Autodesk University 2026
A Seat at the Table: How AI Brings Facility Managers Into the Revit Design Process
Sharon Dar (Data Scientist, Compliance Automation Lead, SWAPP)
Submitted for Autodesk University 2026 — not yet scheduled.
Facility managers are the end users of every building, yet almost never part of the design conversation. The operational requirements exist in Owner’s Project Requirements, institutional facility guidelines, and equipment manufacturer specs, but they reach the design team too late, if at all. This session shows how an AI agent ingests those facility-requirement documents and validates them against the live Revit model during design, so the issues that normally surface six months post-occupancy get caught while the architect can still fix them. The model then arrives at handover with structured compliance metadata that operations tools like Autodesk Tandem can act on directly, rather than reverse-engineering the designer’s intent.
What you’ll be able to do
- Identify which of your firm’s recurring post-handover facility-management complaints could have been caught during design, and map them to checkable requirements.
- Locate and structure Owner’s Project Requirements, institutional facility guidelines, and equipment specs as inputs for AI-driven validation.
- Apply a workflow where facility-requirement documents are ingested by an AI agent and validated against a live Revit model during design.
- Embed facility-compliance metadata into the design model so downstream facility-management teams receive structured, actionable data at handover — not a blank model to reverse-engineer.
Session outline
-
Opening · 5 min
The Requirements That Never Arrived
Every firm has the story: a building handed over, and within months the facility team is logging complaints about equipment you can’t reach and clearances that meet code but not reality. The operational requirements existed — in a document somewhere. They never reached the designer. In AECO, Architecture, Engineering, and Construction are in the design conversation from day one; Operations almost never is.
-
Section 1 · 8 min
Where Facility Requirements Live, and Why They Don’t Reach Designers
Three document types architects rarely see during design, all detailed and checkable: Owner’s Project Requirements (maintenance access, serviceability standards), Institutional Facility Guidelines (ceiling heights, MEP access routes, adjacency standards), and Equipment Manufacturer Specs (serviceability clearances, drainage requirements). The information is there and the standards are specific. The pipeline is what’s missing.
-
Section 2 · 12 min
Facility Requirements Validated During Design (Live Demo)
On a live Revit model: (1) the AI agent ingests facility-requirement documents and surfaces issues against the current model, during design; (2) what gets caught — clearance violations, access-path conflicts, ceiling-height non-compliance; (3) designer interaction: the AI surfaces the requirement, the architect resolves it, and the decision is documented either way.
-
Section 3 · 3 min
From Design to Handover: Facility-Ready Models
When facility requirements are validated during design, the model arrives at handover with structured compliance metadata already embedded, so Autodesk Tandem and downstream operations tools can operationalize it immediately, rather than reverse-engineering the designer’s intent.
-
Closing · 2 min
Your First Step
Identify one building type your firm designs repeatedly, and the facility-operations requirements document that governs it. That document is the input — everything else follows. Q&A.
Speaker
Sharon Dar
Data Scientist, Compliance Automation Lead, SWAPP AI
Sharon Dar is a Data Scientist at SWAPP AI, where he leads compliance automation, building AI systems that ingest regulatory and facility-management documents, automatically extract validation rules, and check them against live BIM models during design. His background spans nearly a decade of applied data science in healthcare and digital health, applying LLMs, RAG, and NLP to compliance-heavy domains. He holds a BSc in Industrial Engineering and an MSc in Statistics, both with a machine-learning focus — bringing the same AI-driven compliance rigor that transformed healthcare into the built environment.