Recent changes in federal funding flows and concerns about fraud have impacted the Supplemental Nutrition Assistance Program (SNAP). Millions of low-income families in the U.S. have lost, or are at risk of losing, access to their food benefits, but groups including CfA are advancing efforts to help support continued access to SNAP benefits for eligible recipients that rely on them.
The partnership announced Friday marks the organization’s latest attempt to support this work by designing, piloting and deploying Claude-based tools to support government caseworkers’ administration of benefits, starting with SNAP.
The SNAP Policy Navigator will be the partnership’s initial focus. It is a Claude-powered integration allowing caseworkers to access real-time information about SNAP from federal, state and county policies to navigate the administrative complexity, which program changes increase.
The tool is built on Model Context Protocol, a standard enabling two-way connections between AI applications and trusted external data sources. It aims to support responsible deployment of AI in high-stakes sectors in which critical benefits determinations are made, because every response is informed by up-to-date government data from trusted sources. The intent is to help SNAP Policy Navigator users get answers to case-specific questions with this tool.
“SNAP caseworkers carry an enormous load, interpreting complex rules under tight timelines for the families who depend on the safety net,” Elizabeth Kelly, head of beneficial deployments at Anthropic, said in a statement, underlining that the new tool aims to help more eligible families get access to benefits.
The pilot is intended to create tools that can be adapted and reused across states and counties to make service delivery more modern and efficient while remaining compliant with various policies. Its duration was not specified in the news release. The partnership is slated to deliver a suite of Claude integrations that support benefits administration; these tools may help answer policy questions, draft communications and aid in the review of eligibility documents.
The announcement, made during the 2026 Code for America Summit in Chicago, closely follows CfA’s release of the 2026 Government AI Landscape Assessment, which evaluated states' AI readiness. That assessment found many states have made progress in AI readiness and implementation, but progress is still fragmented and gaps remain. CfA is working to help government use AI responsibly to be “efficient, effective, and more empathetic,” according to a statement from the organization’s CEO Amanda Renteria.
