From the desk…
Multiple building models are open on my screen: architecture, structure, and services. I can section the model, select elements, and inspect their properties.
What interests me is how quickly this idea became something I could test.
This IFC viewer is an experiment in AI agent-assisted product discovery: how far can I take a focused idea before committing to a larger development effort?
In previous product work, reaching this fidelity could require budget, team capacity, and weeks of development. The agent-enabled prototype shortens that path.
This working prototype still leaves important questions:
– Which user need does it address?
– Which assumptions can it test?
– What evidence would justify taking it further?
My construction experience helps me judge the workflow. My product and program experience helps me decide what to test next.
Faster building makes those choices more critical. I’ve learned that a team can spend a lot of money learning a lesson. The opportunity is to learn it earlier.