Skip to content
HOMETECH DIRECTORYNEWS & DEVELOPMENTSRELOCATEEMPLOYERSNEIGHBORHOODSABOUTSchedule a Call

BlueConduit: The Ann Arbor AI Company Helping Cities Find Lead Service Lines

BlueConduit is the U-M-spinout AI company that helps cities locate lead service lines. Born out of the response to the Flint water crisis. Customers acr...

A2Hub Editorial· Rusty Roof Media· Updated May 4, 2026

BlueConduit

By A2Hub Editorial. Published 2026-04-22. Last updated: 2026-04-22.

At a glance

  • Founded: 2019
  • Founders: Eric Schwartz, Jacob Abernethy, and collaborators from the University of Michigan data science team that worked on the Flint water crisis response
  • HQ: Ann Arbor, Michigan
  • Customers: Cities across more than 50 US municipalities, including Flint, Detroit, Trenton, and Toledo
  • Funding and partnerships: Foundation grants and federal partnerships, including support from Google.org and the Rockefeller Foundation for the LeadOut Map initiative
  • Status: Private, operating
  • Website: blueconduit.com

The pitch in one line

BlueConduit is the company that proved data science could solve a public health emergency that nobody else could solve. The Ann Arbor team uses machine learning to predict which homes have lead service lines, and that capability has reshaped how US cities prioritize a federally mandated, multi-billion-dollar replacement program.

Founding story: from the Flint response to a national company

BlueConduit's origin story runs through the Flint water crisis. After the 2014-2015 lead contamination of Flint's drinking water became a national emergency, a team of University of Michigan data scientists led by Eric Schwartz (Ross School of Business) and Jacob Abernethy (then Michigan EECS, later Georgia Tech) was brought in to help the city find the lead service lines that had to be replaced.

The problem was that no city had an accurate inventory. Service lines (the pipe between the water main and the house) had been installed over a century, often without documentation, often misclassified, and frequently a mix of materials within the same neighborhood. Digging up every line to inspect it would have cost an order of magnitude more than the actual replacement and would have taken decades. The Michigan team's bet was that the existing patchy records, plus property characteristics, plus a small number of physical inspections, were enough data to train a machine learning model that could predict the probability that any given home had a lead service line.

The model worked. The Flint replacement program achieved a much higher hit rate than the alternative methods, the cost per lead-line found dropped dramatically, and the methodology was published. After the Flint work, demand from other cities became substantial enough that incorporating a company made sense. BlueConduit was founded in 2019 as the commercial vehicle for the methodology.

What they build

BlueConduit's platform is a machine learning system that ingests city water service records, parcel data, building characteristics, and historical inspection results, and outputs probability estimates for each home in a city's service area on whether the service line is lead, copper, galvanized, or another material.

The product spans:

Predictive inventory. The core ML capability: a service-line material prediction for every parcel in a customer city's service area, with confidence intervals.

Replacement program optimization. Tools that help cities prioritize digging order, route crews efficiently, and report progress to state and federal regulators.

LeadOut Map. A public-facing map (backed by Google.org and the Rockefeller Foundation) that surfaces lead service line risk information to residents and policymakers across multiple US cities.

Inventory compliance. Tools and methodology that help cities meet the federal Lead and Copper Rule Revisions, which require utilities to develop and submit lead service line inventories.

The customer model is municipal water utilities, with funding flowing through a mix of utility budgets, state revolving fund loans, federal infrastructure dollars, and foundation grants. The customer profile is unusual for a software company: city governments under federal regulatory deadline pressure, working through procurement processes that are not built for SaaS.

Funding and partnerships

BlueConduit's capital story is different from the typical venture-funded AI company. The company has been supported by a combination of foundation grants, federal partnerships, and customer revenue, with Google.org and the Rockefeller Foundation as named partners on the LeadOut Map initiative. Specific total funding figures and round details are not as cleanly disclosed as they are for venture-backed peers; the company has been deliberate about the structure of its capital base, given the public-interest nature of the customer mission.

That capital structure is the right one for the work. A traditional venture-equity company would be under pressure to expand into adjacent markets that pay more per seat than municipal water utilities do. BlueConduit's funding model lets the company stay focused on the lead service line replacement program, which is a finite-but-massive opportunity (the federal mandate runs through 2034 and beyond) without the distraction of pivoting toward higher-margin verticals.

Ann Arbor footprint

BlueConduit is headquartered in Ann Arbor and operates as a small but high-impact team. The recruiting profile favors data scientists with a public-interest orientation, software engineers comfortable with the messy data realities of municipal records, and customer-facing staff who understand water utility operations.

The U-M connection is direct. Schwartz remains at U-M Ross; Abernethy is at Georgia Tech. The company has continued to draw on U-M data science talent and has been a recurring presence at U-M data-for-public-good events.

