As state governments move beyond artificial intelligence (AI) experimentation and toward enterprise deployment, state chief information officers (CIOs) and industry leaders say success will depend on strong governance and practical use cases.
Speaking during MeriTalk’s State Tech Vision 2026 roundtable following NASCIO’s 2026 Midyear Conference in Philadelphia, leaders from Nevada, Virginia, and the private sector said states are increasingly focused on meaningful AI outcomes rather than hype.
“AI is no longer a novelty,” said Tim Galuzzi, CIO of Nevada and NASCIO president. “It is becoming an operational force for government.”
Governance and data strategy drive AI adoption
Throughout the discussion, panelists repeatedly emphasized that successful AI adoption depends on strong governance structures and clear data strategies.
“I look at good governance, good policy as setting the guardrails in order to enable agencies to move fast,” Galuzzi said.
Michael Watson, CIO of the Virginia Information Technologies Agency, said government agencies must establish clear guidance around how AI tools are used while ensuring transparency and trust with citizens.
“The really interesting thing about AI is that it can do just about anything at this point, right?” Watson said.
“One of our primary jobs is to make sure that we have transparency and trust in the organizations that we operate,” he added. “And if we don’t have a clear understanding about how we’re using [AI], whether they’re appropriate situations … we need to make sure that we’ve got those things in place.”
In a separate State Tech Vision session, Chris “CT” Thomas, technical director in the global office of the CTO at Dell Technologies, similarly said organizations need governance boards and clear stakeholder engagement to scale AI successfully.
“The biggest piece … is setting up a governance board,” Thomas said. “If I can show that publicly to the constituents or to the organization internally, I’ll start to establish actual trust for the deployment.”
Panelists also stressed that AI readiness begins with understanding government data and properly classifying sensitive information.
“The data is foundational,” Galuzzi said. “One of the things that we just did is our chief data officer and his governance committee just launched a data classification policy for the entire state.”
“That just helps educate all of our users on what is the sensitivity level of this data? What is public information? What needs to be kept secure and confidential? But that’s the first step,” Galuzzi said. “You have to have that understanding across your entire enterprise on that data classification.”
Watson said governments are still working through challenges tied to identifying authoritative information and determining which data sets are appropriate for AI systems.
“We have to design security procedures and structures to make sure that we’re only doing that with the right type of data,” Watson said.
Greg DeYoung, regional vice president for state, local, and education at Elastic, said many states have established AI policies but still lack data strategies.
“AI is experimenting; there’s a lot of novelty to it,” DeYoung said. “But unless you have a real data strategy around what you have, what do you trust, who owns it, and then how are you feeding those AI workflows, there’s a big gap there that a lot of states have to overcome.”
States prioritize practical AI use cases
Rather than pursuing flashy AI deployments, panelists said states are concentrating on achievable use cases that reduce friction for both employees and residents.
Watson said agencies are already using AI to summarize documents, accelerate research, and improve internal analysis processes as governments contend with ongoing workforce and resource constraints.
“We are all strapped for resources,” Watson said. “Anything that’s going to make the analysis process a little bit cleaner, a little bit clearer, we take advantage of those tools.”
Thomas pointed to emerging use cases ranging from public information kiosks and emergency services assistance to water management and computer vision technologies that support public safety efforts.
Galuzzi said one of the biggest opportunities for AI lies in simplifying how residents interact with government services.
“I think in the next year we’re going to see … less friction in government,” Galuzzi said. “Our constituents shouldn’t have to figure out our complicated bureaucracy in order to get services … They just need to ask the question.”
“They’ll have an agent or a chatbot that can provide them that information, and I think that’s going to be what a lot of organizations focus on first, because that’s where the real impact is going to be. It’s going to help residents and citizens get services faster, better, more efficiently,” he added.
Justin Robinson, cyber CTO at ThunderCat, said agencies should begin with narrowly defined use cases where organizations fully understand the underlying data, identity management, and security controls.
“If you try to boil the ocean, it’s just too much, and it’s an ever-changing field,” Robinson said. “It can be daunting.”
Robinson also warned that employees are adopting AI tools regardless of whether formal governance policies are fully in place.
“The next big challenge when it comes to AI is that our users aren’t waiting for permission,” he said. “They’re running to the tool that can help them do their job.”
