Description
Autodesk Forma Data Management (formerly Autodesk Docs) and Revit hold powerful data, yet teams still spend hours on model QA, quantity takeoffs, and reporting. In this two-presenter session, we build a practical AI assistant for Revit and Forma Data Management on the Model Context Protocol (MCP), driven from both Claude and GPT. A Revit add-in bridges natural language drives structured, deterministic tool calls instead of unreliable prompt-only behavior.
Key Learnings
- Discover where MCP fits in a Revit workflow across the add-in bridge, tool layer, and LLM client.
- Learn how to design focused Revit tools for QA, element queries, and quantity exports with deterministic inputs and outputs.
- Learn how to assess when natural language improves Autodesk Forma and Revit workflows and when deterministic tool contracts are required.
- Learn how to avoid common failure modes, such as over-broad tools, ambiguous results, and untracked model changes.