Leadership and people

BlueConduit's leadership has been built around the founding research team plus operational hires from the water utility and government technology sectors. The company's public profile is anchored on Schwartz, Abernethy, and the broader founding team, though day-to-day operations are run by a CEO and operating bench. Specific current titles should be cross-checked on the company website before citing.

Why it matters

Three reasons BlueConduit matters in the Ann Arbor story.

First, it is one of the cleanest examples of academic data science becoming a company that solves a real public problem. The Flint methodology became a national model for lead service line inventory, and BlueConduit is the commercial vehicle that has scaled that model. That maps onto the University of Michigan tech startups story in a way that few other firms do.

Second, the public-interest AI thesis is strategically distinct. Most AI startups aim at enterprise IT, marketing, customer service, or financial services. BlueConduit aims at public-sector infrastructure, a market that is hard, slow, and underserved. That positioning gives Ann Arbor's AI bench a credible public-interest representative inside the broader Brookings AI Star Hub designation.

Third, the long-cycle opportunity is meaningful. The federal Lead and Copper Rule Revisions and the broader infrastructure investment cycle running through 2034 create a sustained demand environment for BlueConduit's specific capability. The company is positioned to be the durable infrastructure layer behind one of the largest municipal infrastructure programs in US history.

Related

  • Genomenon is the other prominent Ann Arbor AI company applying machine learning to a public-interest problem (genomic interpretation for rare disease).
  • Clinc is the U-M-spinout enterprise conversational AI company.
  • SkySpecs is the Ann Arbor climate-tech company whose enterprise sales motion intersects with BlueConduit's public-sector orientation.
  • The AI category hub covers the full Ann Arbor AI bench.
  • The companies directory is the master index.

FAQ

Where is BlueConduit headquartered? BlueConduit is headquartered in Ann Arbor, Michigan.

Who founded BlueConduit? BlueConduit was founded in 2019 by Eric Schwartz (University of Michigan Ross School of Business) and Jacob Abernethy (then U-M EECS, later Georgia Tech), along with collaborators from the U-M data science team that worked on the Flint water crisis response.

What does BlueConduit do? BlueConduit uses machine learning to predict which homes in a city's water service area have lead service lines. The platform helps utilities prioritize replacement programs, comply with federal regulations, and report progress to residents and regulators.

Who are BlueConduit's customers? Customers span more than 50 US cities, including Flint, Detroit, Trenton, and Toledo. The customer model is municipal water utilities working under federal Lead and Copper Rule Revisions deadlines.

How is BlueConduit funded? The company is supported by a mix of foundation grants, federal partnerships, and customer revenue. Google.org and the Rockefeller Foundation are named partners on the LeadOut Map initiative. Specific funding totals and round details are not as cleanly disclosed as for venture-backed peers.

What is the connection between BlueConduit and the Flint water crisis? The methodology BlueConduit commercializes was developed by University of Michigan data scientists working on the Flint water crisis response in 2016-2018. The model proved that machine learning could predict lead service line locations more accurately and cheaply than the alternative inventory methods. BlueConduit was incorporated in 2019 as the commercial vehicle for the methodology.

Is BlueConduit hiring? Yes. Open roles in data science, machine learning engineering, software engineering (geospatial and cloud), project management for utility deployments, and policy partnerships are posted at blueconduit.com/careers. The work model is hybrid with Ann Arbor as the home base and meaningful remote flexibility for specific roles.

Strategic reading

BlueConduit sits at the intersection of municipal regulation, federal infrastructure policy, and applied machine learning. The market tailwind is specific. The Lead and Copper Rule Improvements finalized by the EPA in October 2024 require water utilities across the United States to develop detailed service-line inventories and to replace all lead service lines over a roughly 10-year horizon, with limited exceptions. That regulation created a multi-billion-dollar spending cycle across every water utility in the country, and every utility needs a defensible method for identifying lead lines. Predictive modeling is explicitly recognized as a permissible methodology.

The competitive landscape is limited. Some large engineering consulting firms do lead service line identification work as part of broader water-utility engineering contracts, and a few other small startups have attempted the same problem. BlueConduit's combination of academic methodology, peer-reviewed publication track record, foundation and federal partnerships, and incumbent position across more than 50 cities makes it the default choice for a utility that wants a defensible predictive inventory. The business does not need to be the only answer in the market. It needs to be the right answer often enough to sustain the company through the regulatory cycle.

What would reorder the read. Federal regulatory pressure could soften under a different administration, though the current regulatory trajectory appears stable. A large engineering consulting firm could build comparable capability in-house and use existing utility relationships to win share. Data-quality issues in specific cities could produce bad press and slow adoption. None of those risks are visible at scale today, but they are the ones worth watching.