Copilot licences are not adoption

Buying Microsoft 365 Copilot is a procurement decision that takes an afternoon. Getting an organisation to use it well is a change programme that takes a quarter. Companies routinely budget for the first and not the second, then conclude the tool does not work.

Where the seats go quiet

The pattern repeats across organisations. Licences land in everyone's account on the same day with an email announcing them. A handful of curious people try it, get a mediocre first answer, and go back to what they were doing. Nobody has told them which of their weekly tasks Copilot is actually good at, so they test it on the wrong things.

Pick the first use cases deliberately

A rollout works better when each department starts with two or three tasks chosen because the tool is genuinely strong at them. Meeting recaps, long email threads, first drafts of recurring documents. Small, frequent, low-risk work where a mediocre draft still beats a blank page. Wins there buy the patience needed for harder use cases later.

• ✳Choose use cases per department, not one company-wide announcement
• ✳Start with frequent low-risk tasks where a rough draft is still useful
• ✳Write the data rules before the licences land, not after the first incident
• ✳Track weekly active use, not seats assigned

Measure something other than the licence count

Seats assigned is a purchasing number, not an adoption number. Weekly active users per department, and hours saved on named tasks, are the figures a board can act on. They also tell you which departments need a second session before the renewal date arrives.