Can AI Safely Write Changes Back to Revit?
A practical framework for previewing, approving, applying, and auditing AI-assisted Revit changes without hiding model authority.
Applied AI, BIM, modular construction, and manufacturing systems.
Field Notes translates technical developments into operating consequences, control boundaries, product evidence, and practical decisions for teams working in the built environment.
Editorial standard
Each new article uses named authorship, primary-source links, original Matechi evidence or diagrams, explicit limitations, publication and update dates, and a next step relevant to the reader's workflow.
All Field Notes
Clash Detection Is Not Clash Resolution
Why a clash list is only the beginning, and what an evidence-backed resolution loop should include.
From Product Cut Sheet to Parametric Revit Family
The evidence and validation steps between product information and a Revit family that designers can actually use.
What Configure-to-Order Means for Modular Construction
How approved product rules turn modular design options into a validated commercial and manufacturing output.
Can a Clash Issue Survive the Next IFC Export?
A pass/fail IFC and BCF pilot for proving that model versions, element identities, rule intent, and review decisions survive across a coordination loop.
Is Your AEC AI Workflow Ready for Production?
A 12-control readiness checklist for moving an AI-assisted AEC workflow from an attractive demonstration into a governed production pilot.
Who Approves What in AEC Automation?
A practical human-review control matrix for deciding which AEC automation outputs may inform, draft, change, coordinate, or release work.
Governed AI in AECM: 2026 Evidence and Operating Benchmarks
Fifteen primary-source indicators for AI adoption, digital-delivery friction, rework, BIM, safety, workforce, and governance—plus a defensible way to set internal operating benchmarks.
AI-Driven 3D Segmentation Is Changing the Way We Model
Turning 2D images or fused 3D meshes into editable geometry used to be painful. If you've ever converted a sketch or render into 3D, you know the frustration of working with a single, uneditable mesh. You could orbit around it, but you couldn't really refine it.
AI & The Next Frontier for Modular: Teaching Our Systems to Speak the Same Language
Modular construction has made huge progress in manufacturing, logistics, and digital coordination, yet one problem continues to slow growth: our systems still do not understand each other.
From Text to 3D: The Next Leap in Design Automation
If you haven't seen it yet, Zubair Trabzada's recent walkthrough on using NanoBanana and n8n for 3D model automation is worth a watch. With a few lines of text or a simple screenshot, you can now generate fully rendered 3D concepts and even push them into modeling APIs like Tripo3D for complete geometry creation.
The Living Building: How AI, Omniverse, and Tandem Are Redefining Digital Twins
We're entering a new era for the built world, one where AI doesn't just interpret buildings, it helps them think.
Testing MCP with Revit: Where AI Reasoning Meets BIM Data
I came across a great post by Karim Tayel experimenting with connecting Claude to Revit using the new MCP (Model Context Protocol). This prompted us to test a similar setup with our Standards Agent for BIM compliance auditing.
The 5 Most Groundbreaking AI Announcements from AU 2025 - and What Comes Next
Autodesk made it clear at AU 2025: AI is no longer an experiment. It is core to how projects will be designed, built, and operated.
AI for Submittals: How Contractors Can Cut Review Time by 40%
Project managers spend considerable time cross-checking PDFs rather than focusing on strategic work. On large projects, thousands of pages can require hundreds of staff hours, creating ripple-effect delays throughout the schedule.