Mississippi CIO Craig Orgeron comes off an optimist about artificial intelligence, a leader with a can-do attitude about a technology that sparks nearly utopian hopes and dystopian fears that range from lost jobs to software-induced bias.
Such optimism is hardly uncommon in the government technology industry as AI takes on more work and becomes familiar.
Part of it, no doubt, is just “lead, follow or get out of the way” pragmatism, as it seems increasingly unlikely to gov tech leaders that AI is some hyped-up, over-marketed trend destined to fade out.
Even as state and local agency tech leaders spent much of 2025 worried about the impact of federal budget cuts, AI development and deployment charged ahead. A National Association of State Chief Information Officers (NASCIO) survey, for instance, found that 82 percent of state CIOs reported that their employees are using generative AI tools in daily work, with 90 percent crafting pilot projects with AI.
AI “is democratizing tech in a generational way we’ve never seen,” Orgeron told Government Technology recently.
In his view, AI will lead to more “augmentation” in public-sector tech instead of becoming a zero-sum competitor. AI will become a teammate, not a threat, even as the technology forces significant organizational and other disruption.
Getting from here to there, however, will take some hard and well-thought-out work, and potential hurdles have emerged. The upcoming year promises to shed meaningful light on how and where AI will grow beyond public-sector pilots and whether forces outside of gov tech will slow progress.
Last year showed how artificial intelligence can fuel big-time investment in government technology. Public safety attracted much of the fresh capital in what proved to be a robust time for the industry, but areas such as procurement, compliance and permitting also benefited, and the money shows no signs of slowing down.
But 2026 will bring more pressure on AI — from financial backers, government officials and skeptical residents — to produce results. That could mean money or time saved, or reduced stress on public agency workforces, or stronger ties between community members and their elected representatives.
The general goal is to move past AI experiments and one-offs and into permanent projects that can scale and offer solid proof that artificial intelligence improves government operations.
“We’re seeing AI-driven improvements in areas like traffic optimization, fraud detection and citizen service automation,” said Will Weatherford, a former Florida House of Representatives speaker who is now managing partner of Weatherford Capital. “However, many of these are still pilots or narrowly scoped deployments.”
BUBBLE WORRIES
The pursuit of solid and sustained AI gains in the public sector comes as the technology faces its “bubble” stage — that is, worries that investment in AI is running wild, and that the tech will ultimately succumb to the gravity of financial reality.
As 2025 ended, speculation about an AI bubble was rampant, with those insisting that “trillion-dollar bets” on the technology might not pay off mixing with people who say large language models, not the whole of AI, present more risk.
Other observers pointed to AI’s massive hunger for energy as perhaps the biggest weak spot in the tech’s growth, with still others poking holes in or dismissing the bubble notion.
An AI bubble could knock some suppliers out of the market, change product offerings, shift the flow of investment and cause other headaches just as the fledgling technology is poised to gain genuine wings in the public sector.
AI agents, specifically voice AI, is the great equalizer. This is the means by which both digitally native and non-digitally native residents can both use software to interact with their government.
The lack of anxiety probably has something to do with putting on a good face as AI grows, but it also accounts for the relative financial stability of the gov tech market, the ongoing high demand for better tech from public agencies, and the memory of the dot-com bubble of the late 1990s and very early 2000s.
When that bubble burst, it took down numerous companies and ruined multiple investments but didn’t destroy e-commerce and related industries. In fact, those areas of the economy, cleared of low-performing operations and other inefficiencies, soon achieved a fresh level of growth that led directly to our current cloud-based, mobile, quick-delivery era.
That’s the theory, at least — one advanced during earlier bubbles involving railroads, radios, TVs and even home computers and video games during the 1980s. That’s part of the foundation for AI optimism in the public sector going into 2026: Bubbles come and go, but worthwhile technology survives and becomes ever more widely used.
“Generally, I don't foresee a huge impact of the potential AI bubble,” said Eyal Feder-Levy, CEO and co-founder of Zencity, which sells community engagement software and has recently been “doubling down on AI.”
If the supposed AI bubble does burst, suppliers and public agencies might feel pain. That’s because, Feder-Levy said, many of those suppliers are subsidizing the cost for government users — good news in this era of reduced funding and other economic uncertainties. But good things don’t always last.
“If and when the bubble crashes and the economics look different,” he explained, “public-sector customers might find themselves paying a lot more.”
HYPE AND ROI
Like so many technologies, AI is going “through a hype cycle before maturing,” according to Weatherford, the investor.
But gov tech offers shelter from that potential storm because adoption is “typically more deliberate and mission-driven, which should temper the speculative excess we see in other sectors,” he said.
No matter the relative stability of gov tech — a common theme offered by investors when asked why they continue to park private equity and other forms of investment into this space — those backers will eventually demand solid returns.
A report circulating in the later half of 2025 highlighted the need for ROI, albeit in a broad way that doesn’t focus on the public sector.
The Massachusetts Institute of Technology (MIT) found that despite up to $40 billion worth of investment for generative AI so far, 95 percent of organizations using the tech “are getting zero return” even as wealthy stockholders enjoy the rewards of all that AI investment.
Not only that, but just 5 percent of integrated AI pilots have returned meaningful value.
MIT said that “adoption is high, but disruption is low,” with many pilots but “very few [that] reach deployment. Only a small fraction of organizations have moved beyond experimentation to achieve meaningful business transformation.”
