Overview
- Google Cloud’s DORA study surveyed nearly 5,000 technology professionals and conducted 100+ hours of interviews, finding 90% now use AI at work with a median of two hours daily.
- More than 80% report productivity gains and 59% see better code quality, with this year’s data linking AI use to higher software delivery throughput after last year’s slowdown.
- Concerns over reliability persist as a trust paradox emerges, with 24% expressing high trust in outputs versus 30% reporting little or no trust.
- DORA introduces a seven-part AI Capabilities Model and identifies seven team archetypes, concluding that outcomes hinge on clear policies, strong data practices, version control, small batches, user focus, and robust internal platforms.
- Usage patterns skew toward writing new code (71%) via chatbots and IDEs, autonomous agent modes remain uncommon, and Google reports roughly 30% of new code is AI‑generated with an estimated ~10% engineering velocity gain under human review.