2025 Dora State of AI-Assisted Software Development (cloud.google.com)

Google’s 2025 DORA Report: State of AI-Assisted Software Development (based on ~5,000 survey responses and 100+ hours of qualitative research) finds that AI is now near‑ubiquitous in engineering—90% of respondents use it and over 80% say it boosts productivity—but it’s an amplifier, not a fixer. Adoption correlates with higher throughput and better product performance, yet increased change velocity often reduces delivery stability unless teams have strong control systems (automated testing, mature version control, fast feedback loops). Teams working in loosely coupled architectures with rapid feedback see clear gains; tightly coupled systems and slow processes see little benefit. Trust in AI-generated code remains a concern for ~30% of practitioners. The report introduces seven team archetypes to help diagnose where AI will help or harm, and the DORA AI Capabilities Model identifying seven organizational capabilities that magnify AI’s positive impact. Technical implications are concrete: invest in platform engineering (90% of orgs have at least one platform), prioritize user‑centric problem definition, harden safety nets, and treat AI adoption as organizational transformation. Recommended starter steps include clarifying AI policies, connecting AI to internal context, prioritizing foundational practices, fortifying safety nets, investing in internal platforms, and focusing on end users—so AI amplifies strengths rather than magnifying existing weaknesses.