Description
What if one AI prompt returned live QTO, carbon, cost, compliance, and a 4D schedule from your Revit models in Autodesk Forma (formerly Autodesk Construction Cloud) no exports or plug-ins? This session will present a working BIM Intelligence Agent built at the APS Accelerator Bangalore 2026 in four days, connecting to the AEC Data Model GraphQL API via Model Context Protocol (MCP) and Azure GPT-4.1. Ten BIM intelligence tools, all natural language. Real WSP project data live on screen will include: walls 1,860 square meters; floors 6,089 square meters; 380 tCOâ‚‚e carbon; $1.44 million cost; ISO 19650 compliance check; and a 38-week 4D schedule. One conversation, zero manual work. Learn the architecture, undocumented GraphQL patterns, 3LO OAuth flows, and 10 production bugs fixed. Code shared.
Key Learnings
- Learn how to build an MCP-powered AI agent using Azure GPT-4.1 and FastMCP SSE transport connected to the AEC Data Model GraphQL API.
- Learn how to write AEC Data Model GraphQL queries to extract element properties, apply EMEA headers, and handle pagination correctly.
- Learn how to implement automated QTO, embodied carbon, construction cost, and ISO 19650 compliance checks from live Revit model data.
- Discover and learn how to resolve the top 10 production failure patterns in Autodesk Platform Services 3-Legged OAuth, GraphQL, and MCP tool integration.