AI System for Checking the Distribution of Project Revisions

The problem

An outdated revision may be used on site during day-to-day operations. When documents, revisions, dates, disciplines, distribution records and receipts are scattered across different screens, documents or team records, issues do not become visible in time.

This fragmented data delays preparation of the revision distribution discrepancy list and requires staff to search for supporting records again by hand. Before making a decision, the document controller and project manager have to complete missing fields and verify conflicts one by one.

A common exception is that a cancelled work package may have been deliberately removed from the list. If an incomplete or incorrect workflow runs without checking this context, work may proceed from an outdated drawing or an unnecessary distribution may be made; the output therefore cannot be applied directly.

What we set out to improve

• Number of verified missing distributions
• Rate of outputs corrected or rejected by the document controller and project manager
• Time between the source record and the review output

How the system works

1. 01 Define the review scope and the authorised document controller and project manager roles
2. 02 Receive the inputs: document, revision, date, discipline, distribution record and receipt record
3. 03 Check source identity, date, version and data freshness
4. 04 Identify the current version
5. 05 Compare recipients with the distribution record
6. 06 Check the exception: a cancelled work package may have been deliberately removed from the list
7. 07 Place low-confidence or conflicting records in a separate human review queue
8. 08 Write the source identity, rule, output and time to the audit trail
9. 09 Have the document controller and project manager approve, correct, defer or reject the list

Methods we used

• A revision state machine and role-based distribution
• Source, format, version and required-field checks for the document, revision, date, discipline, distribution record and receipt record
• A missing-data warning instead of a prediction when inputs are absent or conflicting
• A separate check for the exception outside the main rule
• Joint storage of the source identity, rule result, revision distribution discrepancy list and human decision

Where people stay involved

Data and security

Access to documents, revisions, dates, disciplines, distribution records and receipt records is restricted to the minimum permissions required for each role. Identifiers and sensitive fields are masked where possible, and only the necessary section of a file is sent to the model rather than the complete raw file. Access, output and the reviewers' decision are logged; production use does not begin until the organisation has approved the retention period and data location.

Who this suits

Frequently asked questions