How Artisan Is Designed to Centralize, Route and Track Engineering Comments
Artisan is proposed as an AI-agentic workflow for capturing comments from engineering drawings and documents, assigning them to the appropriate owners, and tracking every item through acknowledgement and resolution.
Pilot at a Glance
30 – Drawings and documents in the proposed pilot
Approximately 1,500 – Comments across multiple engineering disciplines
Days to Hours – Target reduction in acknowledgement time
Project Overview
Industry: Engineering, Industrial and EPC
Solution: Artisan – AI-Agentic Comment Resolution Workflow
Status: Proposed Pilot
Pilot Scope: 30 drawings and documents
Estimated Volume: Approximately 1,500 comments
The Challenge
Engineering comments were distributed across drawings, models, specifications, datasheets, emails and multiple document revisions.
Key Challenges
- Comments distributed across multiple documents and systems
- Manual comment identification and logging
- No centralized register of ownership and status
- Risk of using outdated document revisions
- Time-consuming assignment and follow-up
- Delayed or unresolved comments
- Potential delays to engineering and procurement
- Increased expediting costs and commercial exposure
- Potential effects on revenue recognition and liquidated damages
Before Artisan
Documents and Emails → Manual Comment Extraction → Spreadsheet Register → Manual Assignment → Repeated Follow-Up
This fragmented process limited visibility, weakened traceability and increased the risk of comments remaining unresolved.
The Proposed Solution
Drishya AI proposed Artisan as an AI-agentic comment-management workflow.
The system is designed to monitor documents, extract and classify comments, assign them to the appropriate owners and provide centralized status visibility.
The Proposed Solution Includes
- Monitoring documents from Wrench and email inboxes
- Automated comment extraction
- Classification by discipline, type and document creator
- Human review of low-confidence classifications
- A single canonical comment register
- Links between comments, documents and revisions
- Automated routing to mapped owners
- Context-rich owner notifications
- Acknowledgement and status tracking
- Automated reminders, escalations and reassignment
How Artisan Would Work
Documents → Comment Extraction → Classification → Owner Assignment → Status Tracking → Resolution
Artisan would monitor incoming documents, identify and classify comments, route them to the appropriate owners, and track each item through acknowledgement and resolution.
| Before | Proposed Workflow |
|---|---|
| Manual extraction | AI-assisted extraction |
| Spreadsheet tracking | Canonical comment register |
| Manual classification | Automated classification with human review |
| Unclear ownership | Rules-based owner assignment |
| Repeated follow-up | Automated reminders and escalations |
| Limited revision traceability | Comments linked to documents and revisions |
| Delayed reporting | Live ownership and status visibility |
Pilot Scope and Expected Outcomes
30 Drawings and Documents
The proposed pilot would evaluate Artisan across 30 engineering drawings and documents.
Approximately 1,500 Comments
The pilot would include comments across piping, mechanical, electrical and instrumentation, structural, process and other engineering disciplines.
Acknowledgement Time: Days to Hours
The intended outcome is to reduce comment-acknowledgement time from days to hours.
Additional Intended Outcomes
- One canonical register for engineering comments
- Live visibility into ownership and status
- Traceability across documents and revisions
- Automated routing, reminders and escalations
- End-to-end acknowledgement tracking
- Faster responses to comments affecting engineering and procurement
- Greater accountability across disciplines
Important Status Note
This initiative is a proposed pilot. The stated improvements are intended outcomes and have not yet been demonstrated or validated at scale.
Key Takeaway
Artisan is designed to transform engineering comment management from a manual review and follow-up process into a centralized, traceable AI-agentic workflow—subject to validation through the proposed pilot.
