Work out what your unused Copilot seats are costing. Enter your licence count, what a seat costs you, and an honest estimate of how many people actually use it, and see the annual waste, the cost over your remaining term, and what that money would otherwise buy. No sign up, nothing leaves your browser.
Assumptions last reviewed August 2026
The calculator runs entirely in your browser, so none of these figures are sent anywhere.
The salary figure is used only for the comparison line in the results and is deliberately conservative. All assumptions are illustrative and editable.
This tool gives a directional estimate from one input you had to guess: real usage. The AI Shelfware Audit replaces the guess with evidence. In four weeks, for a fixed fee agreed up front, we measure actual usage across your Copilot seats, AI API consumption and the AI line items inside your vendor agreements, and hand you a board ready report with a prioritised recovery plan. Some of the savings need no negotiation at all. The rest becomes leverage at your renewal, and leverage requires runway, so the earlier you start, the more of this number you keep.
Prefer email? Reach us directly at hello@c4cgroup.co.uk.
The maths is deliberately simple, because the problem is simple. Seats you pay for, minus seats in genuine use, is shelfware. The calculator multiplies your unused seats by your per seat monthly price to give the monthly waste, then projects it over a year and over the remainder of your contract. The salary line converts the annual waste into something a board understands instantly: what else that money would fund.
The only input that needs care is the usage estimate, and it is the one most organisations get wrong in the optimistic direction. A licence that was used once in the rollout month is not in use. For Microsoft 365 Copilot the admin centre usage report shows active users against assigned seats, and that report, not the rollout plan, is the number to put in the slider.
One honest caveat cuts the other way: a seat in daily use by the right person can be worth many times its cost. The point of measuring is not to conclude that Copilot is bad value, it is to stop paying for the seats that return nothing and to defend the ones that earn their keep with evidence.
The defaults are indicative UK list estimates, quoted per seat per month excluding VAT, and were last reviewed in August 2026. Your negotiated rate may differ, so overwrite the price field with the figure from your own agreement for a realistic result.
| Licence | Indicative list price, £ per seat / month |
|---|---|
| Microsoft 365 Copilot | around £24.70 |
| GitHub Copilot Business | around £15 |
| GitHub Copilot Enterprise | around £31 |
GitHub Copilot figures are converted from US dollar list prices at an indicative rate and will move with exchange rates. Microsoft prices its UK list in pounds.
Copilot licences that are assigned and paid for but not in genuine use. A seat used once during rollout and untouched since is shelfware. Because Copilot is priced per seat per month, every unused seat is a recurring cost that compounds quietly over the contract term.
The published UK list price is around 24.70 pounds per user per month excluding VAT, around 30 US dollars in the United States, normally on an annual commitment. Enterprise agreements can move that number, which is why the calculator lets you enter your own negotiated rate.
The Microsoft 365 admin centre includes a Copilot usage report showing active users against assigned licences, broken down by app. That report is the ground truth. Measure genuine weekly active use over at least a month, not the rollout spike, and compare it with what you are paying for.
Measure usage first, then act in two stages. Seats that are unused can be unassigned and left to lapse, a saving that needs no negotiation. The wider gap between licensed and active seats becomes your evidence at renewal, where the licence count, the rate, or both can be renegotiated from data the vendor cannot argue with.
Usually not for a refund. Copilot seats are typically bought on an annual commitment, so the cash saving lands at the renewal or anniversary rather than the day you unassign the licence. That is exactly why timing matters: the earlier you measure, the more of the next term you can correct.
For the people who genuinely use it, often yes. The honest answer is per cohort, not per company: measure who is actively using it and what for, keep and defend those seats, and stop paying for the rest. The mistake is licensing everyone on day one and never checking again.