Key Takeaways

• Autodesk unveiled agentic AI for its three industry clouds—Forma, Fusion, and Flow—at Autodesk University 2026.
• The new AI layer, dubbed Autodesk Assistant 2.0, works across projects, products, and teams, leveraging “project intelligence” drawn from connected data.
• A stand‑alone AI hub is slated for release in 2027, allowing users to start from a business goal rather than a specific tool.
• Autodesk’s edge lies in grounding large‑language‑model reasoning with real‑world geometry, physics, constraints, and historic project metadata.

Autodesk Introduces Agentic AI for End‑to‑End Project Workflows

A unified AI layer across three industry clouds

During Autodesk University 2026, the company previewed a next‑generation AI that spans its three flagship industry clouds:

The table illustrates how the upcoming assistant moves from a tool‑centric helper to a project‑centric collaborator.

From “What tool?” to “What outcome?”

Today’s Autodesk Assistant already assists designers inside Fusion, architects inside Forma, and editors inside Flow. The upcoming version flips the paradigm: users state the business intent—e.g., “reduce wall assembly weight by 15 % while meeting fire‑rating codes”—and the AI surfaces the right geometry, simulation settings, and manufacturing instructions across all relevant clouds.

Key components of this shift include:

• Project‑level data aggregation – over 200 TB of BIM, CAD, and media assets from typical enterprise deployments are indexed in real time.
• Contextual reasoning – the AI interprets constraints such as load‑bearing capacity, material availability, or rendering pipeline limits.
• Unified UI – a web‑based dashboard (planned for 2027) that replaces the need to open separate desktop apps for each discipline.

Why “project intelligence” matters

Design and manufacturing AI must grapple with more than generic language. Real‑world projects involve:

Autodesk’s three industry clouds each curate this domain knowledge:

• Forma links architectural models, civil‑engineering schedules, and construction‑site logistics.
• Fusion ties CAD geometry to CNC tool‑paths, simulation results, and supply‑chain data.
• Flow connects storyboards, asset libraries, and rendering farms.

Together with four decades of Autodesk’s sector expertise, the AI can turn a high‑level intent into a concrete, manufacturable design that respects all underlying physics and regulatory constraints.

Real‑world impact: a quick scenario

A structural engineer asks the assistant to “lighten the steel truss without sacrificing stiffness.” The AI:

1. Retrieves the truss geometry from the Forma model.
2. Runs a finite‑element analysis using Fusion’s simulation engine.
3. Suggests alternative cross‑section profiles that cut weight by 12 % while keeping the first‑mode frequency above 15 Hz.
4. Updates the project schedule to reflect the new fabrication lead time.

The result is a single‑click workflow that would previously require a team of analysts and multiple software packages.

Bottom Line

Autodesk’s agentic AI preview at AU 2026 signals a decisive move from isolated, tool‑specific assistants to a holistic, project‑aware collaborator. By marrying large‑language‑model flexibility with deep, industry‑specific data—geometry, physics, schedules, and history—Autodesk aims to let engineers, architects, and creators start from a business goal and let the platform surface the right tools, data, and recommendations. The planned 2027 standalone hub promises to make this vision accessible across enterprises, positioning Autodesk as a leader in AI‑augmented, end‑to‑end product development.