NASCIO 2026 Midyear: States Shift From AI That Assists to AI That Acts, Tennessee CTO Says

Building the Foundation Before Scaling AI

Tennessee’s first priority was establishing governance and structure before expanding AI use.

Jones emphasized that states must start with enterprise security policies aligned to frameworks such as those from the National Institute of Standards and Technology, along with formal oversight bodies.

The state created an AI advisory council in statute, bringing together IT leaders, lawmakers, universities and industry partners to guide strategy. It also established a review committee to vet proposed AI use cases, requiring agencies to justify business value, funding and data readiness before launching pilots.

“Everybody has a great idea for AI,” Jones said. “You have to have a way to filter that.”

The process helped Tennessee avoid early missteps, such as approving projects before confirming whether the underlying data was accessible or usable.

READ MORE: Strategic planning can overcome AI adoption challenges.

Pilots Focus on Real-World Impact

Rather than pursuing a single use case, Tennessee launched a portfolio of pilot projects across legal, citizen services, IT operations and back-office workflows.

One effort uses AI to streamline public records requests by retrieving and redacting information from emails, chats and agency systems — a process that previously required significant manual effort from attorneys and IT staff.

Another pilot involves a statewide chatbot framework designed to help residents navigate services such as benefits eligibility, avoiding a patchwork of one-off solutions across agencies.

The state is also applying AI to IT service management, analyzing hundreds of thousands of annual support tickets to identify patterns and recommend resolutions. In the governor’s office, AI is being tested to help manage appointments and workflows for boards and commissions.

Jones said these varied pilots are intentional, allowing the state to test different technologies and integrations while building a foundation for broader adoption.

Workforce Trust Is the Biggest Hurdle

For Jones, the biggest challenge is not the technology itself but preparing the workforce to use it.

“It’s not IT, it’s the people,” he said. “You really need to ensure that you have a strong change management process in place.”

Tennessee is investing in AI literacy training, requiring employees to learn the basics before gaining access to tools. The state has also built a network of “AI champions” embedded across agencies to share knowledge and identify opportunities.

Clear policies are critical as well, particularly around which data can and cannot be used in AI systems.

“You’ve got to have transparent ground rules,” Jones said. “Well-intentioned use can still get you in trouble.”

Rather than replacing workers, the state is positioning AI as a tool to augment existing processes. Leaders are also rethinking staffing, reallocating positions to build AI expertise without adding headcount.

“We’re not getting new positions,” Jones said. “So, we’re retooling what we have.”

LEARN MORE: Government agencies harness AI to improve citizen services.

Moving Toward Automation, Carefully

Tennessee is now beginning to explore more advanced capabilities, including AI-driven automation of workflows.

One goal is to use AI to provision IT environments automatically after a request is approved, reducing turnaround time from days to hours. Longer term, the state is working toward systems that can analyze citizen data and proactively identify services that residents may need.

Still, Jones stressed that automation must be introduced gradually, starting with low-risk back-office functions.

“Start in the back office,” he said. “If it goes sideways, nobody’s going to die.”

Strong guardrails are essential as systems become more autonomous. Tennessee is prioritizing capabilities such as full auditability of AI decisions, strict access controls and “circuit breakers” that can halt processes if something goes wrong.

“Everything has to be traceable back to the decision,” Jones said. “If you can’t replay it, you can’t trust it.”

Ultimately, Jones said, the shift to agentic AI represents a broader transformation in how government operates.

“We’re moving from systems of record to systems that act,” he said.

In that model, AI does more than store or retrieve information. It becomes an active participant in workflows: gathering data, initiating tasks and supporting decision-making with human oversight.

NASCIO is beginning to frame that transition as the next phase of AI maturity. In a recent report, “Beyond Generation: The Rise of Agentic AI in State Government,” the association outlines how states are evolving from assistive tools toward systems capable of managing multistep processes and taking limited actions on their own. 

Amy Glasscock, NASCIO’s program director for innovation and emerging issues and author of the report, said most states are still in the early to middle stages of that progression, with few fully embracing autonomous capabilities. The path forward will require balancing innovation with governance, security and trust, she said.

For Jones, that balance is already shaping how Tennessee approaches the next phase.

“Agentic is coming,” he said. “But you’ve got to build the foundation first.”

Bookmark this page for our coverage of the NASCIO 2026 conference.