Workbench / Issues

Human Review and Automation Disagree

A human review result and an automated AI or workflow result disagree without a declared rule for resolving the difference.

What This Looks Like

A human reviewer and an automated AI or workflow step produce different results for the same case. The human may approve what automation rejects, automation may escalate what a reviewer clears, or the AI may classify a case differently than a human decision without a clear resolution rule.

Why It Matters

Human review and automation often sit in the same control path. If they disagree without a declared authority rule, the workflow cannot reliably know which result governs. This can lead to duplicated review, inconsistent escalation, skipped approvals, or unresolved cases.

Structural Signal

Two review authorities produce non-equivalent states, and the workflow lacks a stable reconciliation rule. The issue is not simply that one party is wrong; it is that disagreement between human and automated review is not structurally resolved.

Common Triggers

When to Use This Issue

Use this Issue when human review and automated AI or workflow review disagree and the system does not clearly declare which result controls or how the conflict is resolved.

When Not to Use This Issue

Do not use this Issue when the human reviewer is simply correcting an AI mistake through a clear override path. Do not use it when only one review source exists. This Issue applies when disagreement between review sources creates an unresolved workflow state.

Category

Workflow & Handoffs

Primary Pattern

PAT-0100 — Authority Collision

Declared Patterns

Derived Primary Lenses

Derived Secondary Lenses

Search Intents

Ontology Metadata

Code
ISS-0036
Version
ISS-0036@0.1.0
Ontology release
0.1.0
Updated
2026-05-10T00:00:00Z

History

No public history entries recorded.

View full ontology changelog