Los Angeles is housing more people than ever, and building lots more low-income housing, yet it can't keep pace with this ever-rising number of people who end up in cars, tents and shelters. "It's a bucket with a hole in it, so we've got to do something ... to fill that hole," says Dana Vanderford, who helps lead the department's Homelessness Prevention unit. With that goal, the pilot program is using artificial intelligence to predict who's most likely to land on the streets, so the county can step in to offer help before that happens. From a report: The program tracks data from seven county agencies, including emergency room visits, crisis care for mental health, substance abuse disorder diagnosis, arrests and sign-ups for public benefits like food aid. Then, using machine learning, it comes up with a list of people considered most at-risk for losing their homes. Vanderford says these people aren't part of any other prevention programs. "We have clients who have understandable mistrust of systems," she says. They've "experienced generational trauma. Our clients are extremely unlikely to reach out for help." Instead, 16 case managers divide up the lists and reach out to the people on them, sending letters and cold calling.