AI Review
AI Review analyzes your engineering drawings and identifies potential issues before they reach manufacturing.
How It Works
When you run an AI review on a drawing version:
- The AI analyzes each page of your drawing
- It identifies potential issues like missing tolerances or unclear dimensions
- Issues appear as comments pinned to specific locations on the drawing
- Each issue includes a confidence level and suggested fix
What It Checks For
The AI review looks for common issues in engineering drawings:
- Missing or incomplete tolerances
- Unclear or ambiguous dimensions
- Missing surface finish specifications
- Inconsistencies between views
- Missing material or heat treatment callouts
Credits
AI reviews consume credits from your team's subscription. The number of included reviews depends on your plan. See pricing for details
One credit is used per review, regardless of the number of pages. Follow-up questions with Pen are included — no extra credits. Credits reset monthly on paid plans.
Review Profiles
Review profiles let you tailor the AI's focus to your specific manufacturing process or industry. Choose from starter profiles like CNC Machining, Sheet Metal, Aerospace, and more — or create your own custom profiles on paid plans.
Select a profile from the dropdown next to the AI Review button in the drawing viewer, and manage your profiles on the AI Review page. Browse all profiles in the Profile Library.
Confidence Levels
Each AI finding includes a confidence level to help you prioritize:
Very likely a real issue that should be addressed
Probably an issue, but please verify
Possible issue, may be a false positive
Follow up with Pen
Once an AI Review has run, you can tag @pen in any comment on that drawing to ask follow-up questions — "Is this tolerance achievable on our CNC mill?" or "What does this GD&T frame mean?" Pen answers with full context of the drawing and the review findings. Learn more about Pen.
AI Review is a tool to assist human reviewers, not replace them. Always verify AI suggestions with engineering judgment. The AI may miss issues or flag false positives.
