Reducing Frictions in Access to ChildCare: A Pilot Evaluation of AI-Guided Family Support

Access to subsidized childcare in the United States is hampered by informational and administrative frictions. In New Haven families must navigate a fragmented landscape of programs -- including Head Start, school choice magnet programs, and subsidy programs for private providers like Care4Kids and School Readiness -- each with different eligibility rules, application procedures, and providers, often without coordination or accessible guidance. These barriers may disproportionately affect lower-income and non-English-speaking families. This project conducts a randomized evaluation of a bilingual AI-powered agent designed to help families in New Haven find, compare, and access childcare programs. Families who applied to New Haven Public Schools pre-K programs but did not receive placement will be randomly assigned to receive access to the tool or standard informational resources during the 2026 enrollment cycle. The AI agent integrates a comprehensive database of licensed childcare providers with subsidy eligibility information and delivers personalized, conversational guidance to support families through the application process. The study will measure effects on childcare enrollment, subsidy program applications, search effort, and the quality and match of childcare arrangements. The project provides rigorous evidence on whether AI-based tools can reduce inequality in access to early childhood education at low marginal cost.