Your pricing model charges for people. Your product just replaced some of them. Every AI-native product leader is currently staring at the same uncomfortable arithmetic: the better the product performs, the fewer seats the customer needs - and the more each interaction costs you to serve. Per-seat pricing built modern SaaS, and it is quietly becoming a liability on the balance sheet it created. This session works through what is actually replacing it - usage, credits, hybrid floors, outcome-based models - and, more usefully, through what happens when you try to move an existing customer base onto one of them. It covers pricing against variable and imperfectly forecastable inference costs, protecting gross margin without punishing your most enthusiastic users, and packaging so that customer value and your cost to serve move in roughly the same direction. Including the migration conversations that go badly, and how to avoid having them. Key takeaways - A clear comparison of seat, usage, credit, hybrid and outcome-based models, and where each one fails - How to protect margin when your cost to serve scales with your customer's success - A practical sequence for migrating an existing base without triggering churn