I spent more than twenty years in retail leadership. Teams, budgets, performance reviews, the whole apparatus of getting work done through other people. Then I built and shipped a software product with a team of approximately zero. Just me, a clear intent, and AI systems doing work that would once have needed developers, designers, analysts and copywriters.
I expected the technology to be the story. It wasn't. The story is what happens inside your own head when the constraint you have organised your entire working life around, other people, suddenly disappears.
The pace is not a bigger number. It is a different physics.
When you manage people, work has a natural rhythm. You brief on Monday, you check in on Wednesday, you review on Friday. The gaps are not waste; they are where your own thinking happens. You use the waiting to reflect, to reconsider, to prepare the next decision.
Working with AI removes the gaps. The work comes back in minutes, sometimes seconds, and it comes back finished-looking. Which means the next decision is due immediately. And the one after that. For the first few weeks this feels like a superpower. Then you notice something uncomfortable: the constraint on output was never really the team. It was the quality and speed of your decisions, and now that constraint is fully exposed, all day, with nowhere to hide.
Managing people, it turns out, had been subsidising my thinking time. Nobody tells you that.
What transfers from people management
More than I expected, and not what I expected.
The brief is everything. Twenty years of learning to delegate well, to describe outcomes rather than tasks, to give context so people can make good decisions inside the work, transfers almost perfectly. A vague brief to a human costs you a day. A vague brief to an AI costs you thirty seconds and then produces something confidently wrong at scale. The managers who were always good at briefing will be shockingly good at this. The ones who managed by hovering and correcting will struggle, because there is nothing to hover over.
Knowing what good looks like. A career of reviewing other people's work builds a quality instinct that becomes your single most valuable asset. The machine produces; you are the only judgment in the loop. If you cannot tell good from plausible, pace and scale just mean you ship mediocrity faster.
The feedback loop. Correcting course, being specific about what to change and why, iterating without ego. Same skill, compressed from weekly one-on-ones into minute-by-minute cycles.
What breaks
The absence of pushback. A good team argues with you. Someone frowns in the meeting, someone says “have we thought about,” someone quietly does it a better way than you asked for. Machines, by default, do what you say, brilliantly and immediately, including when what you said was a bad idea. The dissent that a healthy team provides for free now has to be deliberately engineered. I have learned to explicitly ask for the counter-argument, to instruct the system to attack my plan before executing it. It feels strange. It is also the single practice that has saved me from the most mistakes.
The emotional read. With people, half of management is sensing. Energy in the room, hesitation in a voice, the gap between what someone says and how they say it. I wrote recently about how none of our error-detection instincts work on machines, and the same is true of our management instincts. There is no morale to read, no body language, no gut feel. Everything you need to know is in the output itself, and you have to learn to find it there.
The pride shift. This one caught me off guard. For twenty years, the satisfaction of leadership was watching people grow, seeing someone do the thing they could not do a year ago. That is gone, and something has to replace it, or the work hollows out. What replaced it for me was craft: the quality of the thing itself, and the compounding capability of my own direction. It is a different pride. It took a while to feel legitimate.
The mindset in one line
Managing people is about getting the best out of others. Developing at pace and scale is about getting the best out of yourself, continuously, with the volume turned all the way up.
Your attention becomes the resource you manage. Your clarity becomes the bottleneck you optimise. Your standards become the only quality system in the building.
Why this matters even if you never build alone
Very few leaders will do what I did and work solo. But every leader is heading toward a hybrid version of it: teams made of people and agents together, where part of your organisation responds in minutes and never gets tired, and part of it is human and needs everything humans have always needed.
Running that hybrid well requires both skill sets at once. The old craft of leading people, fully intact. And the new craft of directing machine work: briefing with precision, engineering your own dissent, judging output without behavioural cues, and making decisions at a pace no management career has ever demanded before.
The leaders who treat the second craft as an IT topic will be managed by events. The ones who treat it as a management discipline will run the organisations everyone else wonders how to catch.