The State Department has released a generative artificial intelligence (GenAI) playbook for federal agencies looking to responsibly adopt and scale enterprise GenAI.
In the playbook, the department shares its experience developing StateChat and its enterprise AI platform, aiming to help agencies accelerate innovation and deliver better services to the American people.
“We hope this playbook can be a companion to you in your AI journey,” State Department CIO Kelly Fletcher and acting Chief Data and AI Officer Amy Ritualo wrote in a letter accompanying the playbook.
“Whether you’re exploring AI’s potential or scaling existing initiatives, the frameworks, tools, and insights included here reflect our lessons learned along the way and we hope they will help you anticipate challenges, seize opportunities, and deliver meaningful results for your agency and the American public we serve,” they said.
StateChat is the State Department’s first GenAI chatbot. The department deployed the tool in 2024 with support from the General Services Administration’s Technology Modernization Fund.
As of June 2026, StateChat has over 62,000 users and adoption across 98% of U.S. diplomatic posts worldwide.
The Bureau of Diplomatic Technology’s Center for Analytics led the design, development, and deployment of StateChat, as well as the broader enterprise GenAI platform.
In parallel to StateChat, the Department of State created the infrastructure to support safe, scaled GenAI development and adoption. This included novel testing procedures, tailored training programs, and responsible use policies.
“Critical additions to the enterprise GenAI platform came next, including secure APIs to generative AI models and sandbox environments where technical experts and ‘citizen developers’ could prove new ideas,” the playbook says. “In every chapter of the story, users led the way.”
The document also highlights the importance of change management after deployment, including executive sponsorship, targeted training, communications campaigns, and performance measurement.
Looking ahead, the playbook outlines the State Department’s plans to evolve beyond standalone AI tools toward “agentic AI systems” that coordinate multiple specialized AI agents to complete complex workflows.