The License Report · Contoso

Read from your Microsoft subscription list under one read-only consent. A subscription’s bought and assigned counts are procurement totals — seats, never people — so this page is safe to project in any room.

counted Procurement totals read from your Microsoft subscription list — a SKU's bought and assigned counts are seats, never people. Free and trial subscriptions are left out: an unbought seat is not money. The bar above puts the same two totals on one scale: seats in somebody's hands, and seats that have never been given to anyone.

So what: most organizations answer this from a spreadsheet somebody last touched at the previous renewal. This is your live subscription list, read from Microsoft, in the order that matters — the subscriptions with the most seats waiting for an owner, first.

You can now decide: you can state your true licence estate in one sentence, and you can stop buying a seat type you already own spares of.

Cold seats come from Microsoft’s own Copilot usage report — a seat with no genuine activity in the last 28 days, not merely a signed-in session. Shelf seats come from the subscription list. The traffic light is lifted, unchanged, from the same engine that runs your consumption console.

Assigned, quiet this period — $38,400 a month. Someone holds the seat and has done nothing with Copilot in the last 28 days. These seats are ready to work — a warm activation push is the first move, and it is the cheapest one.

Never handed to anyone — $4,800 a month. Bought and paid for, sitting on the shelf. Nobody has to give anything up for these — they can go straight to the people already asking.

Working and quiet counts are counted per team — Microsoft's own Copilot usage report joined to your directory, aggregated to team totals. No row here is derived by dividing an org number.

So what: the villain here is the missing view, not your people. Nobody chose to leave a seat quiet; until now nobody could see which seats were. These are seats ready to work, and the teams below are where the next warm introduction lands.

You can now decide: which teams get the next activation push, and which shelf seats can be handed out this week without anyone giving anything up.

A seat becomes reroutable only after it sat in the rescue arm for 45 or more days of genuine, warm help and stayed cold — the never-nudged control group is never judged this way, because it was never helped. Demand is counted from people who raised their own hand.

How the reroute protects your people. The teammate who gives up the seat is never named here — only the seat type, and the recipient (and them only if your org turned on named recognition). A leader approves a seat type moving to a person who asked for it; the ledger keeps the receipt.

So what: a reclaimed seat with nowhere to go is just a smaller bill. A reclaimed seat with a name waiting behind it is the same money doing work. This page only ever moves a seat toward counted demand — someone who asked, unprompted.

You can now decide: whether to keep the plan running as it is. The target was frozen on approval, so every later read of this report measures the same promise — and each individual seat move stays a separate, verified approval.

Microsoft exposes no renewal date, so the countdown runs on the date your own team provided and is recomputed on every read — it can never quietly age. The one expiry signal Microsoft does expose is the post-lapse grace window, and that is counted straight from your subscription list.

The countdown. provided Recomputed on every read from the date your team gave Aristo, so it can never quietly age. Microsoft's subscription list carries no renewal date to read — the one forward expiry signal it does expose is the post-lapse grace window, and that is counted.

The recommendation is an upper bound, on purpose. Renewing 3,497 of 4,200 seats is a recommendation, not an instruction: the rescue-resistant and surveyed signals can overlap, so Aristo deliberately recommends releasing fewer seats than the evidence would allow. Verify before you cut. Nothing in this report cancels a licence on your behalf.

So what: procurement posture is the whole prize here. Walk into the renewal with counted keep-and-release evidence per team, and the conversation stops being a negotiation about a guess.

You can now decide: how many seats to renew, and which teams you can say that about with evidence rather than instinct.

The left bar is counted seats multiplied by your seat price. The middle is list price from this platform’s own pricing constants. The right is a capped monthly envelope for the people whose counted usage already says they are ready — you set the cap, in your own admin center.

Station 1 · Idle licences recovered — $43,200 a month. 1,440 paid Copilot seats doing no work. counted $518,400 a year.

Station 2 · The AI strategy — $2,237 a month. $2,000/mo platform base + $3 × 79 holdout-proven activated seats.

Station 3 · Cowork credits for 1,120 ready people — $33,600 a month. A capped monthly pilot for the 1,120 people whose own counted usage says they are ready for metered agent work. estimate Capped by you, in your own admin center.

So what: this is the line a CFO can defend out loud: the money that pays for the AI work is already inside the Microsoft bill, pointed at seats nobody is using. The recovered spend then funds the people your own numbers say are ready for metered agent work — so the saving turns into capability instead of evaporating into the general fund.

You can now decide: approve the reassignment plan, and point the recovered spend at the credit-ready cohort rather than back into the general fund.