For four years, we’ve asked engineering leaders the same question: how do you actually know your team is performing? This year’s answers are messier than any dashboard suggests.

AI has changed what engineers do each day faster than anyone has worked out how to measure it. Metrics built for manual pull requests and predictable cycle times now have to explain a workflow that barely resembles the one they were designed for.

This report, based on 384 responses, shows what leaders are measuring, what they’ve stopped trusting, and what they can’t see yet.

Inside the report:

• Why reporting feels like a scorecard. 74% of managers and executives have to report on team performance. Fewer than half of everyone else does.
• Which metrics are earning trust, and which aren’t. 68% avoid lines of code. Incidents, lead time, and cycle time are rising. Only 44% say their metrics reflect real value.
• How leaders and engineers see L&D differently. 81% of CTOs and managers of managers say retention justifies the spend. Only 40% of engineers agree.
• Which skills AI is upgrading, and which it’s wearing down. Understanding unfamiliar codebases improved for 74%. Debugging without AI declined for 33%.
• Where the delivery bottleneck moved. 52% say it’s now code review and validation. Only 7% say writing code. Only 31% measure the impact of AI tools.

Preregister for The LeadDev Engineering Team Performance Report 2026!