Measuring Microsoft 365 Copilot adoption beyond license activation

We stopped paying much attention to active users pretty quickly. Someone who tried Copilot once in a Teams meeting counts the same as someone who now writes every status report with it, so past the first month that number doesn't tell you a lot.

What worked better was picking a handful of recurring tasks per team (meeting recaps, first drafts of customer replies, the weekly report) and asking about those directly. We just send a quick survey every month or so asking "which of these did you use Copilot for this week" with checkboxes. The usage reports can tell you someone was active in Word, they can't tell you what for.

For repeat use I'd look at active days per person over several weeks in the Copilot Dashboard rather than first use. And the people who were busy in week one and gone by week five are the best ones to go talk to. Usually they didn't know it worked in a particular app, or the output needed so much fixing it wasn't worth it.