Operate and recover Human Review
Follow a case from assignment through the application’s reported result. Inbox, email, delivery records, and audit history show different parts of that process. Only the decision recorded in VeriProof is authoritative.
Operating flow
Confirm readiness
Review reviewer eligibility, routing, callback registration, signing keys, SDK version, and application health. Test the complete path with non-production data.
Route the case
VeriProof assigns the case to an eligible person or claimable group. If assignment fails, keep the case open for administrator repair.
Notify the reviewer
Inbox is the authoritative work list. Optional email links to the authenticated task but does not grant decision authority.
Record the decision
The eligible reviewer examines the current evidence and submits the supported outcome. Application-visible answers remain separate from private reviewer notes.
Store the callback
The application verifies the signed raw request, validates every binding, stores the event durably, and returns its receipt. A receipt means stored, not applied.
Apply and acknowledge
The application worker revalidates policy and current state before applying the outcome once. It then reports the actual execution result to VeriProof.
Find the responsible team
| Condition | Owner | Response |
|---|---|---|
| Case cannot be assigned | Administrator or review coordinator | Repair eligibility, assignment, or fallback without changing the decision. |
| Reviewer email failed | Administrator | Keep Inbox active and repair only the notification path. |
| Callback failed | Application owner or continuation operator | Repair the endpoint, key, network, or receiver before replay. |
| Application confirmation is missing | Application team | Check the stored receipt, worker, revalidation, protected action, and acknowledgment. |
| Signing key needs rotation | Administrator and application team | Add, verify, and test the replacement before retiring the previous version. |
Repair and replay
Before replay, confirm the tenant, application, environment, event, delivery, current case state, and callback registration. Correct the receiver, network, or signing configuration that caused the failure. If connectivity or signing changed, run a test that cannot resume customer work.
Replay the original event identity. Do not create a second decision or protected side effect.
Never repair delivery by disabling signature checks, widening tenant scope, accepting an expired key, or reporting work the application did not complete.
Retirement and offboarding
Before retiring an application or organization:
- Stop new review requests.
- List open cases, paused work, callbacks, keys, incidents, and retention duties.
- Agree who will complete, reroute, cancel, or preserve each open item.
- Revoke delivery and application credentials at the approved point.
- Export and retain authorized evidence.
- Check the final records against that plan and record the accountable person’s approval.
Current workflow boundary
The workflow uses Inbox, optional personal email, one final employee reviewer, a signed application callback, durable receipt, and execution acknowledgment.
Advanced approval patterns and decisions submitted through external collaboration tools are not part of this workflow.