Working with AI Agents in the Public Sector: What Works (and What Doesn’t)

AI agents are systems that can autonomously retrieve information, use tools, and carry out multi-step actions to achieve goals. In government, early use cases include agents helping public professionals conduct research, prepare case files, analyze applications, coordinate workflows, and complete recurring administrative tasks. At the same time, residents are beginning to use agents to navigate government services, search for benefits, complete applications, understand eligibility requirements, and access personalized assistance.

Yet the term "AI agent" is often used loosely. Not every chatbot is an agent, and not every workflow needs one. As governments begin experimenting with these tools, important questions remain: What exactly are AI agents? How are they different from copilots, automation tools, and traditional software? Where are they creating value for public professionals and the public? What risks and limitations are emerging? And what does responsible adoption look like in practice?

This four-part webinar series explores how governments are beginning to use AI agents to support both internal operations and public service delivery. Through real-world examples, participants will learn where agents are proving useful, where limitations and risks are emerging, and what agencies are learning about governance, oversight, and responsible implementation

Learning Goals

• Explain what AI agents are and how they differ from chatbots, copilots, workflow automation, and traditional software tools.
• Identify current and emerging use cases for AI agents in government operations and public service delivery.
• Assess when agentic approaches are likely to add value and when simpler technologies may be more appropriate.
• Recognize key risks and limitations, including reliability, privacy, security, transparency, and accountability concerns.
• Apply practical lessons from government experience to support responsible experimentation, governance, procurement, and deployment of AI agents.