Date & time
17:00
Register for the panel discussion
Login or join LeadDev.com to view this content
Code review was built as a checkpoint for a small number of changes, from people who could explain their reasoning if asked. It wasn’t built for the sheer volume of edits today, from authors who can’t explain their reasoning at all. AI-co-authored pull requests carry roughly 70% more issues than human-only ones.
It’s easy to add another review at the end of the pipeline but far more effective to catch mistakes when they happen. Handing development over to a non-human collaborator means safeguarding against its weaknesses by adding more context up front, more frequent check-ins, and more accountability at the point code is written.
In this session, we bring together engineering leaders who are rebuilding the software development life cycle (SDLC) to account for the new requirements of verification at machine scale. Participants will learn where to introduce verification throughout the SDLC to manage much higher volumes of change.
Key takeaways:
- How to move verification earlier in the SDLC: from a pre-merge gate to a continuous check at every stage where an agent acts
- How to preserve accountability when a change has multiple authors, human and otherwise
- How to give agents the context to produce reviewable work upfront, rather than relying on review to catch what they got wrong