Skip to content
All sessions

Autodesk University 2026

Revit AI: Down the Rabbit Hole

Adi Shavit (CTO & Co-Founder, SWAPP)

Technical Deep Dive30 minIn-person + DigitalProfessionalAIA LU

Submitted for Autodesk University 2026 — not yet scheduled.

AI can save time on repetitive Revit tasks. But can it enforce your firm’s standards while it works? Can it share what one team learns with every other team? Can it scale across offices without retraining? The answer depends on the depth of AI integration. This session maps increasing levels of AI automation in Revit, from scripting to firm-aware embedded agents, and shows what each level unlocks for your practice. Attendees leave with a framework to assess where their firm stands today and what the next level of value looks like.

What you’ll be able to do

  • Map the capability tiers of AI automation in Revit, from task-level scripting to firm-wide embedded agents.
  • Identify the specific value each tier unlocks: execution safety, firm-standards enforcement, cross-team scalability.
  • Assess where your firm’s current AI usage sits using three diagnostic questions.
  • Determine what the next capability tier requires and what business value it delivers.

Session outline

  1. Opening · 5 min

    The Depth Question Nobody Asks

    Every AI demo shows the easy part: type a prompt, something happens in Revit. The gap between “AI does a task” and “AI runs your documentation workflow” is not incremental — it is architectural. This session maps exactly where that gap is.

  2. Section 1 · 8 min

    Six Levels of AI Depth in Revit

    A capability ladder from shallowest to deepest: (1) access & basic automation, (2) conversational assistance and action, (3) domain-grounded controlled execution, (4) firm-aware shared intelligence, (5) enterprise workflows and governance, (6) learned production intelligence. For each: what it requires, what it delivers, and what breaks when you skip levels.

  3. Section 2 · 12 min

    The Two Transitions That Matter

    Level 2 → 3, where safety enters: a generic LLM generates and executes code that works on this project but applies the wrong choices on another; a domain-grounded system validates before acting and rolls back on failure. Level 3 → 4, where firm-level value begins: the system applies your firm’s conventions without being told, configured once by an admin and inherited by every team.

  4. Section 3 · 3 min

    Where Does Your Firm Sit Today?

    Three diagnostic questions that place your firm on the capability ladder. Most firms sit between Levels 1 and 2, experimenting individually, without standards integration or cross-team scalability.

  5. Closing · 2 min

    Your Next Level

    A one-page capability ladder: where you are, what the next level requires, and what it delivers, plus a firm assessment scorecard to take back to your team. Q&A.

Speaker

Adi Shavit

CTO & Co-Founder, SWAPP AI

Adi Shavit is a serial entrepreneur and CTO & Co-Founder of SWAPP AI, where he leads the technical architecture behind AI-driven BIM automation deployed across firms in North America, Europe, and Australia. He has built real-world AI and machine-learning systems for over 25 years, previously headed the Autodesk Mobile Innovation team, and co-founded Homestyler (acquired by Autodesk in 2012). He holds an MSc in Computer Science and has multiple patents and publications. He has spoken at NXTBLD, Trimble Dimensions, and other industry events.

Want this session at your firm?

We bring the same material to private firm sessions and webinars. Tell us what your team is working through.