Step 12 months into the future - where AI agents create data products, enforce rules, and act in real time. In that world, anyone can request a new data product with a simple form and natural language prompt: “I need to predict customers likely to drop after purchasing” - and get it instantly. Who controls the pace of production so value isn’t buried under maintenance? Are data products even relevant in this new paradigm? How are AI-built assets folded back into a trusted delivery lifecycle? This session reverse-engineers what might soon be reality, exposing the issues these capabilities could bring - and the guardrails needed to keep them in check.