AI-assisted PRs counts pull requests in which at least one commit was authored on a day the developer recorded AI tool activity (Copilot, Claude Code, etc.). It's the simplest signal of how much AI-touched work flows through the team.
What it measures
The number of PRs merged in the period that have at least one commit authored on an "AI day" for the commit author.
How Leanmote calculates it
ai_assisted_prs = count(prs where merged_at in period and exists commit c in pr such that ai_active(c.author, c.committed_at))
ai_active(user, date)means the user had at least one AI-tool event on that calendar date.AI events come from connected AI tool integrations — Copilot usage telemetry, Claude Code OTLP sessions.
The classification is binary at the PR level: a PR is AI-assisted or it isn't. There's no partial credit for "small AI use."
How to interpret it
Rising count — adoption is growing. Compare to the team size to see whether it's broadening or just deepening among existing users.
High AI-assisted share with AI Lead Time advantage — your team is getting measurable value from AI.
High AI-assisted share without measurable Lead Time advantage — adoption without impact. Investigate whether AI tools are being used effectively.
What to do about it
Pair this metric with AI Lead Time Comparison and AI Intensity for a complete adoption-and-impact picture.
Don't optimize for higher AI-assisted PR counts in isolation. The goal is delivery improvement, not AI usage.
Related metrics
AI-assisted Commits
AI Intensity
AI Lead Time Comparison
AI Governance metrics overview