Plenty of engineers were experimenting with AI on their own. Getting the whole organization to ship better software because of it was a different problem: without shared practice, every team relearns the same lessons and adoption stalls at the enthusiasts. A memo wasn't going to close that gap. It needed a person on each team.
What we set up
Three reinforcing structures. The AI Engineering Champions initiative — which I founded — borrows the Security Champions model: every engineering team formally designates a champion who leads practical adoption by influence, not mandate. A seat on the enterprise AI Center of Excellence connects team-level practice to organization-level policy on safe, effective use. And hands-on enablement — workshops, demos, and 1:1 coaching — meets teams at their actual maturity level instead of assuming one.
Decisions
Let teams pick their champion. Mandates get you compliance at best. A champion the team picked themselves gets people actually using it.
Whole-lifecycle integration. We aimed AI at the full development lifecycle: planning, architecture, implementation, testing, and code review, not just code completion. In practice that means Claude Code workflows with each team's conventions encoded as standards and skills the agent loads before touching code — one champion's team turned its Spring 3.5→4.0 upgrade into a reusable agent skill. That's where the gains in velocity and defect detection came from.
Prove it on real partner problems. Working sessions with partner teams prototyped AI into their actual workflows: accessibility remediation, security-vulnerability fixes, automated test generation, documentation. Skeptical teams came around once their own backlogs started shrinking.
Where it landed
AI went from a handful of enthusiasts to an org-wide engineering practice: every engineering team has a named champion, agent-assisted workflows run across planning, testing, and review, and we track the velocity and defect-detection gains with real numbers. It stuck after the novelty wore off because the teams own the practice themselves. The migration story shows what that practice looks like when I apply it myself.
Founder and lead of the champions program, member of the Center of Excellence, and the hands-on coach for much of the enablement work itself.