Abstract
Field
Computer-implemented systems for maintaining employee competency credential validity. This publication describes a two-state model (valid/invalid) with document-level change detection and cryptographic hash verification. It does not use proposition-level granularity, multi-state lifecycle management, or validity bitmaps.
Background
In regulated industries, organizations must keep competency records showing that employees have been trained on the current version of authoritative reference materials: standard operating procedures, regulatory guidelines, product specifications, and similar documents. When a reference document gets updated, employee credentials linked to it may need invalidation followed by retraining. Commercial systems like Veeva Vault Training Impact Assessment and MasterControl Training Management handle this at document-level granularity. Any change to any part of a document triggers retraining for all employees linked to that document, regardless of what specifically changed.
Technical Description
The system maintains one credential record per employee-document pair. Each record contains: credential_id, employee_id, document_id, document_version_at_certification, certification_timestamp, credential_status (an ENUM with two values: valid or invalid), content_hash (SHA-256 computed over the full document text at certification time), and invalidation_timestamp (set to NULL when status is valid).
Here is how document change detection works. When the system ingests an updated document version, it computes a SHA-256 hash over the full text of that updated document. It then queries all credential records where the document_id matches and credential_status is valid. For each one, the system compares the stored content_hash against the new hash. If they differ, the document changed, and the system transitions credential_status from valid to invalid, writes the current time into invalidation_timestamp, and generates a retraining obligation record. That record includes: obligation_id, employee_id, document_id, new_document_version, old_document_version, and a due_date calculated from a configurable per-document-category deadline (30 days for safety-critical content, 90 days for general content, and so on).
The retraining obligation gets routed to the employee's learning management system or to a manager notification queue. When the employee finishes retraining (confirmed by an LMS completion event, a passing assessment score, or a manager acknowledgment), the system creates a fresh credential record with the updated document version, computes a new content_hash over the current document text, sets credential_status to valid, and archives the old invalidated record.
Cryptographic verification runs separately from document change events. A periodic verification job (daily, for instance) iterates through all active credential records. For each one, it recomputes SHA-256 over the stored document text matching that credential's document_id and document_version_at_certification. If the recomputed hash doesn't match the stored content_hash, something has been altered since certification time, whether through unauthorized modification or data corruption. The system immediately invalidates the credential regardless of whether any formal document update event was received.
An organizational compliance metric is computed as: compliance_rate = (count of valid credentials) / (total active employee-document pairs). The system calculates this per department, per document category, and organization-wide, then surfaces it on a manager dashboard.
Distinguishing Characteristics
This is a binary two-state model. Valid or invalid. There is no intermediate review state where a materiality evaluation determines whether a change was editorial or substantive. There is no proposition-level validity bitmap that would allow partial invalidation of a credential while keeping other parts intact. There is no asymmetric transition complexity where degradation takes one trigger but restoration requires multi-factor verification.
Document change detection looks at the whole document through full-document hash comparison. The system does not segment documents into authority units. It does not compute per-authority-unit embeddings. It does not run section-level semantic differential analysis. And it does not perform reverse-index traversal to figure out which employees were specifically assessed on which specific sections that changed. Everyone linked to a changed document gets invalidated, period. If 200 employees are assigned to a document and only 2 paragraphs changed affecting 4 of those employees, all 200 still get retrained. That is the over-assignment problem. But the implementation is simple, and there is no dependency on embedding models or section-level indexing infrastructure.
Retraining requires the employee to complete the full updated document. There are no targeted delta challenges scoped to just the propositions that changed. No invalidation proof object specifying exactly what changed and what the employee previously answered. No append-only transition journal with hash chain integrity.
This approach works best for organizations with moderate document update frequency, where the cost of retraining employees on unchanged content is acceptable compared to the complexity of building section-level tracking.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Davis, Kenneth, "Binary Credential Validity Model with Document-Change-Triggered Invalidation and Cryptographic Hash Verification", Technical Disclosure Commons, (March 25, 2026)
https://www.tdcommons.org/dpubs_series/9624