This short transcript captures Jason M. Lemkin discussing why so-called "hero" AI purchases—large, attention-grabbing buys intended to showcase AI—often fail in organizations. He argues CMOs often pick conspicuous projects they're already reasonably good at, making training/onboarding hard for a first agent and yielding limited gains; instead he recommends deploying AI where the organization previously wasn't solving a problem to secure an initial win. He also notes timing considerations (e.g., some improvements may be better suited for 2026 rather than 2025).
Chapters
Yep.
And that's when you look at AI and you're like, oh my gosh, like, I could just do so much better.
It is. But I think the failure that I see, especially CMOS make, is they're looking for a hero purchase. They want to go to their boss, to the CEO or whoever and say, I bought AI.
Yeah, I think everybody right now wants that hero purchase.
And listen, I'm not saying you shouldn't save your job, but I guarantee it's probably going to fail because that hero purchase is probably going to be aligned with something you're decent at. The bar is so high and you're not going to know how to train it because it's your first agent. So you're looking for, you're solving medium high hanging fruit because it's already working and you've never done it before and you don't know how to train and onboard it. I know you're going to get that. There are a bunch of apps we know that have done really well as the hero apps and maybe buy it. It's just a little bit of money. But I would not deploy it first. I would leave it on the shelf and deploy something where your org just isn't doing it. Just it's not happening. Right. And then you get a win. Right. But I just, you know, some of these hero apps, we just watch them fail again and again. And I, and I hadn't fully reflected until. One of the reasons is that it's not necessarily the most broken problem in their org. Right. Just making. If AI is only going to make something 20% better, maybe 2025 is not the year.
Right.
Maybe that's a 2026 project. So this is about, this is about picking.
Everybody, I think has started. I think like we're sitting here in the fall of 2025. Very few people are like, I haven't even thought about this. There's probably been some experimentation like there has at your organization throughout the course of the year. But as all of this is coming together, you said something on a recent interview that was about, you know, what's happening to SaaS and that's what people are kind of wondering. Is SAS dead? Is like kind of the, the mega theme of the moment right now and how are AI agents changing all that?
