States Cautiously Explore Agentic AI's Potential

Report finds tech leaders are taking a gradual approach to advanced AI that can automate tasks without human intervention.

Published on Mar. 10, 2026

A new report from the National Association of State Chief Information Officers (NASCIO) indicates that state governments are exploring the use of agentic AI, a more advanced form of AI that can automate tasks and make decisions independently. However, tech leaders are taking a cautious approach, starting with pilot projects and focusing on security and privacy concerns before wider deployment.

Why it matters

As governments face workforce shortages and the need to do more with fewer resources, agentic AI has piqued the interest of state leaders as a potential force multiplier. But the technology also poses greater security and privacy risks than previous AI tools, requiring a gradual and thoughtful approach to implementation.

The details

The report outlines five phases of agentic AI maturity, with the most advanced phase involving AI agents initiating work without prompting. While no states have fully reached this level yet, some are taking initial steps, such as Virginia's agentic AI pilot program and Delaware's AI sandbox initiative. To mitigate risks, the report suggests strategies like monitoring AI actions, limiting permissions, and requiring human oversight.

  • In July 2026, former Virginia Governor Glenn Youngkin signed an executive order to launch the nation's first agentic AI pilot program.
  • Also in July 2026, Delaware lawmakers launched an AI sandbox initiative to allow startups and industry to develop agentic AI solutions in a controlled environment.

The players

National Association of State Chief Information Officers (NASCIO)

An organization that represents state chief information officers and IT executives.

Amy Glasscock

Program director for innovation and emerging issues at the National Association of State Chief Information Officers.

Glenn Youngkin

Former Governor of Virginia.

Tennessee Department of Finance and Administration

A state agency in Tennessee that released a request for information on how generative and agentic AI could influence its enterprise resource planning system.

Got photos? Submit your photos here. ›

What they’re saying

“Like with any AI, we always recommend starting small with pilot projects [or] doing things internally before using it externally on citizen services … before it's scaled and could cause widespread problems.”

— Amy Glasscock, Program director for innovation and emerging issues at the National Association of State Chief Information Officers (Route Fifty)

What’s next

The judge in the case will decide on Tuesday whether or not to allow Walker Reed Quinn out on bail.

The takeaway

This report highlights the cautious approach state governments are taking as they explore the potential of agentic AI, balancing the technology's promise as a force multiplier with the need to address security and privacy concerns through a gradual, thoughtful implementation process.