As AI products evolve from simple models to complex, agentic systems capable of autonomous decision-making, the need for robust evaluation frameworks has never been greater. This session will demystify the role of evals in the AI product lifecycle, equipping product managers with practical strategies to define, measure, and iterate on quality for agentic AI applications. Attendees will learn how to collaborate with engineering, select the right metrics, and use real-world evaluation data to ship reliable, high-impact AI products that delight users and drive business value