Agentic Architecture Now GA

Copilot code review has transitioned from public preview to general availability with a new agentic tool-calling architecture. This means the code review agent can now dynamically gather broader repository context—including relevant code files, directory structure, and architectural references—to provide feedback that reflects how changes fit into the larger system.

What Developers Will Experience

The agentic approach delivers three key improvements:

• Higher-quality findings that prioritize correctness and architectural integrity
• Lower noise by surfacing meaningful signal rather than volume of comments
• More actionable guidance that helps developers resolve issues quickly and confidently

This results in more targeted review comments that help teams merge with confidence and ship cleaner code faster.

Availability and Setup

The feature is generally available for all users with Copilot Pro, Pro+, Business, and Enterprise plans. Agentic code review runs on GitHub Actions, so:

• Most users: No setup required; reviews run on GitHub-hosted runners by default
• Self-hosted runner users: Complete a one-time setup to receive agentic reviews (see self-hosted runners documentation)
• Alternative option: Use GitHub-hosted larger runners if preferred

Get started with the Copilot code review documentation or read more in the technical blog post.