Government coders save 28 days a year with AI tools in landmark trial

Government software engineers saved the equivalent of almost six working weeks a year thanks to AI coding assistants, according to the results of the UK’s largest public sector trial of the technology.

More than 1,000 developers across 50 departments took part in the pilot, run by the Department for Science, Innovation and Technology (DSIT) between November 2024 and February 2025. The results suggest that AI assistants such as GitHub Copilot and Google’s Gemini Code Assist could play a major role in boosting digital productivity across Whitehall.

On average, participants reported saving nearly an hour per day – equivalent to 28 working days a year – by using AI to draft and review code. The time released allowed teams to focus on higher-value tasks such as developing new services, including the Humphrey AI assistant for civil servants and digital tools in healthcare.

Technology Minister Kanishka Narayan said the findings showed how government can catch up with the private sector’s use of digital tools: “For too long, essential public services have been slow to use new technology – we have a lot of catching up to do. These results show that our engineers are hungry to use AI to get that work done more quickly, and know how to use it safely. With a £45 billion jackpot at stake, it’s not an opportunity we can pass up.”

The £45 billion figure refers to the government’s wider ambition to find efficiencies through automation and AI, as part of its “Plan for Change” strategy. Ministers argue that smarter use of digital tools can cut costs, reduce backlogs and improve service standards for citizens.

However, despite the strong efficiency gains, the trial also revealed that developers remain cautious about AI outputs. Only 15% of AI-generated code was accepted without edits, with most engineers using the tools to create first drafts or to check existing work.

Guidance issued by DSIT emphasised that coders remain responsible for all outputs, whether or not they were originally generated by an assistant. Training and support were provided throughout the pilot to ensure staff could use the tools responsibly.

Reception among DDaT teams

Survey data suggests that government developers were broadly enthusiastic about the technology. Nearly three-quarters (72%) felt AI assistants delivered good value for their organisation, 65% said tasks were completed faster, and 56% reported they could solve problems more effectively. More than half (58%) said they would prefer not to return to coding without an AI tool.

Industry partners also welcomed the results. Tara Brady, President of Google Cloud EMEA, described the trial as a “landmark” moment: “This underscores the transformative potential of AI in enhancing productivity and problem-solving for coding professionals, and highlights the successful collaboration stemming from Google Cloud’s Strategic Partnership Agreement with the UK government.”

Findings from the trial will help shape decisions on whether to expand AI coding assistants across government. Ministers see the technology as a foundation for wider adoption of AI in the public sector, positioning Britain as a leader in digital innovation while delivering tangible service improvements for citizens.