Builders, Not Roles

For a long time the product manager job came with a flattering lie baked in: you're the CEO of the product. In reality it usually meant you wrote documents, ran meetings, and acted as a human router, moving context between people who didn't talk to each other directly. That version of the role is quietly dying, and AI is the thing eating it.

The parts of the PM job that used to justify an entire career ladder are exactly the parts AI is very good at consuming. Summarizing customer calls. Turning research into a tidy theme. Drafting the PRD. Writing the ticket. Generating "here's what's going on" in minutes from a pile of inputs. When a system can sit in on every call, read every ticket, and produce a competent synthesis on demand, the value of "I summarized discovery" collapses. That work doesn't disappear, it just stops being a job.

What's left is the part that was always hard, and always the actual point. Deciding which intent matters most right now. Seeing where user pain, business reality, and technical leverage actually intersect. Holding flows, systems, and consequences in your head at once. That's not administration. That's judgment, and systems design, and it doesn't belong to a title. It belongs to a behavior.

I call the behavior product building, and the person who does it a Product Builder. Not a new title to slot into the org chart. A description of how someone works. A Product Builder starts in the user's world and ends at a working system. They spend time with raw signal, calls, tickets, field visits, usage patterns, and turn it into a clear statement of what the user is actually trying to do. Then they work shoulder to shoulder with design and engineering from the beginning, with no theatrical handoffs, no disappearing for a week to reappear and critique. They think in systems rather than one-off features, and they define success in the user's language first and metrics second.

The way most organizations split product and design into separate lanes no longer matches the work. Product "owned" the problem and the requirements. Design "owned" the experience and the flows. It looked reasonable on an org chart and created a nasty gap in practice, between how a problem was framed and how it was actually shaped. The job that decides whether a product is any good cuts straight through that boundary. It doesn't care what your title is. It cares whether someone is holding the whole thing.

Most designers won't say this out loud in front of their PMs, but many of them already live closer to the real job. When no one else is protecting the user from the system, they're the ones walking through edge cases until everyone's exhausted, asking "what happens next" long after everyone wants to stop. The good ones already behave like builders. The same is true of engineers who can talk in user stories instead of just tickets. The convergence isn't a threat to design and engineering. It's a recognition that the people who can think end to end were always the ones doing the load-bearing work.

No company will send a memo announcing that the PM role is dead. It will look quieter than that. The PM who exists only in roadmaps, tickets, and status updates slowly stops getting invited to the rooms where direction actually changes. The designer who can think in systems and the engineer who can speak in user outcomes get pulled into those rooms more often. Tools absorb more of the coordination and documentation that used to fill a calendar. And underneath, the real role in the middle changes into something else: someone who lives at the intersection of user reality, system behavior, and business outcome, and is genuinely competent in all three.

This is uncomfortable, because it removes the hiding places. When AI clears away the busywork that used to justify the role, what's left is the part that was always hard and never optional, and you no longer get to blame the process for bad outcomes. The process is finally out of the way. What remains is whether you can think clearly about what matters, hold the whole problem, and stay in the work from the user's first complaint all the way to the behavior changing in production.

That's not a downgrade for anyone who got into this to build things people actually use. It's the job finally being about the job. The title matters less every year. The behavior matters more.

This is one of the central shifts in my book, Intent: How to Build Products That Last in the AI Era.