Why do Copilot and ChatGPT Enterprise licenses go unused, and what fixes it?

It loses them in the weeks after go-live. A typical enterprise AI launch is planned as a technology project: licenses bought, identity and data controls set, a launch message sent, and a training session offered. All of that is necessary, and it ends on the day the work is supposed to begin.

Gallup's indicator for U.S. employees, as of May 2026, found that 47% say their organization has integrated AI, while 30% use AI a few times a week or more (Gallup, 2026). AI is arriving faster than the habit.

It leaves out the decisions that sit with the business. Which recurring task in each team should AI help with first? What does good use look like on that task? What may people put into it? When it saves an hour, what is the hour for? A launch plan does not answer these, because the people who write it do not own the work.

The AI Profitability Gap™ explains why this matters for the return. The AI Profitability Gap™ is the distance between what an organization spends on AI and the commercial return it produces. A license report measures the first two links of the chain, Investment and Usage. The return is decided by the next three: Direction, Skill and Reinvestment. An idle license is a Direction and Skill problem showing up as a usage number.

Run it team by team, with the leader who owns each team's work in charge.

1. Set a goal for the team. Name the business number AI is meant to help move this quarter, and the recurring task where you will start.
2. Take a baseline. Count how many people use AI on that task today, and how long the task takes.
3. Learn on the job. Give people a short weekly task on their real work, and let the person who uses AI best show the others.
4. Change one step of the process. Put AI inside the work, so using it removes a step from the task.
5. Decide where the time goes. Write down what the saved hours are for, and check it at the next team meeting.

If you need to show something within a month, the 30-day plan for one team is the short version of these steps.

The leader who owns each team's work owns that team's adoption. IT owns the platform, the access and the data rules, and holds the usage data everyone will ask about. The CEO owns what the AI spend is for. When these sit with one function, the usage number gets reported and nobody is in charge of moving it. The CIO role page sets out the split from the technology side.

Measure what changed in the work. For each team in the second launch, bring three numbers to the renewal conversation: how many people use AI on a recurring task, before and after; how long that task takes, before and after; and where the saved time went. Teams where the work changed have a case for renewal. Teams where it did not have a reason to reduce seats, decided on evidence.

Finance will see the renewal as a cost line with a usage figure beside it. Our article on the AI license renewal question nobody can answer covers how to answer it. In the AI Profit Bridge, this is the AI That Earns Its Keep step: every AI investment tied to a goal.

Less than the launch does. Each product has strengths, and your IT team will have good reasons for its choice. The pattern above applies to any AI assistant a company licenses, because the missing piece sits in the work around the AI. Switching vendors without fixing it reproduces the same usage report a year later.