Nov 18, 2026 — 12:00 – 1:00 PM PT
How the Claude Code team does code review
When agents write the code, pull requests pile up faster than anyone can read them.
In this session, Morgan Westlee Lunt, PM on Claude Code, walks through how Anthropic cut its revert rate by more than 40% even as merged pull requests more than doubled, with most of the reading moved to machines and human attention focused where the risk is.
We’ll cover each layer of the review process, where humans stay in the loop, and how to adopt these practices on your team.
Featured speakers
Morgan Lunt
Claude Code @ Anthropic
What you’ll learn
How to verify work before a pull request exists, with checks Claude can run and a fresh-context review of the diff
How to tune AI code review with CLAUDE.md and REVIEW.md so findings stay accurate and review rounds converge
How to scale human approval by risk, with tiers set by the change rather than the author
How to turn past incidents and reverts into an eval that improves your reviewer over time
How to measure human review time, with revert and incident rates as the guardrail