Close

New Workspace Modernization Research from CDW

See how IT leaders are tackling workspace modernization opportunities and challenges.

Feb 19 2026
Artificial Intelligence

Agentic AI Turns Government Workflows Into Autonomous, Governed Systems

Artificial intelligence agents can help state and local agencies improve service delivery and extend resources.

Generative artificial intelligence (GenAI) can power chatbots and help draft emails. Agentic AI goes further, giving state and local governments an automated means to plan and take actions and coordinate tasks across systems.

CIOs and agency leaders must pair these autonomous capabilities with strong guardrails. When they do, agentic AI can reduce manual work, speed up citizen services and inspections, and free staff to focus on higher-value decisions.

Where GenAI merely answers questions, agentic AI can trigger workflows. “Agentic AI is the next-step change above what we saw in the initial language models, where you’re just able to get summaries and all those kinds of interesting pieces back out of your data,” says Google Public Sector Field CTO Chris Hein.

“Agentic AI is that, plus the ability to use tools, to actually have memory and context and reasoning,” he says.

Click on the banner below to explore the first steps of AI adoption.

 

How Can Government Agencies Use Agentic AI?

Agentic AI in government has powerful potential for state and local agencies.

“If you have a goal to be fulfilled, these agentic AI systems can help subdivide those goals into smaller units of work, and can sequence them. They can see what has to be done on each step,” says IEEE Senior Member Arpita Soni. In some cases, the AI agents can take the steps needed to meet those goals.

“It is not just responding to our queries, like traditional chatbots. It is doing something. These agentic AIs are like digital co-employees, digital coworkers,” she says.

In the realm of citizen services, that digital coworker can carry some of the load. “There’s a wealth of opportunity when we look at how to make citizens’ experience of government better,” Hein says.

Agentic AI use cases include:

Automated Permit Prepopulation

When residents start the permitting process, agentic AI “can guide the applicant who is filling out those application forms: ‘This belongs to this, and it has to be completed with this.’ As a guide or assistant, it will help the applicant to fill out the forms with fewer errors and a higher chance of getting it right the first time,” Soni says.

READ MORE: AI is a top management priority for governments in 2026.

Intelligent Service Request Routing

Requests may come in via phone or online, or even through paper forms. Agencies need a way to get them into the right hands, and an AI agent can speed that process. “You can build out a custom workflow that really pertains to that particular agency and what it’s trying to accomplish,” Hein says.

Benefits Application Processing

Benefits applications can be daunting. “What if I am recently unemployed? What are my options?” Hein says. “If I need healthcare, what are the things that I need to know that the government can help me with?”. An AI agent “has the full contextual understanding of me and what my background is with the government, as well as access to these full sets of tools and services. It can then go on my behalf” and get the wheels turning.

Chris Hein
There’s a wealth of opportunity when we look at how to make citizens’ experience of government better.”

Chris Hein Public Sector Field CTO, Google

How Does Agentic AI Improve Government Processes

By applying agentic AI to administrative workflows, agencies can help make workers more efficient, effective and responsive to constituent needs.

AI use cases in government workflows include:

Caseworker Support and Follow-Up

To support case workers, agentic AI can help keep cases moving forward, for example by identifying missing information and automatically contacting clients for updates, as well as by creating drafts follow-up messages, reminders or requests for verification. “For every single knowledge worker task that sits in front of you, the question is: How do you streamline that?” Hein says. Agentic AI offers “expertise that you can trust, and it’s based in your own data set. It’s not off hallucinating based on what it read on the internet somewhere.”

Inspection Scheduling and Prework

In support of inspections, agentic AI can draft notes and reports and follow up on compliance. Google, for example, is tapping that potential in a partnership with the New York City Metropolitan Transportation Authority, leveraging agentic AI to understand where there might be defects on the rails and identifying how to fix them faster. “We’re also building out a track inspection program that is intended for the people who are out physically inspecting these trains and these train lines. They can take pictures of things, they can do a ticketing system,” Hein says. “It helps them to get trained up faster on that entire process.”

Meeting Preparation and Documentation

Meeting prep can be tedious. “You are trying to book an appointment, managing calendars. All of those things are done manually today,” Soni says. An AI agent can gather relevant files, case notes and emails; it can generate a meeting brief summarizing key issues and deadlines; and after the meeting, it can draft minutes and highlight follow-up tasks. “It reduces that tremendous amount of administrative effort,” she says.

DIVE DEEPER: Governments must brace for the transformational impact of AI.

Can Agentic AI Optimize and Extend Government Resources?

Agentic AI supports state and local financial management efforts.

“In budget analysis and resource optimization, agentic AI can learn from the data and do the analysis, whether you are spending or saving. It will check your patterns over time,” Soni says. “And when your budget becomes tight, that spending analysis will help the project teams to understand what resources they can shuffle and how they can distribute their workload.”

AI agents also can be trained to spot fraud and abuse and to trigger remediations. “That’s one huge category for cost savings,” Hein says.

State and local leaders need to double down on governance when implementing agentic AI: With the ability to deliver information and take autonomous action, strong guardrails are needed.

“You should have ethical and responsible agentic AI principles: clearly defined boundaries around what agentic AI can do, what it cannot do and when to raise an alarm around it,” Soni says.

And leaders will need to implement training in order to bring that governance to life. “You’re going to need to upskill those users,” Hein says.

“The city of Los Angeles has been doing a phenomenal job with this, as they’ve enabled Gemini throughout their entire 27,000-employee workforce. They’re making sure that everybody goes through training so that they know what it means to give an AI agent access to specific tools or access to specific data,” he says.

Workers will need training on agentic AI, he says, “so that they are doing it responsibly and ethically, and they understand what bias can exist in the systems and how to look out for it.”

Boy Wirat/Getty Images