SuperAPI AI Directory

usage-metering-accuracy-review

Office & productivity

View repository

Updated 10-16 02:29

About this skill

Usage Metering Accuracy Review

Use this skill to convert usage-metering, billing accuracy, quota, credit, overage, invoice, and reconciliation questions into a concrete artifact with owners, gates, metrics, and recovery paths.

Workflow

  1. Identify billable units, source events, customer identifiers, plan rules, credit/commit logic, quota windows, aggregation grain, invoice timing, dispute process, and current accuracy risks.
  2. Read references/usage-metering-accuracy-patterns.md.
  3. Classify metering as raw event, derived usage, rating, quota check, invoice line, credit deduction, correction, dispute, or customer-visible estimate.
  4. Define source-of-truth events, idempotency, dedupe, late-event policy, reconciliation, customer transparency, alerts, correction workflow, and audit trail.
  5. Produce metering accuracy plan, state machine, decision table, event schema, control checklist, and dispute/recovery path.

When not to use

  • Do not use for generic advice the base model already handles without this skill's specific artifact contract.

Guardrails

  • Do not invoice usage that cannot be traced from customer-visible billable event to rated invoice line.
  • Do not silently correct usage without audit trail, customer impact assessment, and finance/support notification.
  • Do not mix quota enforcement, billing estimates, and final invoice truth without labeling freshness and accuracy.
  • Do not launch usage-based pricing without dispute handling and metering backfill/replay controls.

Output format

Metering context:
Audience / source of truth / risk boundary:

Metering accuracy plan:
| Area | Decision | Evidence | Risk | Owner |
| --- | --- | --- | --- | --- |

Reconciliation, correction, and dispute handling:
- <trigger> -> <policy, metric, edge case, support note>