What product leaders must rethink in the AI era

For decades, the product management playbook was built on a simple premise: identify a problem, ship a deterministic solution, measure the output, and iterate. But in the era of AI agents, that playbook is being rewritten in real time.

At Sierra, building AI agents for some of the world’s most iconic brands has required a fundamental rethink of product management. Roadmaps no longer stretch cleanly across quarters — they operate as continuous “Vision Loops,” where teams must re-earn product-market fit continuously as models evolve and customer expectations shift. Trust in AI systems isn’t about being perfectly right — it’s about being predictably aligned with user expectations. And succeeding in this environment requires developing “Model Sense”: the judgment, taste, and informed “vibe checks” needed to build and evaluate systems that are probabilistic by design.

In this talk, Sachi Shah shares the core mindset shifts required to lead in a world where products learn, adapt, and sometimes behave unpredictably. You’ll leave with a clearer framework for ownership, trust, and what great product management looks like in the AI era.