Responsible Agentic AI Playbook: A Practical Guide for State and Local Government Agencies

Agentic Artificial Intelligence (AI) can perform tasks and make decisions with less direct human involvement than traditional AI tools, creating new possibilities for government operations—and new risks. This Urban Institute playbook provides state and local governments with practical guidance for using these systems responsibly, with attention to oversight, security, transparency, and accountability.

For Temporary Assistance for Needy Families (TANF) agencies, the playbook can help leaders think through where AI may genuinely improve operations and where human review remains essential. It provides a framework for considering issues such as privacy, bias, reliability, and accountability before putting AI into practice. For TANF programs handling sensitive information and making decisions that affect families, planning before adoption is critical to using new technology without compromising trust.