Case Study
Skanska: Turning the Jobsite Photo Archive into a Safety Signal
Machine learning made existing BIM 360 photos and narrated walks searchable for risk and progress evidence.
Before buying more data, ask whether AI can make the evidence your project already collects easier to find, compare, and act on.
Safety and progress were documented visually, but finding the right photo, spotting repeated risk factors, and organizing narrated walks demanded manual effort.
Construction teams take photos because the site changes every hour. The trouble begins later, when somebody asks for the image that proves a condition existed on Tuesday morning. Skanska's case used machine learning on BIM 360 photos and narrated walks so managers could search evidence and observe safety-risk patterns instead of hunting through folders like digital archaeologists.
Smartvid.io applied machine learning to BIM 360 imagery and narrated site walks to detect risk factors, index visual evidence, and generate searchable progress records.
Autodesk's published description does not provide a financial saving, so this case does not invent one. Its value is operational: index evidence consistently, benchmark signals over time, and make progress records accessible to internal and external collaborators. Good AI sometimes solves the unglamorous problem of finding the right fact before the meeting ends.