Governments don’t exist to make profits, of course. But elected and appointed officials are on the hook for how they spend taxpayer dollars, and they have to justify their AI expenditures to voters who are often deeply skeptical about the technology.
“The public souring on technology is always possible and a lot of the trust rests on individuals building and deploying that technology to take it seriously,” Kyle Patel, vice president of engineering at Polimorphic, told Government Technology via email. “From what I’ve seen to date, on both the vendor and government sides, great strides are being taken to ensure this technology is rolled out responsibly.”
Adobe Stock
A PATH TOWARD SCALE
New York-based Polimorphic offers an example of how ROI could be measured for public-sector AI in the coming years — a way for it to scale and give tangible daily benefits to all involved.
The company sells artificial intelligence-backed chatbots and search tools, voice AI for calls, constituent relationship management and workflow software, and permitting and licensing tech. In 2025, it raised $18.6 million in a Series A funding round.
Time and time again, the more that people get hands-on with [AI], the fear drops. The hesitancy drops. And so I think it’s real important that we continue to talk about AI literacy, education, what it means.
“AI agents, specifically voice AI, is the great equalizer,” he said. “This is the means by which both digitally native and non-digitally native residents can both use software to interact with their government. Using AI for customer service will also open up a much greater view into how residents engage with their government,” which in turn can lead to ever more data-driven insights for public officials.
Anti-fraud efforts also appear likely to drive growth in public-sector AI in 2026 and help it to scale.
For instance, federal cuts and changes to Supplemental Nutrition Assistance Program (SNAP) benefits — once known as food stamps — will “encourage state food assistance [SNAP] programs to accelerate the use of advanced analytics and AI to transform program operations and increase accuracy,” according to John Maynard, a principal solutions architect for SAS, a data analytics and AI company.
Public health, taxes, identity verification, human resources, law enforcement and emergency dispatch — those are areas cited by gov tech experts on the supply and demand sides when asked where AI is likely to scale in the near future. Such areas typically account for many of the core functions of government, and that’s where public-sector AI probably has the biggest chance to prove itself and shine in these early days of the technology.
And anything that helps public agencies improve workflows also gets high marks when it comes to that potential.
THE APPEAL OF AUTOMATION
Consider this: A study released by Accenture in late 2025 found that 62 percent of government workers think AI will reduce their workloads via the automation of “routine tasks.” Still, 86 percent of residents said they want to retain access to humans even if there are digital options, especially for “complex” civic issues. That suggests the hype around AI has yet to break through some of the public skepticism that remains a barrier to scale. But it also indicates opportunity.
“AI is most likely to scale first in areas where it can automate high-volume, repetitive tasks and improve service accessibility, such as permitting and customer service,” Eyal Darmon, a managing director at Accenture, told Government Technology via email. “The research finds that initial AI deployments in state government have focused on employee-facing applications, such as automating administrative backlogs and streamlining internal processes. As confidence grows, the focus is shifting toward resident-facing solutions.”
Identifying the targets for scaling public-sector AI and getting to that point are two different things, though. And right now, it can seem more art than science in achieving scale, though more refined guidance is becoming available as more people get used to artificial intelligence — and as more public officials absorb seemingly countless sales pitches about the wonder and magic of the fast-moving technology.
AUDIENCE KNOWLEDGE
For one thing, suppliers and public agency backers of AI need to follow a rule that dates back to ancient times, when solar eclipses and tea leaves offered state-of-the-art data analysis: Know your audience.
One size doesn’t always fit all, especially given the differences in revenue and tax bases among local and state governments — to say nothing of the political differences via which certain tools — including AI-backed police technology — could be lauded or condemned.
For instance, a public-sector AI pilot program should answer the question of “how is this solution going to maximize our services to residents, but also save us money over the long term?” according to comments from Christopher Rodriguez, assistant city administrator for Washington, D.C., at the second annual GovAI Coalition Summit* in November.
Smaller towns might need their own specific reason to test and eventually deploy AI, according to Nichole Sterling, mayor pro tem and Board of Trustees member in Nederland, Colo., and founder of the company My Town AI, who spoke at the same conference.
After all, such places are often “under-resourced,” and a successful AI pitch might highlight its ability to improve operations.
Sterling helps lead a town of 1,500 residents. Conversations with constituents are vital when trying to advance the cause of AI and work toward scaling the technology, even if some people fear it.
“If it’s a doomsday thing, then yeah, it’s going to be a doomsday conversation,” she said. “Time and time again, the more that people get hands-on [with AI], the fear drops. The hesitancy drops. And so I think it’s real important that we continue to talk about AI literacy, education, what it means.”
Data hygiene also stands as a vital precursor to scaling public-sector AI. Starting with even a small amount of curated data can lead to bigger results, Sterling said.
AI is nothing without data, she noted, and many cities “don’t have a good data hygiene, or a good data framework that we’re operating from.”
In the early days of this new year, so many AI efforts remain pilots or glorified experiments, or even reactions to the hype. Bubble or not, however, public-sector artificial intelligence stands ready for a burst of growth — assuming it’s done right. Scale takes thought and skill.
“We’re still in the early innings on AI integration in gov tech,” said Weatherford. “Early returns have been modest but encouraging. The real gains — measured in productivity, cost savings and safety outcomes — will emerge as AI tools become embedded across systems and workflows rather than added on top of them.”
Senior Staff Writer Skip Descant contributed reporting to this story.
*The GovAI Coalition Summit is hosted by Government Technology in partnership with the GovAI Coalition and the city of San Jose.
