AI for design
Two agents engineers can sign off.
AI agents that work inside the tools engineers already use. They read the model, call proven automation through MCP rather than improvising, and hand every change back to an engineer for approval.
BIM AI agent
An agent that understands the model — and runs the work.
Engineers inspect views, create elements and query references in plain language. The agent plans the steps and executes them through proven automation tools across the Autodesk and CSI ecosystems.
- AskA request in plain language, inside Revit
- PlanThe agent sequences the automation tools
- ApproveThe engineer reviews before any model change
- Revit · Navisworks · CAD
- SAP2000 · ETABS · SAFE
- MCP connectors to proven tools
Design AI agent
A team of specialised agents on top of the structural model.
It reads the BIM and analysis models side by side, checks them against codes and company standards, and flags discrepancies before a reviewer does — with answers referenced to their source.
- ReadBIM and analysis models, plus the knowledge base
- CheckDesign review, quality and code-compliance agents
- ValidateThe engineer remains the design authority
- Model understanding
- Code compliance & QA
- RAG query over standards
- Learns from expert feedback
SUniform data schema
Automation chained to one indexed data structure — a library, not a scavenger hunt.
DCentralised design data
Everything stored with project and model indexing, and every change tracked.
CCorrelation is key
When data speaks one language, AI stops guessing and starts connecting.
HHuman in the loop
Validation belongs to engineers and architects. Approval cannot be skipped.
