Every read Aristo needs is granted (5 of 5); connect the Cost Management export next and your dollar figures turn from estimate to measured.

How many seats show qualifying Copilot activity, how many do not, and the activation percentage every workspace leads with.

Seats bought versus seats assigned, per SKU, and the seats already inside a renewal grace window. The denominator under every cost-per-seat figure in the product.

Which department each person belongs to, so the program can be reported by team. Aggregate only — no per-person row is ever built from it.

Which teams hold the seats and which of those seats are working — the whole Users & teams workspace and the department table on the money pages.

The Copilot rollout radar. Two feeds join one table and every row is stamped with which of the two it came from.

How many people signed into a third-party AI application with a work identity — the unsanctioned-AI reading on Security & risk.

The tenant's own catalog of declarative agents on the AI agents workspace.

The actual dollars for Cowork and credit consumption, per month, from the customer's own billing export. This is the source that upgrades the money pages from an envelope estimate to a measured figure.

The seat position per team month by month (the department table's change and trend columns) and paid seats per product month by month. The dollar stack by product remains the sample organization's instrument: metered consumption is not attributed to a product family for a real organization.

Credit consumption per billing policy — the run rates behind the spend lights and the credit share of the money trend.

The seat cost every return figure is multiplied by, the monthly Cowork cap, and the enforcement mode beside it.

Every month-over-month comparison in the product: what changed, which direction, and the number of the newest stored run the reading on screen is compared against. Since 21 August 2026 each run also records the seven chargeback quantities the KPI strips report — paid seats, seats waiting for a user, department rows, averageable departments, charged-back headcount, charged seats and the seat price they were costed at — so those cards can draw a movement instead of naming a missing figure. Each comparison identifies whether its records are measured or illustrative.

Which recommended moves this organization has taken up, and how many have been completed.

How many agents this organization has built, who made each one, and the capability scope each maker attested to.

How many skills are published for this organization and how many times they have been taken.

Every license the organization holds beyond Copilot, so the chargeback view can show what else is being paid for.

The organization's name, as it appears at the top of every page and on every report.

The "as of" stamp in the top bar and on every figure that has no fresher stamp of its own.

The estate score and its band, computed from the pillars above it. Each pillar carries its own provenance — measured, or self-reported — and the score states what it was computed over.

Who is credit-ready, the suggested capped funding, and the tier a leader is asked to act on.

The narrative sentence at the top of the product and the ranked moves under it, each with the counted figure that justifies it.

Seats waiting for a user, the cost of licenses waiting for a user, and the cost-per-working-seat figure the chargeback pages are built on.

Returned value, protected spend and the proof boundary used by the Value Report. Counted, attested and causal inputs remain distinct rather than being totalled into one unsupported claim.

The budget traffic lights, and the month-by-month spend trend under them. Each light carries the basis of the number it was computed from, and each month of the trend is drawn solid when measured and hatched when estimated.

Where this month's spend is heading, on the analytics workspace.

The "on pace to cross into the next band" sentence, where there is enough history to support one.

Every permission table in the product: what this tenant has actually consented to, decoded from the roles claim in its own app token.

The lifecycle of every finding this product has raised — its status, its owner, and every change made to it.

Every generated report and its frozen figures, plus the schedules that generate them. Each stored report carries its own per-figure freshness and a digest of the figures it froze.

The auditability page: every act that changed money, licenses, governance or configuration, with the role that carried it out.