usecase
Board Reporting
Fleet reports built from receipts, not from asking around
The board asks what the organisation's AI actually does and what it costs, and in most companies the honest answer is a survey. CueCrux replaces the survey with a query. Because every agent action mints a signed receipt and every credit spend mints one too, the board pack derives from the spine: your agents did 4,183 things this period, and every one of them resolves to an agent, a verb, an identity, and a cost. Nothing estimated.
The interesting rows are often the exceptions. Signed RefusalReceipts show where policy stopped an agent; gate records show which high-risk actions waited for a named human and how long approval took; revocations show as signed pulls with per-agent acknowledgements. Together they give a board the thing governance codes actually ask for: evidence that oversight operates, not a statement that it exists.
Period-on-period comparison is a diff over signed records, so trends in fleet activity, spend, and refusal rates are defensible numbers. And because the finance line reconciles to receipts down to individual metered credits, the audit committee's follow-up questions have answers of the same quality as the headline.