Separate input completion from verification success when deciding whether a workflow is ready…

Separate input completion from verification success when deciding whether a workflow is ready for human review. Required evidence can be present even when an automated check rejects it or a provider fails. Reusing a success-sensitive completion predicate can strand the very cases that need review. Conversely, having all inputs is not always proof that the user pressed final submit.

For staged rollouts, distinguish explicit final submission from legacy records with no marker. Backfills should require positive proof of incompleteness before demoting legacy review states, preserve current authoritative decisions, and update derived status projections atomically without a global version bump when a targeted repair suffices. This applies to workflows whose review action requires a completed submission; it does not imply delaying independently terminal safety failures.

Illustrative example — 2026-09-15

A synthetic form requires a photo and consent. A submitted photo rejected by an OCR provider is still present input; missing consent is unfinished input. Final submission with both inputs can proceed to review despite the OCR rejection. A legacy record lacking a newly introduced submit marker needs additional evidence before a repair calls it unfinished.

This is a design lesson with a synthetic example, not a report of a deployed system or production repair. Compatibility rules and authority invalidation must be checked against the specific workflow.