AI procedure revision review
Produce a traceable change list between document revisions while keeping approved content separate from drafts and missing approvals.
Synthetic evaluation input
Document DOC-17, revision A: approved; drawing DR-1; appendix X included. Document DOC-17, revision B: draft; drawing DR-2; appendix X removed; approver not stated.
Original Inzonex training example, version 1.0. No customer document or real equipment data. No model has been scored on this page.
Expected fields
| Field | Reference answer |
|---|---|
| document_id | "DOC-17" |
| new_revision | "B" |
| new_status | "draft" |
| drawing_before | "DR-1" |
| drawing_after | "DR-2" |
| appendix_x | "removed" |
| approver | null |
Check a structured response
The comparison runs in your browser against this answer key. It checks exact fields and values, not the accuracy of an explanation or a model's general ability.
Recommended workflow
Compare controlled document fields and classify changes before generating prose. An absent approval remains absent even when the revision number is higher.
A newer revision is not automatically an approved revision. Before asking AI to summarise a procedure change, include document status, approval evidence and the precise scope of the comparison. Otherwise the output can turn a draft into an apparent operating instruction.
The sample changes a drawing reference and an appendix while leaving the document identifier unchanged. The result should identify those changes and retain the draft status of revision B. It should not invent a reviewer or recommend using the new revision.
For production work, keep line-level references and obtain review from the document owner. This workflow organises a change review; it is not an authorisation process and must not directly overwrite a controlled document.
Failure checks
- Do not describe revision B as approved.
- Keep before and after drawing references separate.
- Do not invent an approver.
What this exercise does not prove
Document comparison only. No operating, maintenance or safety procedure is generated or approved by this exercise.
For an actual evaluation, keep a separate held-out set, record tool/model version and settings, and log raw outputs, corrections, elapsed time and actual charges. Do not compare tools tested on different inputs as though they ran the same benchmark.
What is checked in AI procedure revision review?
Produce a traceable change list between document revisions while keeping approved content separate from drafts and missing approvals.
What does AI procedure revision review not prove?
Document comparison only. No operating, maintenance or safety procedure is generated or approved by this exercise.
Related tasks
Match this workflow to your data requirements
Methodology and reuse
These packets and answer keys are original Inzonex educational material, licensed CC BY 4.0. Attribute Inzonex and link to this task page when reusing the packet. The licence does not cover third-party material linked from this site.
NIST AI 600-1: background on generative AI evaluation and risk. This exercise is not NIST-certified.