Developer productivity is the most contested metric in engineering. DORA metrics (deployment frequency, lead time for changes, MTTR, change failure rate) are the gold standard — but they require instrumentation that most teams haven't built. In the absence of good data, teams fall back on activity proxies: commits per day, PRs opened, lines of code written. These are not productivity metrics. They're work theater metrics.
The AI coding tool era has made this worse. METR's 2026 randomized controlled trial found experienced developers were 19% slower with AI tools while believing they were 20% faster — a 39-point confidence gap. If your productivity dashboard is showing "commits per day up 40%" because AI tools are generating more code, you're measuring the wrong thing and drawing the wrong conclusions.
GitIntel's productivity dashboard separates signal from noise. It measures PR cycle time (time from open to merge), review bottleneck points (where PRs stall), rework rate (commits that revert or substantially modify previous commits), and AI tool contribution vs verified human productivity. It's not about blaming AI — it's about understanding where AI is genuinely helping vs where it's adding volume that creates review debt.
Deploy the dashboard as a static page in minutes: `gitintel dashboard --deploy`.