Business ownership
- What outcome does the handoff support?
- Who owns the source, destination, and exception queue?
- Which system wins when records disagree?
Interfaces and access
- Are supported APIs, events, exports, files, or database interfaces available?
- What authentication, authorization, rate limits, and vendor approvals apply?
- Which environments and test accounts exist?
Data contract
- What fields, identifiers, formats, enumerations, and relationships move?
- How are duplicates, updates, deletions, and historical records handled?
- Which sensitive data should not cross the boundary?
Timing and reliability
- Is the transfer synchronous, event-driven, scheduled, or manual?
- What latency is actually required?
- How are idempotency, retry, ordering, and partial failure handled?
Validation and reconciliation
- What makes a record valid?
- Where do rejected or ambiguous records appear?
- How can an operator compare systems and safely retry?
Monitoring and change
Define useful alerts, health evidence, audit history, vendor-version monitoring, credential rotation, deployment ownership, and a procedure for changing the contract. An integration without visible exceptions is an operational blind spot.