybody can kind of make a dashboard of some of sorts. Now anybody can make an app. What happens with, you know, you talk about technical debt, I don't even know what that, it's just sprawl, I think, is what you get. And so what happens when everybody built their own app to do their own thing and now that person quits, is fired, whatever, like what do you do? You know, that's, I haven't even, I mean, I've given it a little bit of thought, but that's coming.
0:06:47.0 Sara Faatz: Yeah, and who's commenting that code? Who's doing, you know, all of those things? And again, when you think about integrations and having an API-first mindset, if you're not, if you've just been given this power, right, the great power comes great responsibility, right? You have this power to build these apps and to do these things. But yeah, I mean, I think it's going to be really interesting to see, can we use those same tools to modernize from there? Or does the role of the technologist all of a sudden become elevated again to say, okay, we have all of these things, now we need to find a way to make it production ready and meeting corporate standards and user standards. I mean, your end user is becoming that much more sophisticated as well, and their expectations are much higher than ever before.
0:07:30.6 Greg Kihlstrom: On the positive side, I mean, you know, definitely there's some things to avoid, but I think on the positive side, I will say, you know, I come from kind of a marketing ops background as well. So, you know, I consider myself fairly, I'd like to think through processes and all that kind of stuff. But I will say like using, whether it's the vibe coding tools or even, you know, prompting, you know, ChatGPT, Claude, whatever, it has forced me to think through things in a depth that I often, not everything, you know, some of those things I was already thinking that way, but like it forces you or else you just get terrible results. So like if you're going to get good results, it forces you to think three steps ahead or whatever, which I think is something that not everyone like naturally does. So I think that's a potential positive.
0:08:16.0 Sara Faatz: Oh, I could not agree more. You know, I think we're going to have a combination of some technical debt and some what I have kind of lovingly being referred to as digital waste, right? Especially when you think of some of the early implementations, people were using AI to create content. Well, that's just a regurgitation or a refactoring of existing thoughts and ideas, right? You know, I think a lot of people have that conversation that, hey, we don't necessarily need to use it for creating new content. But I think it's exciting. And it's also, you know, I think those are just a couple of things that we need to be thinking about and cautious of, perhaps.
0:08:50.5 Greg Kihlstrom: Yeah, I guess it's comfort to me as a human that it's not coming up with like bold new ideas. But, you know, you could make the argument, I guess, that there's nothing new under the sun as well. But like humans still have the edge when it comes to newer ideas or even synthesis of existing ideas. And, you know, if we think of it that way, then, you know, AI can be a good partner to us in helping us think through the synthesis of those ideas and things like that.
0:09:21.3 Sara Faatz: What piece of advice would you have for somebody who's just now saying, okay, I've been nervous about this. Maybe they're more on the cautious side, right? What piece of advice would you have for them as far as getting started on their AI journey?
0:09:33.8 Greg Kihlstrom: Yeah, just pick a small use case and follow through with it. And don't be afraid. If you're the kind of person that's reluctant to do something that might fail, like try to be okay with getting it wrong the first two or three times. But pick something really small that would be meaningful enough to help you, but not is going to take you three months to figure out if it worked or not. Like that's just, you know, you'll never get through it. So, you know, start small, anticipate that you're going to fail twice or something like that and build that because those are good muscles to build regardless.
0:10:09.5 Sara Faatz: Sure. Yeah. Who do you follow right now for inspiration or information?
0:10:13.5 Greg Kihlstrom: Yeah, that's a great question. I mean, I just read and listen a lot to a lot of things. So, I mean, you know, some of the publications I write for, I also read. So, you know, CMSWire, Martech, some of those as well. And the Marketing AI Institute folks, they're, you know, always a good resource for me as well.
0:10:34.5 Sara Faatz: And do you have a Martech hot take right now?
0:10:36.7 Greg Kihlstrom: I think in a year from now, we're going to be talking a lot more. Like right now it's about agentic from a business standpoint. I think we're going to be talking about consumer agents. I think a lot of people are not going to be ready to talk about it. Just like there's not a lot of use cases around agentic internally. I think the divide between leaders and laggards is just going to keep growing and growing and growing. And but, you know, just to get ahead of that somehow.
0:11:07.4 Sara Faatz: Awesome. Well, Greg, thank you so much. I really appreciate your time today.
0:11:11.3 Greg Kihlstrom: Yeah, thanks so much.
0:11:13.0 Sara Faatz: Listeners, thanks for tuning in. Make sure to like and subscribe wherever you get your podcasts. Until next time, I'm Sara Faatz, and this is 10 Minute Martech.