Product Discovery with AI

In this 60-minute session, Andreas Stephan, freelance agentic product manager and trainer at hackers&wizards and the Digitale Leute School, explores how product discovery changes when teams can build faster than they can decide what is worth building.

Many companies are building software factories: agents that turn specifications into tested code. Fed with weak assumptions, the factory produces the build trap at machine speed. Product managers alone cannot feed it fast enough, so discovery becomes a team game.

The question shifts to what is really worth building next, and whether a team is building to learn or building to earn. Access to users, the quality of evidence and the team’s ability to decide become the real constraints.

The session follows one discovery cycle on a demo product: from strategy and a KPI tree to an opportunity, a test card, a test with a real person and a decision about the next bet. It shows where agents and skills help (organizing signals, synthesizing interviews, exploring alternatives, analyzing results) and where people stay responsible: the conversations with users and the calls the evidence cannot make on its own.

You leave with a starting point for your own agentic discovery pipeline: what to delegate, what to keep human, and how to turn cheap building into evidence before you build to earn.