Definition
Percentage of documented information within the management system that has a verified provenance trail identifying whether content was human-authored, AI-generated, or AI-assisted, and whether it completed the defined review and approval process per Clause 7.5.
The Measurement Gap
ISO 55001 Clause 7.5 requires that documented information be created, updated, and controlled to ensure it is suitable, adequate, and available when needed. The standard requires that documented information be reviewed and approved for suitability and adequacy. When AI tools generate or substantially modify management system documents (risk assessments, inspection reports, procedure revisions, competence evaluations, management review summaries), they introduce content whose origin, basis, and approval status may be indistinguishable from human-authored content. Legacy document control metrics such as document review currency rates and approval workflow compliance percentages measure whether documents went through the defined process but cannot detect whether AI-generated content entered the document control system without provenance metadata. An AI tool that drafts a risk assessment, which a human then reviews and approves, has followed the workflow. But if the organization cannot identify which documents contain AI-generated content, it cannot evaluate whether its review process is adequate for AI-produced material, whether AI-generated content introduces systematic biases or errors, or whether AI is producing documented information that meets the standard's "suitable and adequate" requirement. The consequence is a document control system where the organization does not know the authorship basis for its own controlled documents, making it impossible to assess the reliability of its documented information foundation.
Decision Supported:
Does the organization know which management system documents were AI-generated or AI-assisted, and can it demonstrate that all such documents completed the proper review and approval workflow?
Formula
Plain Language: Count all documents in the management system's document control scope that have verified provenance metadata (authorship type identified as human, AI-generated, or AI-assisted) and a completed approval workflow. Divide by the total number of controlled documents.
Notation:
DPR = (Fully_traced_documents / Total_controlled_documents) x 100
Variables:
| Variable | Definition | Unit |
|---|---|---|
| Fully_traced_documents | Number of documents in the document control scope that have both: (1) identified provenance (human-authored, AI-generated, or AI-assisted as a classified attribute) and (2) verified approval workflow completion appropriate to the document type. Both conditions must be met for a document to count as fully traced. | Count |
| Total_controlled_documents | Total number of documents within the management system's document control scope as defined by the document register. Includes policies, procedures, work instructions, risk assessments, audit reports, management review records, competence records, and any other documented information required by the standard. | Count |
Sign Convention: DPR is a percentage from 0% to 100%. A score of 100% means every controlled document has identified provenance and verified approval. Higher values indicate stronger document control integrity in the context of AI content generation.
Proposed formula, pending validation pilots.