DORA: Team Profiles and Capability Model for AI-Assisted Software Development

At QCon London, in a talk titled 'Scale High Performance, Not Dysfunction,' DORA research lead Nathen Harvey discussed how AI acts as an amplifier: its benefits are maximized only when organizational systems are well-established. DORA combines survey data with qualitative research, and in 2025 published the 'State of AI-Assisted Software Development' report. The research identified seven team profiles, including 'process-constrained' teams, which show high burnout and low throughput despite stable delivery. DORA also summarized seven key capabilities that, when combined with AI, lead to better outcomes, such as clear AI usage policies, healthy data ecosystems, and small-batch development. Harvey noted that AI significantly enhances teams' ability to experiment, allowing them to build multiple prototypes simultaneously. DORA's AI Capabilities Model, released in December, provides a framework for teams to assess their current state and prioritize improvements. In an interview with InfoQ, Harvey addressed findings that while AI improves perceived productivity and code quality, delivery stability declines, possibly due to misaligned incentives and lack of visibility into production impacts. He also highlighted the role of internal platforms in standardizing AI capabilities and abstracting complexity, allowing teams to focus on business needs.