After spending years testing chatbots and exploring generative AI capabilities, federal agencies are beginning to deploy AI agents.

Cameron Groves, director of rapid innovation and specialist engineering at Google Public Sector, said agencies are increasingly focused on using these AI agents to complete tasks and automate workflows.

“It’s been unbelievable what’s unfolded the last couple of years — where we’ve seen a lot of federal entities move from this notion of pilots or testing out chatbot experiences, to last year looking at agents and exploring how agents can ultimately help them be able to do certain tasks. Now, what we’re seeing is enterprise-ready AI using full end-to-end agents,” Groves said during Federal News Network’s Cloud Exchange 2026.

Leading federal agentic AI initiatives

The Defense Department is leading the federal adoption of AI tools and has partnered with Google on its GenAI.mil initiative. Groves said that DoD, using Google’s Gemini for Government platform, has built over 100,000 AI agents within a short period of time.

“What we’re seeing there is well over a million users using that front door,” he said.

Federal agencies are also beginning to use AI agents to tackle complex mission critical processes. At the Food and Drug Administration, drug approval reviews can take years, and scientists often must analyze lengthy submissions from pharmaceutical companies.

FDA is using AI to speed up that review process, which improves productivity and gives researchers more time to do impactful work.

“What that allows them to do is refocus the scientists to do more scientific research types of tasks related to being able to enhance the clinical trial process,” Groves said.

The Transportation Department is embedding AI capabilities into its workplace productivity tools. Groves said the department’s workforce of about 50,000 employees is using Google Workspace, which includes its Gemini platform, to be able to review and automate tasks.

Preparing for AI expansion

As agencies face budget and workforce reductions, Groves said AI enables them to do more with less. But, he stressed, these tools shouldn’t be viewed as a replacement for the workforce.

“It’s not about replacing humans in terms of individuals, with regard to the use of AI. It’s augmentation,” he said.

Upskilling employees to work with AI tools is critical to moving forward with adoption. Groves said agencies should help their employees understand both the capabilities and limitations of these tools.

“It’s a really important point, in terms of being able to be aware of the power of these tools,” he said. “There’s a lot of fear out there, in terms of the unknowns, fear regarding job replacement and so forth.”

Security and compliance remain foundational

The rise of AI is also changing cybersecurity. Threat actors can now use AI to identify software vulnerabilities at machine speed, forcing defenders to adopt similar capabilities.

“AI itself has upset the paradigm shift in terms of security,” Groves said. “We have to be able to move at that machine speed and move beyond the human in terms of being able to review it but also be able to then remediate it.”

Google recently launched its AI Threat Defense platform, which uses AI to address network security vulnerabilities before they’re exploited by malicious actors. Groves said the rapid rise of adversarial AI has led to a paradigm shift in cybersecurity.

“We have an approach that we have adopted internally to be able to look, in terms of the code itself, be able to scan and prioritize, be able to then harden the actual code itself, and then from there monitor,” he said. “We have agents that are actually being applied to this new paradigm shift that can proactively do this.”

Scaling AI beyond pilot programs

Despite widespread interest in AI, many agencies continue to struggle with moving from pilot projects to enterprisewide deployments. Groves said agencies often move too quickly into deployment before getting a full understanding of the processes they’re looking to automate and before employees are familiar with these tools.

“They need to understand what exactly it is, and be able to give context, in terms of the actual capability that’s been given to them,” he said. ‘“It’s really understanding the personas, looking at the workflows, and then from there being able to then work with them to be able to train them in how to use the technology to do certain things.”

For agencies just beginning their AI journeys, Groves said some low-hanging fruit offers the most immediate return-on-investment is automating administrative tasks. That includes helping acquisition professionals write and evaluate requests for proposals, assisting with permitting reviews and preparing daily briefings.

The Uniformed Services University, for example, is using AI to streamline medical coding and documentation, reducing administrative burdens for clinicians.

“We’ve partnered with them to be able to use AI for the surgeon, at the completion of their case, to be able to speak into an application. And AI then takes the speech that’s generated. From there, it extracts the codes themselves to be able to help them streamline their work,” Groves said.

Building trust through responsible AI

As agencies deploy AI more broadly, trust remains a critical component of successful adoption.

“I don’t think they’re just buzzwords, I think they’re very important,” Groves said. “At Google, we take this very seriously, in terms of responsible AI. When you interact with the actual technology, the results that are returned have grounding in them, and have references and sources related to them, so you’re able to see very easily where the large language model — in this case, Gemini — is actually referring to how it got to its answer, and that allows them to be able to gain trust.”

Agencies must also focus on change management and identify employees who can demonstrate the value of AI to their peers. Looking ahead, Groves said AI capabilities will continue to advance rapidly across government.

“The world that we’re in today is the worst it’s ever going to be, meaning it’s going to get increasingly better by the moment,” he said. “It’s truly exciting what going on, and we certainly feel very privileged to be at the forefront of all this.”

Because of the rapid pace of AI breakthroughs, Groves said agencies need to stay informed on what’s happening with this technology and understand what works best for them.

“If you don’t, it’s going to pass you by almost in a matter of minutes,” Groves said.

For federal agencies seeking to improve efficiency, strengthen cybersecurity and modernize mission delivery, he said AI is becoming an increasingly essential tool.

“You’ve got to be able to move mission speed. With the world that we’re in today, and the speed at which it’s moving, being able to look at things from a new light — and using the technology that’s there — truly is revolutionary.”

Discover more articles and videos now on the Cloud Exchange event page.

Copyright © 2026 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.