01What is a 'surrogate model' in early-stage performance tools like Forma or cove.tool?
02An early ML energy prediction is most appropriate for which use?
03In computer-vision site-progress tracking, what does the system compare the 360 camera capture against?
04A site-progress tool reads a room as 100% complete. What can you correctly conclude?
05In a scan-to-BIM workflow, what is the point cloud and how accurate is the AI step that turns it into a model?
06Where does AI scan-to-BIM extraction most commonly go wrong?
07Why must every quantity from an AI takeoff be verified before it goes into a BOQ?
08What is the right division of labour for AI clash and defect detection?
