Skip to content
Founderz: Responsible Use of AI home

in collaboration with Microsoft

All case studies

Case study · Government evaluations

UK government: evaluating an AI assistant across departments

A large trial found time saved and people who wanted to keep going. A smaller evaluation found that results varied by task.

Principles involved

When and where

2025

United Kingdom

What happened

From October to December 2024, 20,000 UK civil servants tried Microsoft 365 Copilot in their everyday work. In the Government Digital Service’s findings, published in June 2025, participants reported saving 26 minutes a day on average, and 82% said they wouldn’t want to go back. The report also notes that the tool was weaker on complex or nuanced work, and that human oversight was needed at all times.

A separate evaluation by the Department for Business and Trade, published in August 2025, found 72% of its users satisfied. Observed tasks told a mixed story: people summarized reports faster and better with the assistant, while spreadsheet analysis was slower and less accurate. The department found no evidence that the time saved led to higher productivity.

What the sources establish

The cross-government time savings are self-reported, from 7,115 survey responses, with no control group. The department’s evaluation had a 32% response rate, small samples for observed tasks and no before-and-after baseline. Both describe one product at one moment.

Implications

AI assistants can give people time back, and people notice. The benefit isn’t spread evenly: some tasks improve, some get worse, and time saved doesn’t automatically become better work. So measure task by task, keep people checking and share what you find.

Recommended practice

  1. 01Try AI on your real tasks, and keep a simple note of which ones it improves.
  2. 02Keep checking where the evaluations found weaknesses: complex analysis and nuanced work.
  3. 03Decide what you’ll do with the time you save.

Sources

What each source establishes, and its limits. The practices and recommendations on this page are ours, and the facts come from the sources. See every source we use.

  1. Microsoft 365 Copilot Experiment: Cross-Government Findings Report UK Government Digital Service · June 2, 2025 · Government evaluation 20,000 civil servants used the assistant from October to December 2024 and reported an average saving of 26 minutes a day; 82% wouldn’t want to go back. The tool was weaker on complex or nuanced work, and human oversight was needed at all times. Limits: Time savings are self-reported, from 7,115 survey responses, with no control group.
  2. Microsoft 365 Copilot pilot: DBT evaluation report UK Department for Business and Trade · August 28, 2025 · Government evaluation 72% of users were satisfied. In observed tasks, report summaries were faster and better, while spreadsheet analysis was slower and less accurate. It found no evidence that time saved raised productivity. Limits: A 32% response rate, small observed samples and no before-and-after baseline.

Cite this page

Founderz (2026). UK government: evaluating an AI assistant across departments. The RUAI Standard, 2026 edition. Developed by Founderz in collaboration with Microsoft. https://responsibleai.founderz.com/toolkit/case-studies/uk-government-ai-assistant-evaluations

Licensed under CC BY 4.0: share and adapt with attribution.