Microsoft has unveiled a new integration that lets engineering teams ask natural-language questions about their Autodesk project data directly inside Microsoft Teams. The connection, built by Italian engineering group MAIRE, links Microsoft Foundry and Microsoft Copilot Studio with Autodesk Forma and Autodesk Revit, enabling quick, conversational access to complex building information models without leaving the collaboration app.

The Integration at a Glance

The setup marries two distinct Microsoft AI services with Autodesk’s design software. Microsoft Foundry—a data integration and analytics platform—pulls structured and unstructured data from Autodesk Forma (a cloud-based early-stage design tool) and Autodesk Revit (building information modeling, or BIM, software). That data is then surfaced through a custom-built Copilot, created in Microsoft Copilot Studio, that understands engineering terminology and can generate answers, summaries, or visualizations from the underlying models.

Engineers type questions like “What is the total square footage of the lobby on floor 3?” or “Show me all structural steel specifications for the east wing,” and the Copilot responds in seconds. It can also handle more complex requests, such as comparing material quantities across different design revisions or identifying clashes between architectural and structural elements.

The Copilot is surfaced inside Microsoft Teams, meaning team members stay in their flow of work—no switching between apps, no exporting files, no waiting for a BIM specialist to pull a report. The integration works with natural language, so there’s no need to write SQL queries or know the exact field names inside the Revit database.

What This Means for Your Daily Workflow

For engineers, architects, and construction managers who spend hours extracting, formatting, and validating data from design models, this integration promises a significant reduction in busywork. Instead of manually digging through Revit schedules or exporting CSV files, team members can simply describe what they need in plain English. The Copilot understands context, project-specific terminology, and can even be trained on organizational standards.

Home users and small firms may not see immediate benefits, as the integration likely requires enterprise licensing for Foundry, Copilot Studio, and Autodesk’s AEC (Architecture, Engineering, Construction) Collection. However, the technology sets a direction that could trickle down to smaller-scale tools in the future—Microsoft has a track record of eventually bringing Copilot features to more affordable tiers.

For IT administrators, the integration raises new governance questions. Copilot Studio offers granular controls over who can access which data sources, and all queries are logged. Admins will need to configure data connectors, set up authentication between Microsoft’s and Autodesk’s clouds, and ensure that model data is properly indexed and secured. The good news is that the architecture relies on existing identity frameworks (Entra ID) and can be managed through familiar Microsoft 365 admin centers.

Developers at large engineering firms may use this as a template for similar integrations. MAIRE’s solution is not a one-off; Microsoft is actively encouraging other partners to build domain-specific Copilots. The Copilot Studio platform allows low-code customization, so a firm could replicate the pattern for other data sources like SAP, Oracle, or project management tools.

How We Got Here: The AI Inflection Point in AEC

For years, engineering and construction have lagged behind other industries in digital transformation. BIM data is notoriously locked inside bulky desktop applications, and sharing information often means exporting static PDFs or DWG files. The pandemic accelerated the shift to cloud collaboration, but the real enabler for this kind of natural-language querying is the maturing of large language models (LLMs) and Microsoft’s Copilot ecosystem.

The timeline leading to this moment:

  • 2021: Autodesk acquires Spacemaker (later rebranded as Forma) to strengthen its cloud-based conceptual design capabilities.
  • February 2023: Microsoft announces the general availability of Azure OpenAI Service, giving enterprises access to GPT models.
  • November 2023: Microsoft Ignite showcases Copilot Studio, a tool for building custom AI assistants on top of organizational data.
  • Early 2024: Autodesk and Microsoft deepen their partnership, integrating Autodesk Construction Cloud with Microsoft 365.
  • Late 2024/early 2025: Microsoft Foundry (the data platform) becomes more prominent, pitched as a way to unify analytics, AI, and business processes.

The MAIRE project is one of the first tangible examples of stitching these pieces together. It shows that foundational AI services have reached a maturity where they can ingest, interpret, and reason over the highly specialized, graph-like data structures found in architectural models.

The announcement also underscores a broader industry trend: moving from “AI as a chat tool” to “AI as a data concierge.” Instead of asking a generic chatbot, professionals query their own authoritative data sources—with the AI doing the heavy lifting of converting intent into technical queries and then formatting the results in a human-friendly way.

What to Do Now: Steps for Teams Interested in This Workflow

If your organization uses Autodesk and Microsoft tools and wants to explore similar capabilities, there is a practical path forward.

  1. Check your licenses. You will need Microsoft 365 E3 or E5 (for Teams and Copilot Studio), Foundry capacities, and the Autodesk AEC Collection with Forma and Revit. Confirm if your Autodesk data is already in the cloud—Revit workspaces should be saved to BIM 360 or Autodesk Construction Cloud.

  2. Assess data readiness. The Copilot is only as good as the data it indexes. Standardize your model naming conventions, parameter definitions, and family libraries. If your Revit models are a mess of inconsistent parameters, the AI won’t be able to surface reliable answers.

  3. Start a pilot with a controlled scope. Instead of connecting to every model at once, pick a single project and a small team. Work with your IT department to set up the data connector in Foundry and build a simple Copilot Studio bot that answers a few pre-defined questions correctly. Measure time saved on information retrieval before expanding.

  4. Train the Copilot on your vocabulary. Copilot Studio allows you to upload a glossary of terms, abbreviations, and classification systems. Teaching the AI that “W16x31” is a steel beam shape, or that “phase 3” refers to a specific construction sequence, will dramatically improve accuracy.

  5. Establish a governance framework. Decide who can modify the Copilot, which data sources are allowed, and how data access is audited. The integration can surface financial, structural, and proprietary information; treating it like any other line-of-business app is essential.

  6. Watch for official Microsoft guidance. As of now, the MAIRE integration is a partner showcase, not a packaged solution you can download from an app store. Microsoft may release a reference architecture or accelerator in the coming months, which would simplify deployment.

What to Watch Next

This is likely just the opening move in a larger strategy. Microsoft has been aggressively courting the AEC sector, and Autodesk is investing heavily in cloud APIs. Expect to see first-party Copilot extensions for Autodesk products announced at future Build or Ignite events. Also, keep an eye on whether Microsoft evolves Foundry into a more developer-friendly integration platform that could make such connections no-code. For now, the MAIRE example proves the concept works, and it’s a sign that the way we interact with complex design data is about to get a lot more conversational.