I manage 50+ product managers at Zillow. Last weekend I sat at my kitchen table and built a working prototype in a few hours.

Not a mockup. Not a deck. A thing that runs.

AI coding tools have gotten good enough that a product leader who understands what needs to exist can actually make it exist. The gap between “I have an idea” and “here, try this” used to be weeks and an engineering team. Now it’s a Saturday morning.

What I didn’t expect

Building things myself changed how I lead my team.

When you show up with a working version instead of a spec, the conversation is completely different. Your engineers aren’t reacting to a document — they’re reacting to a thing they can poke at and break. You get to the real tradeoffs faster. The discussion shifts from “what should we build?” to “what’s wrong with this, and how do we make it production-ready?”

That’s a better starting point for everyone.

The uncomfortable part

Most of my PMs use AI tools every day. And most of them are using AI to do the same work faster — write PRDs quicker, summarize research faster, generate slides. That’s a natural starting point, but it’s not where the leverage is.

The real shift happens when AI makes execution so fast that your judgment and taste become the bottleneck, not your throughput. That’s a different kind of problem. It means the differentiator for a great PM is no longer “can you ship efficiently?” It’s “do you know what’s worth building, and can you recognize quality when you see it?”

That requires a different kind of PM — one who has strong opinions about what good looks like, who can evaluate a prototype and know instantly what’s missing, who treats building as a way of thinking rather than a task to delegate.

Closing the gap

I’m not saying PMs should become engineers. I’m saying the gap between “strategist” and “builder” is closing, and the leaders who close it themselves are going to pull their teams forward.

I’ve been doing this on nights and weekends. It’s making me a better product leader in ways I didn’t expect — not because I’m replacing my engineers, but because I can finally prototype at the speed of my own judgment.

The question I’d ask any product leader: when was the last time you built something yourself?