Dive Brief:

• Following up on a study in which Smartvid.io’s artificial intelligence engine “Vinnie” learned from contractor Suffolk's data to predict roughly one in five safety incidents with 81% accuracy, nine major construction firms are partnering with the software company to further develop predictive analytics and therefore prevent incidents and reduce risk across the industry, according to a Smartvid announcement from CEO and Founder Josh Kanner.
• Suffolk will chair the Predictive Analytics Strategic Council, which also includes Barton Malow, Clayco, DPR Construction, JE Dunn, Messer Construction Co., Mortenson, Shawmut Design and Construction and the Bouygues Group, as well as insurance broker Aon. The council is still accepting member applications.
• During the first, 12-month phase, Suffolk fed 10 years of photo and project data from internal systems like Autodesk, Procore, Oracle and Oxblue into a machine learning predictive model. Partners in this phase will continue to expand this model by contributing data anonymously, Engineering News-Record reported, and will also discuss the operational, social and business impacts throughout the process, Kanner said in the post.

Dive Insight:

Vinnie is likely to sharpen up as he quickly “learns” from millions of images to identify objects and interpret their context in order to flag risks to workers. The AI system can alert teams to standing water, for example, helping to prevent them from slipping or coming across an electrical hazard.

According to the Suffolk and Smartvid case study, Vinnie’s warning system predicted 20% of all incidents over a three-year period with 80% accuracy, which comes out to four alerts per year with one of those being a false alarm. Managers willing to receive more alerts, though, could get predictions for 40% of all incidents over that period with 66% accuracy, the companies said, with one in three being a false alarm.

Taking the conservative assumption that only 25% of predicted accidents are avoided, Smartvid says that a company with 50 projects each year can avoid 40 to 100 incidents annually, saving between $1.4 million and $3.6 million. In 2018, according to the company, safety-related costs for contractors averaged $36,000 per incident.

Given these promising results, Smartvid continues to attract interest from both contractors and onlookers who see this as a way to advance the AEC field as a whole.

“This work is not about one company’s risk,” said Jit Kee Chin, chief data officer at Suffolk, “it’s about reducing risk across our industry.”

Smartvid closed a funding round last month led by Suffolk, Building Ventures and Leawood Venture Capital, which includes a number of owners and executives from top engineering, design and construction firms in the Midwest.

The funds, along with the council’s contributions, “will accelerate development for the predictive analytics module and more,” said Kanner in the post.