In 2027, someone from your own company will have to decide whether your AI system can continue to operate. They won’t ask for policies or an inventory—they’ll ask for evidence. And evidence for a model isn’t written; it’s measured—and measuring takes time. In this session, using simulated cases of systems that make decisions about people, you’ll learn what you need to start measuring now so you can demonstrate it then—and why that expense belongs in next year’s engineering and data budget, not in the legal counsel budget.