A few days ago, the annual Autodesk University, the "Design and Make" conference organized by Autodesk, a global leader in design and manufacturing technologies based in Munich, Germany, took place. Experts from the fields of design, engineering, construction and creative industries came together this year in Nashville, Tennessee, to exchange ideas, learn from each other and discuss the future of design and manufacturing, among other things. The event offered numerous highlights, especially for participants from the D&M sector.

Reduce routine work, accelerate iterations, achieve production maturity

The main attraction was the AI functions presented, which are set to redefine the entire product development process. The focus was on three innovations in particular: generative AI-supported, editable BREP geometry in Fusion, an agent-based Autodesk Assistant and extended collaboration with Microsoft for contextual visualization and seamless export to Microsoft 365. Together, these innovations reduce routine effort, accelerate iterations and connect design and creativity directly with production-ready results.

Forward-looking solutions in challenging times

Autodesk sees manufacturing industries at the beginning of a paradigm shift, driven by AI. Such groundbreaking moments are not new to companies, but artificial intelligence and the ways in which products are designed, engineered and manufactured are changing faster than expected. According to Autodesk, this paradigm shift presents an opportunity to free up capacity, increase creativity and improve productivity. What's more, with a critical shortage of skilled labor, AI is becoming essential as it enables more to be achieved with fewer resources. Autodesk sees itself in a position to lead the upcoming changes and is presenting a number of developments as part of this mission.

Generative AI-powered editable geometry in Fusion

Autodesk is implementing a new form of generative modeling in Fusion: editable boundary representation (BREP) geometry, driven by the company's own base models for manufacturing. A single command is designed to create highly detailed geometries that are output as a fully editable CAD model. The path from the idea to the design-ready file is considerably shortened, as designers can continue working immediately, adapt parametrically and take manufacturing aspects into account without cumbersome conversions or manual rework.

Autodesk Assistant: An agent-based AI partner for productivity

With the new Autodesk Assistant, Autodesk introduces a context-sensitive, agent-based AI partner that is seamlessly integrated into products such as Fusion and Vault. In Fusion, the Assistant automates organizational and technical steps: inviting team members, setting sketch constraints or generating production paths. Time-consuming tasks are outsourced so that users can concentrate on developing innovations. In Vault, the Assistant enhances working with data, including natural language search, automatic follow-up actions such as updating outdated drawings and transferring bills of materials (BOM) into PLM workflows (Fusion Manage). The result: a consistent, reliable data flow across the entire product life cycle.

Contextualized visuals and direct export to Microsoft 365

Autodesk and Microsoft are jointly driving forward the visual preparation and distribution of design results. By using Microsoft's Azure OpenAI Services and models such as "GPT image 1", Fusion designs can be directly contextualized and visualized photorealistically. This allows users to automatically place a newly developed device in a realistic environment, for example. These visuals and metadata can then be automatically exported to Microsoft 365 and transferred to PowerPoint presentations so that marketing and sales teams receive high-quality materials without manual effort.

Production-ready MCP server for seamless automation

Autodesk is empowering the entire design & manufacturing industry with production-ready implementations of Model Context Protocol (MCP) servers. This server no longer just delivers isolated CAD files, but complete, verified working views with all relevant details such as release status, dimensions, material information and manufacturing constraints. AI agents and automation tools should be able to clearly interpret and process these working snapshots. For product development and manufacturing, this means that routine tests for manufacturability can be automated, CAM processes can be triggered directly from the validated model context and NC programs can be generated much more accurately. Simulations can also be started automatically with the correct input data and results can be fed back into the development process. Changes to parts lists can also be tracked via the mapped supply chains.

Video