PilotCognitive automation2024-12-31
A UK cross-government trial gave 20,000 civil servants in 12 organisations Microsoft 365 Copilot for three months; users self-reported saving 26 minutes a day on average, mostly on drafting
Administrative assistantoccupation page →
Event date / reported
2024-12-31 · reported 2025-06-02
Evidence stage
PilotSmall-scale trial in a real setting. Tells us the deployment conditions are being tested, not that they hold.
Tasks this bears on
Scheduling and correspondence
Finding the slot six people can make, drafting the reply, chasing the confirmation, rescheduling when it all moves.
Automating✓ Evidence-backed
Where this applies
UK central government, 30 Sep–31 Dec 2024, self-reported time savings with no control group; scheduling meetings saved about 9 minutes, drafting documents 24. The report notes the tool struggled with complex data and that human oversight was needed throughout.
What this means
A pilot record on the scheduling-and-correspondence task at unusual scale for a government: 20,000 civil servants in 12 organisations used Microsoft 365 Copilot for three months and reported saving about 26 minutes a day — drafting documents about 24 minutes, scheduling meetings about 9. It shows deployment conditions being tested in real offices, and which half of the task the tool helped with: the drafting, far more than the scheduling.
What it does not yet show
The savings are self-reported with no control group, so they measure what users felt, not what happened, and the report itself notes the tool struggled with complex data and needed human oversight throughout. A trial is not a rollout — the government still had to decide whether to buy licences — and 26 minutes across 20,000 people says nothing about assistant roles, which the report does not mention.
What you can check
Open the findings report's task table (page 12) and, for one working day, log your own minutes in the same categories — drafting, summarising, scheduling, searching — then compare; if your day is mostly the categories the trial saved little on, the pilot's 26 minutes is not your 26 minutes.
Does it change the assessment?
No. The impact index is never moved by a single event. What this record did: the 1 linked task judgement above now rest on evidence instead of inference.
Source
UK Government Digital Service — M365 Copilot Experiment findings report · verified 2026-09-10 · Claude (VOLO agent) — source text fetched and cross-checked · interpreted 2026-09-10 · Claude (VOLO agent)