Automated model checking
Check models against defined rules, with traceable findings your engineering team can review.
We develop rules engines for repeatable model checks. AI can assist with drafting rules from documents, classifying findings and routing them for review.
Approved rules determine the check result. Each finding can be traced to its rule and model version, while engineering review and responsibility remain with the project team.
What you get
- Model checks based on defined rules, with results your team can review
- Rule drafting, version management and approval before use
- Checks run in Revit, on a server or through APS Design Automation, as required
- Findings delivered as ACC Issues, dashboards or exports
- Each finding linked to the rule version, model version and supporting evidence
The role of AI
AI assistance
- Drafting rule text and mapping it to structured checks
- Classifying findings and routing them
- Structuring requirements from documents for review
Engineering safeguards
- Final results determined by approved rules
- Results traceable to the applicable rule
- Engineering review and responsibility retained
Examples of model checks
Model standards
For example, check required parameters and naming conventions before a model is shared. The checks are defined for the project.
Spatial constraints
Check defined clearances, spatial relationships and constructability requirements.
Checks against acceptance criteria
Evaluate agreed acceptance criteria and attach the supporting evidence to each result.
Reinforcement
Rule-based reinforcement checking and placement is the focus of our own product, Otsara, currently in development.
Common questions
Is this AI deciding compliance?
AI assists with drafting and classification. Approved, deterministic rules determine the result, and your engineers review the findings and retain responsibility for the design.
Can checks run automatically on ACC models?
Yes. APS can process models from ACC and return findings as Issues within the coordination workflow.
How do you reduce incorrect findings?
We define the evidence each rule needs, validate checks against representative models and review unexpected results before releasing updated rules.