JPMorgan Is Grading 65,000 Engineers on Whether They Use AI

JPMorgan Chase has built a real-time dashboard that tracks GitHub Copilot usage across its 65,000-strong engineering team — and as of early 2026, that data feeds directly into performance reviews. Every engineer gets classified as a heavy user, light user, or non-user, making AI adoption a formal career metric for the first time at a major regulated financial institution. The move reflects a fundamental shift in how performance is being defined. JPMorgan now formally evaluates not just what engineers deliver, but how they deliver it — and whether they lean on AI tools as part of their daily workflow. With documented productivity gains of 10 to 20 percent already on the books, leadership had the data to back the shift from encouragement to expectation. But the policy raises real questions for HR leaders everywhere. Raw adoption metrics without context can penalize strong performers who work in codebases where AI assistance is less applicable. Senior engineers debugging legacy systems may clock fewer Copilot interactions than junior engineers building greenfield apps — but that doesn't reflect their productivity. Building fair, defensible adoption metrics means accounting for role, project type, and codebase complexity. In this episode, we break down what JPMorgan built, why it matters, and what HR leaders should do now — before AI adoption creeps into your performance frameworks informally.