AI: Moving from Small Time Savings to Real Productivity Gains

The Department for Business and Trade has just completed the largest public sector trial to date of Microsoft Copilot. With 1,000 staff involved, the evaluation offers one of the clearest pictures yet of how generative AI is used inside government.

The findings are striking. Copilot helped most with written tasks such as drafting, reviewing and summarising, generating some time savings. It was also where the tool had the most positive impact for neurodiverse colleagues and for those whose first language is not English. That makes sense: AI’s strengths lie in processing and shaping text, and for many people that reduces barriers to getting their work across.

But the trial also showed that saving time didn’t always mean getting more done. In some areas, like generating images or scheduling meetings, the tool slowed people down. The evaluation also highlighted issues of accuracy, environmental impact and a lack of confidence in outputs. The bottom line: it “did not find evidence that time savings have led to improved productivity”.

If we step back, what it really shows (and what the DBT report itself recommends) is that when people are given a tool without structured support or training, they stick to the most obvious uses. It’s the equivalent of giving someone a smartphone and watching them use it only to make calls, as if it were still an old brick-like Nokia. Without guidance on what’s possible, and without understanding how to work around the limitations, the potential remains untapped.

At Teamshaper, we see this every day. That’s why we’ve developed programmes to help public sector teams move beyond “AI for convenience” and start realising productivity gains:
➕Human-Centred Writing in the Age of AI – helps communications teams use AI to support, not replace, their writing. We focus on secure, compliant use while also showing how to avoid the pitfalls of generic or jargon-filled text that undermine trust.
➕Competitive Bid Writing – as AI makes compliance easier to achieve, it has become harder for bids to stand out. Our training shows teams how to use AI for routine drafting and formatting, while focusing energy on the originality, evidence and creativity that actually win bids.
➕AI-Ready Teams – goes beyond writing and into process improvement. With light-touch facilitation, we help teams map workflows, identify where AI can reduce errors and repetitive effort, and design prompt libraries tailored to their needs. This is how AI becomes a real driver of efficiency without costly consultancy.

The DBT trial shows that licences alone won’t deliver productivity. The value comes when teams learn how to apply AI to their actual work, understand its limits and build practices around it. The organisations that invest in this capability will be the ones that see genuine improvement, not just quicker drafts.