Back to Blog

Saying No in the AI Era

A PM friend told me something last week that stuck with me.

"We shipped 14 features in 6 weeks. Nine of them shouldn't exist."

This is the new reality of product management in 2026. With agentic coding tools, engineers can go from idea to working prototype in an afternoon. Features that used to sit in the backlog for months, getting pressure-tested and vetted, are now live before anyone asks, "But should we build this?"

InfoWorld recently called this the biggest shift in the PM role in a decade. And I think they're right.

The Old Playbook Is Gone

The old PM playbook was about fighting for resources. Convincing leadership your feature deserved a spot in the sprint. The backlog was your bottleneck, and it forced proper due diligence.

AI just removed that bottleneck entirely.

So What Replaces It?

The best teams I'm reading about are adopting "kill reviews," basically a 30-day post-launch check where if a feature hasn't moved a core metric, it gets retired. No debates. Just data.

One team cut 6 features after their first kill review. Their NPS went up 12 points.

The Inversion

This is a fascinating inversion for anyone studying product strategy right now. The scarce resource used to be engineering time. Now it's user attention. The PMs who win in this era won't be the ones who ship the most. They'll be the ones with the shortest feature lists and the happiest users.

The hard part used to be building. Now the hard part is restraint.