Date context: publication
The UK trial provided 20,000 civil servants with Copilot licences for three months. Its report uses 7,115 survey responses and anonymised usage data for 14,500 people. The estimate of 26 minutes saved daily uses midpoints of participants' time bands. HR was among the professions. Logs and participant experience offer different views of adoption.
UK Government Digital Service · 2025-06-02
Full HTML methods, time calculation and results checked; raw data not independently analysed.
What the source shows
The UK trial provided 20,000 civil servants with Copilot licences for three months. Its report uses 7,115 survey responses and anonymised usage data for 14,500 people. The estimate of 26 minutes saved daily uses midpoints of participants' time bands. HR was among the professions. Logs and participant experience offer different views of adoption. Source: UK Government Digital Service, 2025-06-02
What the material does not establish
Time savings are self-reported, not a secure causal estimate for each person. The report could not establish how saved time was used. Non-response, coarse time bands and available logs limit the conclusions. Methods and results in the HTML report were read. Source: UK Government Digital Service, 2025-06-02
For HR, AI + JJ
AI + JJ, application: Track use, time, quality and experience separately in a small HR experiment. Does a concrete result improve, or does the task mainly feel easier? Source: UK Government Digital Service, 2025-06-02