How Suffolk Construction Used AI to Cut Safety Incidents on Site

The Challenge

Traditional site safety management is reactive by design. Inspections happen on a schedule, not continuously, so the gap between a hazard appearing and someone noticing it can stretch for days. Incidents get logged after the fact, but the underlying patterns - which crews, which conditions, which times of year carry the most risk - are buried across thousands of site photos and reports that no one has time to analyze by hand.

The Solution

Suffolk Construction built RiskX, a cloud-based mobile app that lets crews capture safety observations in real time from the field. Layered on top, an AI image-analysis model (developed with Smartvid.io, now Newmetrix) scans site photos automatically to flag hazards such as missing PPE, then surfaces the patterns behind past incidents so site teams can act before something happens rather than just documenting it afterward.

• Real-time field observation capture via a mobile safety app
• Automated image analysis to detect PPE non-compliance and other visual hazards
• Predictive analytics identifying which conditions and crews correlate with higher incident risk
• Proactive monitoring of at-risk individuals and projects, rather than after-the-fact reporting

The Results

• 28% reduction in recordable incident rate
• 35% reduction in lost-time incident rate
• Faster identification of high-risk crews, conditions, and time-of-year patterns
• A shift from reactive inspection to proactive, data-driven safety management