As AI systems evolve from features to agents that act and decide, the PM role fundamentally changes. It’s no longer just about shipping capabilities—it’s about defining intent, guardrails, and measurable outcomes. But before architecting agents, teams must establish AI–Value Fit: clear customer pain, decision leverage, and organizational readiness. Without it, even the most sophisticated agent becomes an impressive demo that fails to scale. In this session, I’ll share practical frameworks for identifying AI–Value Fit and translating it into agent design—defining decision boundaries, human-in-the-loop systems, and success metrics that drive durable business impact.