United KingdomPublic sector2024–2025UK Department for Business and Trade
Microsoft 365 Copilot trial with 1,000 staff
UK trade department's Copilot trial saved some staff time but found no evidence of better productivity
Problem
The department wanted evidence on whether a general-purpose AI assistant would help civil servants with everyday writing, summarising and meeting work before buying it widely.
Approach
The department gave 1,000 staff Microsoft 365 Copilot licences from October to December 2024. About 70% were volunteers and 30% were picked at random. The evaluation used usage data, diary studies, surveys and interviews, with a control group for comparison.
Results
Takeaway
Most users reported saving some time, mainly on writing, but the department could not show that this raised output. Productivity had not been a main aim of the trial, so little data was collected on it.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Internal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Company's own claim
Sources
Source summaries
Civil Service World reported the department's evaluation of 1,000 Copilot licences used over three months. Most users saved some time, mostly on written tasks, while scheduling and image generation took longer. The evaluation found no evidence that time savings improved productivity, and users reported hallucinations.
The Register reported 72% satisfaction and an average of 1.14 Copilot actions a day in the same trial. Email summaries were faster, but Excel analysis was slower and less accurate. The assessment found no clear evidence of productivity gains, and productivity had not been a key aim.
Full details are in the original sources.