Key Findings
1. Code review remains the most acute problem
of engineering leaders identify review volume, senior-reviewer burden, or both as a top-two problem. As agents produce more code, teams face growing demands on human review capacity.
2. The performance gap is widening
say the gap between their strongest and weakest engineers has widened since adopting AI. Access to the same tools isn’t translating into the same results.
3. Effective practices rarely become team standards
say effective AI practices get captured and rolled out as a team or org standard. Helping those practices spread is an opportunity to make the whole team more effective.
4. Teams are rethinking human review
already let some agent PRs merge with sampled or no human review. Mandatory review of every change is beginning to give way to new approaches to verification.
Meet with a former CTO on Span’s team to see how your organization compares. We'll help you identify opportunities to improve your software development process.
