CARE: How South Bend is Combining Predictive AI with a Human Touch to Remake City Services

“I would be ecstatic,” said one interviewee about whether a service to connect them with help would be useful to them. “I never knew where to get help.” The City of South Bend’s Department of Innovation and Technology has been working to understand the root causes of individual crises and what we can do to help, by speaking with at-risk residents about what would be most helpful to them. This research has led us to one of our biggest projects yet: launching the Community Assistance Risk Estimator (CARE) with the Bloomberg Mayors Challenge.

TLDR: The City of South Bend was awarded a $1 million grant from the Bloomberg Mayors Challenge to develop and implement CARE, a product that will proactively identify risk factors using resident and public data to help residents before a crisis occurs.

The Problem: A System Built to React, Not Prevent

As South Bend Chief Innovation Officer Madi Rogers puts it, “A lot of what we do in local government is responding after the fact. We often only identify problems when residents call 311. By that point, a missed utility payment may have turned into a water shutoff, a housing challenge into an eviction filing, or a small neighborhood issue into something much larger.”

What if, instead of reacting after the case, the City of South Bend could meet residents where they are, provide them with opportunities to get help, and improve neighborhoods before they get deteriorate further? That’s the question that CARE hopes to answer — if we can use data and artificial intelligence to help residents before they reach a crisis point and, furthermore, to help improve their lives by supporting them, their neighbors, and their block.

Introducing CARE: Proactive Help, Before Crisis Hits

CARE is a data model that draws on existing city data such as utility records, 311 calls, eviction data, code enforcement data, and more to identify residents and places that may need proactive outreach, before they have to ask or before they reach a crisis point. In collaboration with Notre Dame’s Lucy Family Institute Data and Society, South Bend’s Civic Innovation, Resident Success, Data and Performance, Digital Services, and Digital Integrations and Development teams are working together to turn CARE from a vision into a real, useful product. This work couldn’t be done without Bloomberg Philanthropies and the Bloomberg Mayor’s Challenge.

In late 2025, the City of South Bend was selected as one of 24 worldwide Bloomberg Mayor’s Challenge grants. Along with cities such as Boise, Rio de Janeiro, Ghent, and Cape Town, South Bend was awarded $1 million to make our concept a reality. South Bend stands out among the rest of the awarded cities for many reasons: our size (small in comparison with most of the others), our innovative thinking, and our commitment to using human-centered design to proactively help our residents.

High Tech, Deeply Human: How CARE Works in Practice

While CARE is a data model and uses AI, an important distinction is that it is going to be deeply human in its design. AI doesn’t help anyone on its own; it helps the humans behind the product focus so they can show up for people in the right place, at the right time, and with the right resources.

This is apparent in our City’s Resident Success Team, who regularly meet one-on-one with new residents and residents in need of support, and are on the front lines of connecting people with the resources they need. The Resident Success Team has been integral to the project from the very beginning, especially with identifying and interviewing key residents to get their perspective on how this product might serve them.

What Residents Told Us About AI and Their Data

One prototype example we initiated this summer was a proactive SMS outreach campaign connected to planned community clean-up days. We got an overwhelmingly positive response to these messages and found that people were pleased to learn before the fact about cleanup opportunities in their neighborhood.

When developing this concept, it was important to us that we knew what residents would feel comfortable with. We were concerned that our neighbors might feel surveilled or uneasy with the concept of AI using their data this way. After conducting numerous engagement sessions throughout the summer, we learned that residents were glad that their data could work for them. If the data being used is public data or what we already have, and we’re stewarding it with care, residents did not have a problem with it being activated in this way.

This is the most important part of CARE for the City of South Bend team. We are putting guardrails in place and setting principles that we will adhere to in order to handle resident data with the utmost care, privacy and security. The use of data is meant to help residents so we can effectively identify when they need extra support — and it will not be used in ways that cause harm or limit opportunities.

What’s Next for CARE

The tool is currently still in development. A long road is ahead of us with stakeholder mapping, wire-framing, prototyping, beta testing, and more. It’ll involve lots of listening, lots of data analysis, and time and care to get this moving. We’re looking forward to the journey and can’t wait to provide another update soon! In the meantime, if you have thoughts for us, or are interested in learning more about the project, feel free to leave a comment below!