Modern Product Ops is not just a toolkit of apps, it is an operating system for how decisions get made. In this talk I will introduce process as code, using natural-language “How We Work” documents as the source of truth that defines workflows, triggers automations (including AI agents), and improves over time based on real signals from meetings, messaging, and delivery. Chris will share early examples from teams at GitHub, including how they turn working agreements into automation-ready policies, how we capture “what actually happened” through transcripts and artifacts, and how the system can spot repeated manual steps and suggest specific automations, agent workflows, or doc updates. The outcome is a modern Product Ops stack that helps teams move faster, stay aligned, and spin up new teams from a proven template that evolves as they learn.