DORA's 2025 research reveals that traditional delivery metrics cannot capture AI's true impact on engineering systems. AI acts as an organizational amplifier, accelerating healthy practices while intensifying existing bottlenecks. The research introduces five software delivery performance metrics and the DORA AI Capabilities Model, emphasizing that successful AI adoption depends on team capabilities rather than tool selection. Key findings show AI shifts bottlenecks downstream to code review, junior engineers adopt AI more heavily than seniors, and measuring ROI requires qualitative insights beyond dashboards. Leaders need clear AI policies, accessible documentation, and direct conversations with engineers to understand where AI creates friction or value.
Sort: