By Alex Smith |
The Department for Work and Pensions has published the findings of its trial of Microsoft 365 Copilot in a report, titled An Evaluation of DWP’s Microsoft 365 Copilot Trial , revealing that the generative AI tool delivered measurable benefits in task efficiency, job satisfaction and perceived work quality for central office staff.
The trial ran from October 2024 to March 2025 and involved 3,549 staff members. The evaluation used a mixed‑method framework combining large‑scale surveys of Copilot users and non‑users, econometric analysis, and in‑depth interviews. Two surveys collected responses from 1,716 Copilot users and 2,535 non‑users, while regression analysis estimated Copilot’s impact on efficiency and work quality, controlling for demographic, occupational and AI‑attitude factors. Nineteen interviews explored user experiences, training and task usage patterns.
Copilot was used regularly, primarily for summarising documents and meetings, drafting emails, and searching for internal information. According to the report, users generally found Copilot intuitive and well-integrated with Microsoft applications, though output quality depended on human oversight.
“Ninety per cent of users indicated that Copilot helped them save time,” the report says, estimating an average saving of 19 minutes per day across eight routine tasks, with the largest reductions in searching for information and email drafting. Time saved was often reinvested in higher‑value activities such as project delivery and strategic planning.
The evaluation also found that 65 per cent of users reported increased job satisfaction. “Many users reported that Copilot reduced cognitive load and helped them focus on more meaningful aspects of their roles,” the report notes. Medium increases in job satisfaction (0.56 points on a seven‑point scale) and perceived quality of work output (0.49 points) were estimated compared with non‑users.
Quality improvements were linked to clearer, more structured written outputs, though users stressed the need for editing and human judgement. The report emphasises that Copilot is best seen as a “collaborative assistant” rather than a replacement for professional expertise.
The evaluation acknowledged limitations, including non‑random licence allocation, reliance on self-reported measures, and absence of pre‑trial baseline data. Despite these constraints, the report concludes that Copilot was widely adopted and provided consistent evidence of positive effects on efficiency, job satisfaction, and work quality.
Read the full report: https://www.gov.uk/government/publications/an-evaluation-of-dwps-microsoft-copilot-365-trial/an-evaluation-of-dwps-microsoft-365-copilot-trial