DORA 2025: AI Amplifies Your Strengths (and Your Weaknesses)

The DORA 2025 “State of AI-assisted Software Development” report draws on over 100 hours of qualitative data and survey responses from nearly 5,000 technology professionals worldwide. Its central finding fits in one sentence: AI is an amplifier. It magnifies the strengths of high-performing organizations and the dysfunctions of struggling ones.

This finding fundamentally changes the question leaders should ask. It is no longer about whether to adopt AI, but how to prepare the organization to extract value from it.

Adoption is near-universal, trust remains measured

90% of respondents use AI in their daily work, a 14% increase from 2024. The median time spent with AI is two hours per day, roughly a quarter of the workday. Writing new code remains the top use case (71% of developers), followed by technical literature reviews (68%), modifying existing code (66%), and proofreading (66%).

Over 80% of respondents believe AI has increased their productivity. Yet 30% report little to no trust in AI-generated code. This coexistence of massive adoption and measured skepticism is a sign of maturity. Developers compare this “trust but verify” approach to the healthy skepticism they apply to solutions found on Stack Overflow.

Seven team profiles reveal the complexity of performance

The report identifies seven distinct team profiles through cluster analysis, ranging from “harmonious high-achievers” to teams caught in a “legacy bottleneck.” This classification goes beyond software delivery metrics to incorporate team well-being, friction, and burnout.

Clusters 6 (pragmatic performers) and 7 (harmonious high-achievers) represent nearly 40% of the total sample. Their existence proves that high velocity and high quality are not a theoretical ideal but an observable reality. These teams excel simultaneously in throughput and stability, confirming that the “speed vs. stability” trade-off is a myth.

At the other end, cluster 1 teams (foundational challenges) are stuck in survival mode with fundamental gaps in their processes. Cluster 3 teams (constrained by process) run on a treadmill: despite stable systems, their effort is consumed by inefficient processes, leading to high burnout.

AI improves throughput but increases instability

The most notable shift from 2024 concerns software delivery throughput. Last year, AI was associated with reduced throughput. In 2025, the relationship has reversed: AI is now associated with improved throughput, product performance, and time spent on valuable work.

However, AI continues to increase delivery instability. The report suggests that teams have adapted for speed, but their underlying systems have not yet evolved to safely manage AI-accelerated development. AI also has no measurable effect on friction or burnout, which remain properties of the organizational system rather than the individual workstation.

The DORA AI Capabilities Model: seven foundational capabilities

The core of the report is the new DORA AI Capabilities Model. It identifies seven capabilities that, when combined with AI adoption, amplify its benefits on organizational outcomes.

The first capability is a clear and communicated AI policy. Organizations that explicitly define permitted tools, encourage experimentation, and communicate expectations see AI amplify individual effectiveness, organizational performance, and reduce friction.

The second is a healthy data ecosystem. When internal data is high-quality, easily accessible, and unified, AI benefits on organizational performance are amplified. The third, AI-accessible internal data, goes further: connecting AI tools to internal systems amplifies their impact on individual effectiveness and code quality.

Strong version control practices form the fourth capability. Commit frequency amplifies AI’s effect on individual effectiveness, and rollback capability amplifies its effect on team performance. Working in small batches, the fifth capability, amplifies AI’s impact on product performance and reduces friction.

User-centric focus is the sixth capability, and its results are particularly striking. In the absence of user-centric focus, AI adoption has a negative impact on team performance. With user-centric focus, the impact becomes positive. The seventh capability, a quality internal platform, amplifies AI’s effect on organizational performance.

Internal platforms: the strategic lever for AI

90% of organizations have adopted platform engineering. The report shows that users perceive their platform as a single entity: its overall effectiveness matters more than the quality of any individual feature.

The most important finding for leaders is that AI’s positive effect on organizational performance depends directly on internal platform quality. When platform quality is low, AI has a negligible effect. When it is high, the effect is strong and positive. Investing in AI without investing in the platform amounts to local optimization without systemic impact.

Value Stream Management as a force multiplier

Value Stream Management (VSM), the practice of visualizing and improving the flow of work from idea to customer, acts as a force multiplier for AI. Teams that understand their value stream spend significantly more time on work they consider valuable.

AI’s effect on organizational performance is modest when isolated. It is dramatically amplified in organizations with strong VSM practices. VSM ensures that local productivity gains translate into measurable improvements at the team and product level.

What this means for your organization

The DORA 2025 report delivers a clear message: the greatest returns on AI investment come not from the tools themselves, but from a strategic focus on the underlying organizational system. The quality of the internal platform, the clarity of workflows, and the alignment of teams determine whether AI produces positive outcomes or amplifies existing chaos.

For organizations considering or accelerating AI adoption, the report recommends treating it as an organizational transformation. Clarifying AI policy, investing in the data ecosystem, strengthening foundational engineering practices, and building a quality internal platform are the prerequisites for turning AI potential into measurable performance.

📄 Download the DORA 2025 Report (PDF)