More than a decade on, the story of one Michigan city’s struggles is widely known.

In 2014, Flint switched its drinking water supply from the Detroit system to the Flint River to save money. The water wasn’t properly treated, leading to a health crisis in which residents faced lead contamination and other serious health issues.

In the wake of all that, one of the many challenges facing Flint—and hundreds of cities looking to update their aging water infrastructure—is efficiently and accurately predicting the location of lead pipes.

Enter BlueConduit, a startup of the University of Michigan’s Ross School of Business that originated the approach of using machine learning to do just that. Since the company’s foundation in 2016, they’ve rapidly expanded, working with more than 100 customer accounts spanning over 400 different towns and cities in more than 25 states.

On the latest episode of Business and Society, Eric Schwartz, U-M associate professor of marketing and co-founder of BlueConduit, recounts the company’s journey from community startup to expanding and impacting public policy. He also offers insights into the businesses of public utility and infrastructure and shares his optimism in data science and marketing being used as forces for good.

From machine learning startup to shifting public policy

Utilizing pattern recognition and machine learning, BlueConduit started using historical records to accurately predict the locations of lead pipes. As these predicted locations became dig sites, the data was validated and verified in real time.

“At first, this was really just an engaged research project—engaged with a community, engaged with the local government,” Schwartz said. “We started realizing we can use this approach elsewhere, too. And that’s when we started getting inbound requests. ‘Hey, can you do that in our community?'”

In 2020, when the Environmental Protection Agency proposed a new lead and copper rule for the first time since 1991, it referenced the research of Schwartz and his co-founder, Jake Abernathy, and continued to reference BlueConduit in the guidance the agency gave for executing this work in other cities.

Data science as a driving force for public good

Schwartz has always been interested in using data science as a force for good, and speaks to how meaningful it’s been to invite a wide variety of students, Ross alumni and other hires seeking mission-oriented data jobs into the company.

“That’s been our goal from the beginning,” he said. “My co-founder and I have wanted to create not just impact, primarily in communities where these issues are happening, but also to create these opportunities for folks that want to do data for good.”

Looking ahead, Schwartz focuses on continuing to solve optimization problems for public utilities and infrastructure fields, working to improve predictions and prioritize limited resources.

Schwartz and the University of Michigan have financial interests in BlueConduit.

Business & Society is co-produced by JT Godfrey of the Ross School of Business and Jeff Karoub of the Michigan News office. The audio engineer is Jonah Brockman and editorial production is provided by Mads Henke. Listen to all episodes of the podcast.