Are we getting better at managing AI while unlearning how to lead people?

This year I stopped managing people. For the first time in a very long time, there was no team, no 9AM stand-ups, no weekly 1:1s, no banter with people on the same boat.

What replaced it was managing Claude: writing detailed prompts, checking every output, correcting all the little details. At some point last week I realized that I was working in a way that would terrify me if Claude was a real person I was supposed to lead.

You see, for most of my career, my approach to #leadership was #coaching rather than directing. I gave people guidance without giving them instructions. I let them find their own way, even when that meant watching them make mistakes I could see coming because that's where the lessons are.

With Claude, I write out every step. I pre-answer questions that haven't been asked. I leave nothing open to interpretation and assumptions.

Then there's trust. With my teams, I deliberately put people outside their comfort zone, always with a safety net. I would push them to try new and uncomfortable things but they knew if something went sideways, I had their backs, and that built confidence over time.

With Claude, I have to verify and double check everything. Claude doesn't learn from doing new things because there is no comfort zone for #AI. It will just do whatever you ask it to do.

Which brings me to the next thing I miss: creative flow of ideas. Brainstorming sessions where anyone could put the rough, unfinished idea on the table. Retros where we disagreed without it becoming personal. A #culture where someone could say "I think you're wrong and here's why" as a normal part of a working week.

Now, Claude does exactly what I ask. And while that sounds efficient, the pushback I used to find challenging sometimes was also where the most interesting thinking happened.

Moreover, some of the best ideas I've witnessed came from a team member taking things somewhere I hadn't anticipated and asking a question that reframed the whole problem, making a connection I hadn't seen. 🤯

That kind of thinking tends to exhaust a micromanager, who only has room for their own views. With AI, there's no risk of being surprised. That, it turns out, is also its limitation.

And then there's the human part. Bad leaders treat emotions as a problem to manage and contain. However, it's people who send flowers when you're sick, who carry extra weight without being asked, who check in on you when you're having a bad day. Replacing human connection is the hardest part of working on your own. So far, Claude hasn't sent me any flowers.

To be clear, I'm not making an argument against AI tools. In fact, I use them every day (see the AI generated image below). Nevertheless, a few months without a team made me notice that the behaviors that AI rewards are exactly what we'd call micromanagement in a human leader.

So, here's the scary question of the week: Are we getting better at managing AI while slowly unlearning how to lead people? 🤔

---

🌟Sparking Leadership # 51: a weekly series on human-centered, sustainable #leadership. Lead with spark!