Nonprofit puts up $10M to help agencies test AI tools for benefits modernization
Vithun Khamsong via Getty Images
The Center for Civic Futures will offer funds to agencies looking to design and test their tech solutions to operational problems with their benefit systems.
Before any tech solution can be implemented in government, it should be tested. That’s the motivation behind a nonprofit’s funding initiative that aims to help public sector agencies and other organizations test their artificial intelligence-based public benefit improvement projects.
Up to $10 million in funding is up for grabs for organizations looking to explore and experiment with emerging technologies for enhancing the delivery of public benefits, the Center for Civic Futures, a nonprofit dedicated to assisting governments adopt AI responsibly, announced this week.
The funds, administered through CCF’s Public Benefits Innovation Fund, will be divided among awards ranging from $150,000 to $2 million to support benefit modernization projects that leverage AI to improve backend processes. Government agencies, startups, nonprofits, academic institutions and other organizations are encouraged to apply.
The Public Benefits Innovation Fund announcement comes as benefit agencies scramble to comply with federal rule changes to Medicaid and the Supplemental Nutrition Assistance Program. The new rules, established by President Donald Trump’s “One Big, Beautiful Bill” will pressure agencies to revamp their enrollment and eligibility systems in order to comply with new work requirements and cost-sharing rates.
“We're at this truly transformative moment, and what really matters is that government leaders have safe, candid spaces where they can talk to each other about what's going on, and that they're not having to learn things on their own in silos,” said Cass Madison, executive director of CCF.
While improving constituents’ frontend experiences of a benefit system through tools like chatbots or eligibility screeners has been popular across benefit agencies, Madison pointed to a growing implementation gap facing those agencies.
“You can make it easier for people to understand that they’re eligible for SNAP and fill out an application, but if, once their application is submitted, and it gets caught in backend, administrative backlogs [like] verification processes, I think there’s a really strong argument that we actually haven’t made people’s lives better,” she explained.
That’s why the PBIF prioritizes three focus areas: improving backend processes, using data effectively and modernizing tech infrastructure. Proposed projects can, for instance, entail an organization's use of generative AI to review SNAP case files and flag errors for remediation, according to CCF’s request for proposals.
Other acceptable project examples include tools that anonymize or otherwise prepare sensitive datasets used for training AI models for benefit determination and administration, leverage AI to enable system modularization or code modernization, offer sandbox environments to test and train AI models, and others.
“We believe the goal of experimentation is to learn what works and what doesn’t in a safe and contained way, so that the broader field can adapt and improve,” the request for proposals reads.
The PBIF also entails two tracks for applicants to pursue, based on their project status. One track is designed for agencies that have “shovel-ready” prototypes ready to be tested and validated in real-world environments, Madison said. Awardees in the pilot track will have 12 to 24 months to work on their projects and receive investments between $500,000 and $2 million.
The other track will serve agencies that have projects in the research and development phase, enabling them to test “a specific hypothesis in a controlled or sandbox environment,” according to the request for proposals. This track will help organizations demonstrate that their proposed intervention can resolve an administration problem, which will take six to 12 months, Madison said. Awardees in the early-concept category will receive up to $500,000.
Beyond funding, awardees will also have access to a cohort of other organizations dedicated to “maximize support for their projects and their chances of success,” Madison said. For example, CCF has an agreement with Better Government Lab to offer further technical assistance, such as evaluating project impact or ROI.
The Recoding America Fund will also serve as a partner to help agencies address structural and operational barriers to adopting innovative tech solutions for benefit system enhancement.
Applications to participate in the Public Benefit Innovation Fund are due May 15. CCF will host a webinar April 22 for candidates to learn more about the initiative and application process.