Applying AI to planning requires a trove of data, and many power utilities have the operational data necessary to begin the kind of advanced analytics machine learning can provide. National Grid recently announced that it will pilot the machine-learning tools from technology firm Urbint to anticipate safety issues on its construction projects, bringing an AI advisor in to spot problems on jobsites before they lead to safety incidents.

“National Grid was one of our early adopters on our damage-reduction technology, so we’re expanding our solution to worker safety,” explains Lindsay Jenkins, Urbint vice president for strategy and technical operations. Urbint Lens for Worker Safety examines a utility’s project data, making risk assessments based on site conditions and the type of work involved, as well as the relevant history of safety incidents and near-miss events.

“The inputs of our model are connected to a company’s project schedule, so we know the work being planned and can identify the site conditions,” says Jenkins.

Urbint’s model generates usable reports for different levels of users. Construction superintendents get an updated risk score for their project and can use these metrics to plan ahead. “It can drive their on-site decisions, like maybe today is a stop-work day, since these tasks in certain weather are prone to safety incidents. Maybe there is a trend in near misses and a safety briefing is needed,” says Jenkins. Urbint provides a different dashboard at the management level, offering a view across the utilities’ entire risk portfolio to identify safety trends.

“This technology is very exciting. It has the potential to pull all this information into one place so we can do some analytics,” says Walter Fromm, vice president of capital delivery for National Grid Gas. “That way, a worker or superintendent out in the field can see before they start what is happening, and take action.”

“We bring our own biases to the jobsite,” says Jenkins. “This is a form of support to take that veil off and use proven methodologies to identify hazards.”

National Grid’s pilot will cover its gas-related construction across the Northeast. Fromm says that if it’s successful, he’d like to roll Urbint out across the utility’s entire portfolio. “I don’t want these safety alerts to be optional,” he says. “Our workers are out there—a lot is going on. My aspiration is this technology will be a control, so they will have those safety conversations in the morning before going to work.”