Letting AI Approve Low-Risk PRs: What Worked in Production

About 30% of code changes got no reviewer feedback. So the team let an LLM auto-approve the low-risk ones, and median time to merge dropped from about 18 to about 12 hours. Sarah Deitke works on the DevEx AI team at Duolingo, where she helped grow AI tool adoption from about 80% to nearly 100% of engineers, then used that groundwork to build an AI risk-assessment bot that auto-approves low-risk pull requests. Recorded at QCon London 2026. QCon conferences are practitioner-driven events where senior engineers share real production experience. 🔗 Transcript & slides available on InfoQ: https://bit.ly/46IRs8G #AICodeReview #CodeReview #DeveloperExperience #AIAdoption #QCon