Why most Copilot rollouts stall at ten per cent

Copilot gets switched on for the whole company. A few weeks later the usage report tells the same story it always tells: a small group is doing everything, and everyone else has drifted back to the way they worked before.

The reflex is to blame the tool, or the training. Usually it is neither. The rollout stalled because nothing about how people actually work was asked to change.

Access is not adoption

A licence gives someone permission to use Copilot. It does not give them a reason, a habit, or a moment in the day where reaching for it is the obvious move. Those have to be built, role by role, inside real tasks. A salesperson and a finance analyst do not adopt the same way, and a generic session treats them as if they do.

What moves the number

The teams that get past ten per cent tend to do the same unglamorous things. They pick a few high-value tasks per role and make Copilot the default way to do them. They put a visible champion in each team, someone whose own use others can copy. And they measure against real usage, not attendance, so they can see what is landing and double down on it.

Measure value, not logins

Even active use is only a proxy. The number that matters is what the use returns: hours given back, cost taken out, errors avoided. When adoption is tied to a value the business can feel, it stops being an IT project people tolerate and becomes something they ask for.

None of this is complicated. It is just work, and it is the work that turns a licence into a return.