Usage data looks objective, but mostly measures who presses a button. Here's how to measure what people can actually do in five minutes per employee, without an IT project.
The first question we get is almost always: can you see who uses Copilot? You can, through the reporting in Microsoft 365. But that figure says less than it seems. Someone who makes one summary a day counts just as heavily as someone who rebuilt their entire working week.
What we ask instead
- Which task did you do with Copilot last week?
- Where did you stop halfway, and why?
- What do you always check yourself afterwards?
Those answers are subjective. But because we collect them again 60 days later, with the same questions and the same scale, the difference is usable. The single figure isn't, and we don't present it that way either.