What I find most useful is not a leaderboard. It is the feedback loop. Each builder should be able to level-up grounded in their own data, something like:
“You merge 82% of your opened PRs. The top 5% in Engineering average ~90%. Closing that gap adds 3 points.”
That changes the conversation. “Use AI more” is vague and mostly unhelpful. “Ship 4 more PRs this month, give 6 more reviews" is concrete. For me, that is the real shift: moving from measuring AI usage to measuring AI-enabled impact.