Professional learning for member companies
Short videos about tools, methods and experience. Descriptions are public; the videos themselves require access through company membership.
Watch the videos before the discussion meeting.
BIM · Sprint 1
GPT Agent Classifies Revit Models
An agent developed in Codex to classify rooms in Revit models according to the space standards of Norway’s county councils’ property forum (FEF) and Norwegian Standard NS 3457-4. Codex uses Revit Experimental MCP to edit parameters and modify views, and computer control to operate the Revit user interface for actions that MCP does not yet support. The agent finishes by exporting annotated floor plans in PDF format for each classification.
View videoAI · Sprint 1
Trust the answer, not the model, not the vendor
For companies considering running AI on their own server. You learn why local AI gives control over data, cost and access, and how many agents checking each other give answers you can trust, even when a single model guesses wrong. The video shows that a server scales in two directions: how large a model it holds, and how many conversations it runs at once. It ends with a checklist of what a company should settle before it invests.
View videoProcurement · Sprint 1
Does transferring risk to the supplier eliminate it, and is it always the customer who chooses the supplier – rather than the other way around?
magine you are managing a major construction project that requires a supplier to deliver and install a complex technical system. With tight deadlines and significant risks, you negotiate a contract with strict requirements, fixed deadlines, and clear responsibilities, transferring as much risk as possible to the supplier. Does this mean your project is well protected? Not necessarily.
View videoAI and digitalisation · Sprint 1
AI gives us superpowers – what should you ask of a consultant?
As AI gives consultants and in-house specialists new ways to develop solutions, it should also influence what clients ask for. How can a specific assignment leave the client with a tool they can build on, even after the consultant has left the project? The video uses Runa as an example: a solution developed with AI to extract information from Revit models in Autodesk Forma and make the data available through an API for use in Power BI. The walkthrough covers room data extraction, data views and API access. A specific need becomes the starting point for a tool the client can extend and use for further tasks. Runa is the engineer that can keep delivering value after the initial need has been met.
View videoDigitalization and Technology · Sprint 1
Autodesk Forma – From Site Analysis to Building Design
How can digital tools help us explore more design alternatives, make better decisions, and develop building projects more efficiently? In this practical demonstration, we explore Autodesk Forma and how the platform can support workflows from early-stage site analysis to building design and visualization. The video demonstrates how to explore different alternatives, develop building volumes, and communicate design solutions through digital models. Relevant for project managers, architects, developers, client representatives, consultants, and anyone interested in practical applications of digital technology in the construction industry.
View videoLLM, ChatGPT · Sprint 2
Under the Hood of My ChatGPT Setup
Sindre Gundersen walks through his ChatGPT subscriptions, explains why he chose them, and shares a few tips and tricks from everyday use.
View videoAI and digitalisation · Sprint 2
From building documents to insight: local AI and control of your data
How can a building owner extract knowledge from operations and maintenance (O&M) documents and product information while retaining control over where the data is processed? This video explains the principles of OCR with locally hosted AI models: first identifying the page layout, then extracting text and tables in a reusable form. It also compares running the software on your own hardware, a rented dedicated GPU server at Hetzner and capacity within the customer's Azure environment. Topics include data control, operational responsibility, sharing model capacity and the cost of idle resources. An experience with unavailable GPU capacity illustrates why access to computing resources matters when choosing a hosting environment. The presentation concludes with hybrid hosting and a hosting example from Aarhus. An illustrated introduction for building owners, facilities managers and digitalisation leads considering a pilot.
View videoAI · Sprint 2
Too much of a good thing: how context can affect answers from AI
For anyone who uses AI on models, drawings and regulations. You learn what the context is, why everything is sent again at every step of a conversation or an agent, and why more context can give worse answers. In a test with an IFC model and the Norwegian building code TEK17, the AI model approved a fire compartment that is missing a wall, in all ten runs, when it was given the whole building and the rules. Given only the room, it found the gap every time. The video ends with how to tidy the context, both by hand and with agents.
View videoAI and practical problem-solving · Sprint 2
When things get stuck: I built an app with AI
What do you do when your tools do not meet your needs? I wanted to mirror my Windows screen to an Apple TV, but ran into technical limitations, costs and an internal approval process for new apps. In this video, I show how I used AI to develop a local prototype, what I contributed and what AI did. For people working in construction who want to explore better solutions to everyday frustrations in their projects.
View video