Evals are the new PRD: How AI PMs actually build products
Registration
Past Event
About Event
In today’s world, AI PMs don’t build with specs, they build with evals. Evals are the new PRDs: they define “good,” catch regressions, and keep model iteration grounded in reality instead of doc debates.
The hard part is not writing a perfect eval set once, it’s building the flywheel that continually turns real user data into better, broader evals that improve the product over time.
In this online event, you’ll learn how the best AI PMs do exactly that.