# Enterprise AI acceptance brief: account revenue diagnostic

Example task: explain a movement in contracted monthly recurring revenue using billing snapshots and account-level support context.

## Required inputs
Two cutoffs, one currency, explicit contracted-MRR definition, subscription-to-account mapping and a support window. Use the supplied synthetic example for evaluation before connecting production data.

## Demonstration
1. Reproduce opening $10,000 and closing $9,000.
2. Reconcile new +$400, expansion +$300, contraction -$700 and churn -$1,000. Residual must equal zero.
3. Preserve account grain when adding support records. Explain why joining raw tickets yields a false 12% decline.
4. Identify Aster/Cedar as a counterexample: three tickets each, opposite revenue movements.
5. Provide the records behind the conclusion and the questions that remain unanswered.
6. Re-run with a changed source record and show the revised calculation without overwriting the prior approved output.

## Delivery decision
Record which components are covered by an existing product, which require implementation, deployment requirements, code ownership and recurring operating responsibilities. Evaluate the demonstrated work, not a generic agent label.

Source example: https://whiteboxx.ai/work/revenue-diagnostic/
