Using AI-Derived Insights to Improve Safety Behavior on the Jobsite

Across all industries, making better-informed decisions is key to companies realizing enhanced business outcomes. This is especially important in the construction industry, where improving safety, especially during the COVID-19 pandemic, is crucial to success.

Improving safety not only makes sense from a human perspective, as we all want workers to go home unharmed at the end of the day, but it also makes financial sense, as safer projects tend to be more profitable. Safety also plays into winning new business, as owners and developers are likely to give preference to contractors with proven, stellar safety track records.

Making better-informed decisions around safety at its foundation requires data collection on current jobsite behaviors as well as the ability to translate this raw data into insights designed to help decision makers to identify opportunities for safety improvement.

Warfel Construction turned to Boston-based software company SMARTVID.IO and their artificial intelligence (AI) engine, “Vinnie,” to accomplish this.

Why Deploy AI Now?

Warfel’s story began at 5:20 p.m. on a Friday afternoon. The president of Warfel had been reviewing jobsite progress photos and had some questions around safety. To get answers, he placed a call to his operations team. This kicked off a Friday evening scramble to formulate a response, which included sending someone to the jobsite to look things over.

In the aftermath, Warfel’s operations team thought there had to be a better way. How could they utilize technology to not only identify existing safety risks at jobsites, but also aggregate this information into analytics that could be used to understand trends?

From Photos to Analytics

Insights start with data, which in Warfel’s case is largely in the form of photos taken from theirvarious jobsites. Thephotos are sourced frommultiple locations, such as mobile phones and 360-degree photos taken from programs such as StructionSite.

As an AI engine, Vinnie has been “trained” to consume photo data to discover and count several indicators of project risk in the areas of safety, productivity and quality. For instance, Vinnie can recognize if someone is standing on a ladder or if required PPE isn’t being worn.

Photos are organized in Warfel’s Procore database. SMARTVID.IO’s AI engine integrates directly with Procore to consume the image data. This integration was especially important to Warfel, because it meant they didn’t have to change anything in their existing process to utilize photo data from their jobsites.

Safety Analytics

Safety was the first important category of analytics produced by Vinnie. The Warfel team wasn’t really interested in policing their workers and subcontractors, but understands that safety risk is everywhere. As such, they were looking for safety analytics to help mitigate possible risk as they managed a process of continuous safety improvement.

For example, the Warfel team now utilizes the output from Vinnie to look specifically at the safety measures and scoring for their subcontractors to understand what causes certain subcontractors to be more of a safety risk on projects than others.

Warfel also examines the subcontractors that continually receive high safety scores to try to understand what can be learned from their behavior to help educate the poorer-performing subcontractors. In essence, safety analytics produced by Vinnie help uncover lessons that can be utilized to elevate the lower- performing trades.

As an example, one of the trends Vinnie uncovered was that safety ratings improved eight times over when a foreman was present onsite. Vinnie recognizes foremen by the color of their hard hats. This insight led Warfel to advise its lower-performing subcontractors that if they work together to make sure a foreman was present, they would likely become safer subcontractors.

COVID-19 Analytics

Warfel made the decision to deploy Vinnie in their workflow before the COVID-19 pandemic hit—and they’re thankful they did, as Vinnie is also able to recognize when social distancing and PPE guidelines are not being followed. As an example, Warfel performs renovation construction projects at senior care facilities. Vinnie helps the Warfel team monitor whether workers are following the proper COVID-19 safety protocols, which is especially important for this vulnerable senior population.

By leveraging Vinnie’s analytics, Warfel is able to give its customers confidence that workers are following COVID-19protocols. In addition, access to these analytics helped Warfel to stay working, as they could provide concrete evidence to the various stakeholdersthat their workers were wearing PPE and maintaining social distancing.

As an example, Vinnie has a “people in a group” tag that recognizes if workers are likely fewer than six feet apart. This is especially helpful, as construction workers are trained to work together. In the post-pandemic world, workers are having to be re-educated on how to perform their jobs while trying to maintain social distance.

Vinnie can also recognize if a mask is not being worn. Again, for Warfel, access to these analytics is not about “nailing” employees for poor safety behavior. Rather, it’s about identifying opportunities for behavior to be improved, which improves safety for everyone. In the case of not wearing masks, there were situations when it was recognized that workers were simply unable to maintain a six-foot distance, so face shields were provided.

Winning Work

As an added bonus of utilizing Vinnie’s AI analytics, Warfel is able to parlay verifiable safety record to win more business. After all, it’s very appealing for developers and owners to work with a contractor that has access to proven safety analytics to help continue to mitigate safety risks.

For many, artificial intelligence seems the fodder for movie scripts and future-speak. But for Warfel—an innovative leader in the construction industry—utilizing AI to improve safety is now just part of doing business.