1. Identify the mismatch.
Describe the customer action, expected internal result and observed failure. Identify the provider, application owner and authorized project scope.
Process
Trace the money event and the application update.
Review one payment path before changing how the wider system works.
2. Map the records.
Follow the order or invoice, provider result and later events. Note where an update can be delayed, repeated or rejected.
3. Agree the rules.
Define what should happen after success, failure, a duplicate event and a missing response. Name the owner of financial exceptions.
4. Test in isolation.
Use provider test facilities and suitable test records. Check the full application result, including a repeated or delayed event, without charging a real customer.
5. Release and hand over.
Agree the production change and recovery plan. Document verified behavior, monitoring and the steps staff use when records disagree.
Prepare the integration checklist or describe the affected path.