Assumptions as Code: A pattern for platform prioritization

Platform teams struggle to make confident decisions when stakeholder priorities conflict and data is scattered. Without shared ground truth, the loudest voice wins. In this talk, Mina Tawadrous shares how his team is experimenting with a human-in-the-loop approach: AI queries version control for PR metrics, CI/CD for deployments, and incident tracking for patterns, while humans validate findings and add context from interviews and observations. These insights feed a central assumptions engine that all teams use and iterate on. This is helping reduce roadmap conflicts by backing decisions with both hard numbers and developer pain points, not gut feelings or HiPPO (highest paid person's opinion). Attendees will see real examples and practical techniques to try on their teams.