Google: AI Widens Gap Between Strong & Weak Dev Teams
▼ Summary
– Nearly all developers (90-95%) now use AI tools for software development, representing a 14% increase from the previous year.
– AI acts as an amplifier, magnifying the strengths of high-performing organizations while also exacerbating the dysfunctions of struggling teams.
– The study identified seven team archetypes, with the top performers proving that high speed and high-quality code delivery are not mutually exclusive.
– Successful AI adoption depends on seven key organizational practices, including AI policy, data ecosystems, and quality platforms, rather than just the tools themselves.
– Two key factors driving AI success are the widespread adoption of platform engineering and the use of value stream management (VSM) to optimize workflows.
The widespread adoption of artificial intelligence in software development is fundamentally reshaping the industry, acting as a powerful amplifier that highlights the existing strengths and weaknesses within development teams. A new comprehensive study from Google’s DevOps Research and Assessment (DORA) program reveals that while AI tools are now used by the vast majority of developers, their impact is not uniformly positive. The 2025 report, based on a survey of 5,000 professionals and extensive interviews, suggests that AI magnifies the capabilities of high-performing organizations while exacerbating the dysfunctions of struggling ones.
Virtually all developers are now integrating AI into their workflows, with estimates ranging from 90% to 95% adoption. On average, developers spend about two hours per day interacting with these tools. While a significant 80% of programmers report an increase in overall productivity, a smaller majority, 59%, feel that the quality of their code has actually improved. This points to a common experience: AI can accelerate the volume of work produced, but it demands rigorous review and validation. The code generated is rarely perfect on the first attempt, necessitating a careful QA process to catch errors, much like managing work delegated to a human team member.
The concept of AI as an amplifier is central to the findings. The technology does not create competence from scratch; instead, it enhances what is already there. Well-organized teams with strong engineering practices find that AI boosts their efficiency and output. Conversely, teams plagued by disorganization or poor processes discover that AI can accelerate their path to chaos, for instance, by quickly propagating errors from a single malformed prompt. The difference often lies in the foundational systems a team has in place, such as robust version control, which can mitigate potential disasters.
The research identified seven distinct team archetypes based on eight performance factors, including team collaboration, software delivery speed, and burnout levels. These range from teams in “survival mode” to “harmonious high-achievers” who deliver high-quality work quickly. The report challenges the long-held belief that speed and stability are a trade-off, showing that the top-performing teams consistently achieve both. AI acts as a mirror, making the characteristics of each archetype more pronounced.
Success with AI is not merely a matter of choosing the right tool. The study emphasizes that it is a systems problem, pinpointing seven key practices that drive impact. These include having a clear AI policy, maintaining high-quality data ecosystems, utilizing version control, working in small batches, and, crucially, relying on quality internal platforms. Teams that invest in these foundational areas are far more likely to reap the benefits of AI adoption.
Two factors were identified as critical differentiators from the previous year’s study, which had shown AI reducing productivity. First, organizations have grown more familiar and effective with AI tools. Second, and more importantly, the widespread adoption of platform engineering, building strong internal development platforms, has created an environment where AI can thrive. When developers are not fighting against a cumbersome system, AI amplifies their value creation. Additionally, Value Stream Management (VSM), which maps the flow of work from idea to delivery, dramatically amplifies the positive impact of AI by ensuring it is applied to the right problems.
The overarching message is clear: AI has moved from an experimental technology to a mainstream component of software development. The ultimate advantage, however, does not come from the AI itself. Real, measurable gains are achieved by organizations that have built solid, disciplined systems and platforms. AI simply reflects the underlying health of an organization, turning potential into performance for those who are prepared and revealing instability for those who are not. The gap between strong and weak teams is not narrowing with AI; it is widening.
(Source: ZDNET)
