GitHub Copilot Code Review: Repository Insights Track How Feedback Is Received

GitHub CopilotView original changelog

GitHub added a Repository insights page as the first item in each repository's Insights tab. It shows how developers respond to Copilot code review feedback. The page tracks comments addressed, high-severity comments addressed, pull requests with no later human change request, and merged pull requests Copilot did not review. The same metrics are available in the Copilot usage metrics API.

Key Takeaways

  • Repository insights is the first item in each repository's Insights tab, so the data is easy to find.
  • The page tracks four summary signals, from comments addressed to merged pull requests that Copilot never reviewed.
  • The addressed rate is judged by an AI model after merge, and the page shows its trend over time and breakdowns.
  • Comment categories show which kinds of feedback Copilot leaves most often.
  • The same data is exposed through the repos-1-day reports in the Copilot usage metrics API.
  • Teams can use the view to find review coverage gaps and target enablement where Copilot review is missing.

A Scorecard for Copilot Code Review

GitHub added a Repository insights page as the first item in the Insights tab of each repository. The page shows how Copilot code review feedback is received in that repository. It answers questions teams have had about AI review: do developers act on the comments, and do human reviewers still step in afterward?

Summary Signals

The page opens with four headline signals.

  • Comments addressed, the share of Copilot's review comments that developers acted on.
  • High-severity comments addressed, a narrower view of the most important feedback.
  • Pull requests with no later human change request after a Copilot review.
  • Merged pull requests that Copilot did not review, which exposes gaps in coverage.

Deeper Views

Beyond the summary, the page offers a view of the addressed rate, including its trend over time and breakdowns. GitHub's documentation explains that an AI model judges each published Copilot comment after a pull request merges, to decide whether the issue it raised was addressed. A second view shows comment categories, which reveal the types of feedback Copilot leaves most often.

API Access

The same metrics appear in the enterprise and organization repos-1-day reports in the Copilot usage metrics API. Enterprise owners, billing managers, organization owners and roles with the View Copilot Metrics permission can read these reports, as long as the Copilot usage metrics policy is enabled.

Why It Matters

The insights let teams see whether Copilot's feedback leads to changes, whether human reviewers keep requesting changes after Copilot reviews, and where merged work skips Copilot review entirely. Teams can use this to find coverage gaps and target enablement. Viewing a repository's insights needs read access to that repository.