AI-assisted engineering: Q4 impact report

DX works with hundreds of engineering organizations to measure and improve developer productivity. Through this work, we have a unique vantage point into how teams are adopting AI and what results they’re actually seeing.

This new quarterly series on AI-assisted engineering is designed to share that perspective and provide a data-driven, moment-in-time pulse on how AI is reshaping engineering performance across the industry.

This inaugural report draws on data from over 135,000 developers across 425 organizations, combining system-level telemetry with self-reported insights. It highlights where adoption stands, what measurable impact AI is having, and what distinguishes the companies realizing the most meaningful gains.

A preview of the key findings from Q4 2025:

• AI adoption exceeds 90%. Nearly all developers now use AI tools monthly or more—proof that AI has moved from pilot to practice.
• Developers save an average 3.6 hours per week using AI coding assistants, though those gains have plateaued as usage outpaces integration.
• 22% of merged code is AI-authored, far below vendor marketing claims and underscoring the need for rigorous measurement.
• Daily AI users ship 60% more PRs than non-users, revealing a clear correlation between usage frequency and throughput—but mixed results on quality.
• Enterprises lead in impact but trail in rollout. Large organizations report the greatest time savings per developer, even as adoption lags smaller, faster-moving teams.
• Structured enablement drives measurable ROI. Teams that pair rollout with training, governance, and measurement outperform those treating AI as a plug-and-play tool.

AI’s impact, however, is not uniform. While industry benchmarks help contextualize performance, they can obscure the realities within each organization. AI adoption is producing asymmetric results—some teams are accelerating throughput while seeing declines in quality; others are realizing strong gains in areas like migrations but struggling to extend those improvements to testing or release processes.

In other words, there is no “average” experience with AI. Each organization’s trajectory depends on its systems, culture, and approach to measurement. The most effective leaders use benchmarks to inform their strategy, building on external insights while grounding decisions in data from their own organization.