This is the story of how our organization reimagined Threat Models as living, machine‑readable code embedded in our Secure SDLC. By co‑locating YAML‑based threat models in code repos, automating adversarial test cases for every threat identified, and streamlining repo onboarding via self-serve tooling, we empowered 750+ Security Champions to drive developer‑led security at scale. Finally, we’ll show how AI agents accelerate our process while preserving human expertise.