No. 30
One database across the building life cycle
- Context
- Point of view
- Year
- 2026
- Type
- Approach
- Tools
- Layer
Every phase of a building's life re-creates the same information in a different tool. I think a project should be one structured database that every role reads in its own way, and that this is what will make AI useful to a firm.
Today an assessment lives in a spreadsheet and a folder of photographs, the design in Revit and a drawing set, construction in submittal and field apps, and operations in a work-order system. Each handover retypes what the last phase already knew, and each role keeps its own copy.
If every room, element and document is a record with an id that never changes, that stops. The owner, the architect, the engineer, the contractor and the facility manager read the same records in their own views, and the reports each of them issues are printed from the data instead of assembled by hand.
This matters more with AI, not less. An AI agent can only answer a question across projects if the data is structured and the ids hold. With that in place a firm could ask which rooms depart from its standard, or pull every casework specification it has used, and an owner, architect and contractor could exchange by id instead of retyping.
Two pieces of my own work test parts of this. Room Data Sheets carries one database from room standards to issued sheets. From model and site data to a place you can walk carries one id from a Revit element to a playable model and its record.
I have not run a building this way from start to finish. I believe in it, I build it in Layer, and I am curious to try it on a live project.
01
Every role, every phase, one set of records
01