AI-powered fit-out and construction workflow platform

Before this platform, a project moved through drawing review, quantity takeoff, material estimation, supplier outreach and quotation comparison as separate, manual steps. Each handoff meant re-entering the same information into a different document, spreadsheet or inbox, and a mistake made early was invisible until much later in the tender.

A platform that keeps one project record moving through takeoff, procurement and quotation, instead of five separate ones.

Suppliers work inside the same system as the internal team, so a request for quotation, a clarification and a final price all live on one thread instead of scattered emails.

Project documentation moves between English, Chinese and Traditional Chinese without losing structure, which matters for a business operating day to day in Hong Kong.

The system reads a construction drawing and produces a structured takeoff for a human to check, not a number to trust blindly.

The quantity surveyor keeps the final call on anything the system flags as uncertain. The AI’s job is to make the first pass fast, not to make the last decision.

Takeoff accuracy is the platform’s central commercial claim, so it is checked against manually produced takeoffs rather than trusted on its own. The AI counts and measures against the drawing set; the estimator reviews anything flagged, checks complex assemblies the system is less confident on, and signs off before a number goes into a bid.

Independent testing of AI takeoff tools reports 95% to 99% accuracy on clean vector drawings, with 8% to 12% error rates on dense, heavily annotated commercial sets. This engagement’s own agreement rate is a separate figure, still to be measured against a live project set.

Drawings, quantities and supplier pricing are commercially sensitive. Access is scoped by role and by project, so a coordinator on one job can’t see pricing on another, and a supplier only ever sees its own quote, never a competitor’s. What reaches an AI model is limited to the specific document being analyzed in that moment, not a project’s stored history.

Construction AI earns its place when it’s embedded in how a project actually runs, not bolted on as a chatbot beside it. A system that turns a drawing into a checked number, and that number into a quotation, is worth more than one that just answers questions about the drawing.