# AI and the future of management

- **Episode:** [217](https://agents.truthliesandwork.com/episodes/217.md) (2025-07-24)
- **Type:** interview
- **Tags:** ai, future, management
- **Time:** 38:33–42:43 (start 2313s)
- **Listen:** [https://truthliesandwork.com/217?t=2313](https://truthliesandwork.com/217?t=2313)

---

[38:33] Leanne Elliott: AI. We cannot have any conversation about the future of management without talking about AI. Here's Andrew's take on it
[38:43] Al Elliott: You mentioned earlier that Daniel was both simultaneously terrified and fascinated by AI. I want to hear what you think about AI. Is it a good thing? Is it a bad thing? How's it going to change culture, teams, management?
[38:54] Andrew Palmer: I find that really hard to answer, Al. I, you know, The sort of ranges of possible outcomes are so big and the tech is moving so fast that I'm not sure that I could give you a really sensible answer. I could tell you the things that I worry about, which is that the incentives are strongly to look for efficiencies and cost savings, and particularly in that middle management layer that we spoke about. Earlier. So you could well imagine, and we are seeing it in some of the tech giants already and some of the language that's coming out of Duolingo, Amazon, Klarna, and others, that there's a sort of layer of what is being labeled as bureaucratic fat that can be slimmed down quite radically if you've got AI doing certain things. I would be really worried about that, partly because it sort of just reinforces an existing problem, which is that managers already often have to do another job as well as manage. So they're kind of already, so by slimming them, you're not solving that problem. But also some of the kind of behaviors which are bad are likely, I think, to be reinforced. So I don't, AI summaries of meetings are in theory good. In practice, if that's going to be used for performance management as a way of kind of like, what's the level of engagement? What percentage of time did you spend talking during this meeting? And that's somehow going to feed into a dashboard. We're going to have even more people talking for no good purpose in meetings in order to appear on a dashboard. So there's all those sort of unintended consequences from quantification, which sort of just worry me. I mean, you definitely want to know if people are unhappy, dissatisfied with their work, not engaged. And I don't think that 6-monthly surveys are the answer, but one-to-one meetings is the most effective way to get at how people are talking. And it's as opposed to occasional surveys or always-on panopticons that are sort of trying to work out your emotional state at any time. So I think there is quite a kind of well-defined and effective way of working out what your team is feeling, but it does require you to meet them. So, and another problem with AI is that incentive to stay at your desk and not to talk to your team is going to go up. If you can just look at a dashboard and it's going to tell you, oh, well, team engagement is okay, it's like in a range, fine, I'll just sit here and not go out. And I'll come back to Toyota because they used this phrase which was really new to me and striking, which is they talked about the right to digitize. And they're very, very keen on something the Japanese is genchi genbutsu, which is basically sort of, it's the idea of management by walking around. You go to the place and if you can sit at your desk and the data comes to you, the incentives to go out, talk to workers, talk to employees, see what's actually happening, um, is muted. And I could see AI doing that like in a massive way. You know, why would I ever talk to anyone?