How Leaders Can Use AI to Solve Real Business Problems

**Alison Beard** (0:01)
Deal is not just another payroll platform, it's one your team might actually enjoy. HR, IT and payroll together finally, built in-house, built for peace of mind. Visit deel.com/hbr.
AI agents are suddenly everywhere you turn, but without identity, you can't really control what they do.
Okta helps you assign every agent a trusted identity, so you get the power of AI without the risk. Secure every agent. Secure any agent. Okta secures AI.

**Adi Ignatius** (0:50)
I'm Adi Ignatius.

**Alison Beard** (0:51)
I'm Alison Beard, and this is the HBR Idea Cast.

**Adi Ignatius** (1:02)
So, Allison, we talk about AI a lot, everybody talks about AI a lot, but I am interested in your take on AI right now.

**Alison Beard** (1:09)
Well, I'm completely amazed by some of the work that is capable of doing, but I think that most organizations and leaders are still really figuring out how to use it effectively in ways that move their business forward.

**Adi Ignatius** (1:22)
Yeah, I think that's right. I think too many companies are solving for AI. You know, they're pushing AI adoption wherever possible, instead of thinking about AI as a tool that can help them solve an actual problem.
So today's guest is Josh Tyrangiel, who writes about AI for the Atlantic. He's been digging deep into AI, its applications, its potential, its pitfalls, and has advice for managers on how to make it work for them. He's neither an AI evangelist nor an AI doomsayer, but rather someone who admires the technology, but realizes its limitations and risks. Tyrangiel is the author of the new book, AI for Good, How Real People Are Using Artificial Intelligence to Fix Things That Matter. Here's our conversation.
So, look, you've been immersed in the world of AI for a few years now, writing articles for The Washington Post, for The Atlantic, now with this new book. Talk about your sort of arc of understanding about AI. Maybe what you thought about it and where you are now.

**Josh Tyrangiel** (2:19)
Yeah. I mean, listen, I started reporting about it, and I wasn't thinking about its potential. I was barely thinking about the technology because in those first couple of months, the way it was presented was pretty split, right? So, as I was beginning to write what was, I guess, the first AI column for The Washington Post, I went out to the valley, and most of what anybody wanted to talk about was the kind of like, yes, this stuff is magic, everyone can see that, but we're very divided. And half of the people I spoke to were like, we have to move because this stuff is amazing and it's going to cure cancer and mitigate climate change. And then on the other side, the other do-mers were very much like, come with me if you want to live. What I realized a couple months in was like, oh, I'm in like a classic spin cycle. And everybody out here is a little bit drunk. And it was only a couple months into the coverage where I was able to take a deep breath and sort of focus on the tech.
And I've sort of been in the same place since then, which is, yeah, this is pretty dazzling. You can say what you want about the technologists, about how they're comporting themselves in the public space.
Tech's pretty incredible. It's not perfect. It's not going to roll itself out, but this is a significant step up from what we've been dealing with in software.

**Adi Ignatius** (3:38)
What's your advice on how executives in particular can follow the script? Where there is this sense of, oh my God, I have to do everything, but oh my God, it's back and forth, it's optimistic, it's pessimistic. How do you follow the script here?

**Josh Tyrangiel** (3:53)
I think it goes back to this sense of drunkenness, which is that we're getting slammed with marketing and advertising and reports about the need to shift to AI. As with all hype, there's a wave of truth in there. I have been in a bunch of places that have done sometimes really good, sometimes flawed implementations. And the number one thing that comes back for the successful implementations is, hey, did you know what problem you were trying to solve to begin with?
And did you have realistic expectations of how you were going to solve it? And what I mean by that is, I think that if you watch the ads on Sundays during football, so much of it is trying to convince someone with big purse strings at a company that the AI is magic, that you make your purchase, you get your license, the software rolls itself out into the enterprise, and all of a sudden, you're way more productive and much more efficient, and maybe even able to cut costs. And that just isn't right, okay? The software is really, really good, but it requires a scalpel to understand the problem you're trying to solve, and it also quite critically requires human beings, like really talented and a very specific phenotype of human who's able to work with both your system, whatever that system may be, and with the software. The software engineers and makers are not concerned about your problem. You actually need to take a deep breath, and kind of stick to fundamentals, which is not what you're hearing in the hype cycle.