Trusted code at scale from legacy to AI | BRK287

Eric Tooley and Subhash Arja discuss managing code quality debt as AI increases the volume of code teams need to maintain. The session connects quality findings with developer workflows, engineering standards, and reviewable fixes.

Code quality debt across new and existing code

Every codebase carries quality debt, and AI can add to it faster than teams can address it. This Microsoft Ignite 2026 session focuses on three areas:

GitHub Code Quality identifies these issues in both new and existing code, bringing findings and recommended fixes into the places where developers work. It also provides governance capabilities for engineering leaders to enforce quality standards.

From findings to reviewable pull requests

Agentic autofix addresses accumulated findings at scale by turning remediation backlogs into reviewable pull requests. The workflow pairs automated remediation with a review step, rather than treating generated fixes as changes to accept without inspection.

Session details and resource