Understanding AI's Impact on Developer Workflows

A JetBrains HAX team mixed-methods study analyzed two years of anonymized IDE telemetry from 800 developers (400 AI users, 400 non-users) alongside surveys and interviews to understand how AI coding assistants reshape developer workflows. Key findings: AI users typed significantly more code (~600 extra characters/month vs. ~75 for non-users), performed far more deletions/edits (~100 more per month), and showed increased context switching — contradicting the common belief that in-IDE AI reduces interruptions. Notably, developers largely failed to perceive these behavioral shifts themselves, especially the surge in code editing. Code quality perception improved slightly but debugging behavior showed no significant change. The study highlights a consistent gap between what developers perceive and what telemetry reveals, underscoring the value of combining behavioral data with self-reported measures.