State CIOs Explore the Transformative Potential of AI Agents to Enhance Productivity and Efficiency
In a significant development, state Chief Information Officers (CIOs) are delving into the transformative capabilities of agentic artificial intelligence (AI), a frontier that promises to revolutionize government operations and citizen interactions. This exploration comes on the heels of extensive efforts over the past two decades aimed at reducing bureaucratic procedures within state governments. The hope is that the integration of agentic AI might finally bring these promises to fruition.
A recent report by the National Association of State CIOs (NASCIO) has shed light on the various ways states are beginning to explore the adoption of agentic AI across their agencies. The potential applications of such technology could initiate a fundamental transformation in the way governmental functions are performed and how citizens engage with state services. Agentic AI distinguishes itself from past generative artificial intelligence (AI) by its ability not only to generate content but also to orchestrate tasks and autonomously manage workflows.
According to Amy Glasscock, who serves as the program director for innovation and emerging issues at NASCIO, the key differentiator between generative AI and agentic AI lies in functionality. “Generative AI was generating content, writing things, summarizing things,” she elaborated. “Agentic AI is AI that can actually do things. It can take steps to move processes along.” This distinction underscores the shift from mere content creation to actionable intelligence capable of streamlining complex government operations.
State governments have increasingly adopted generative AI technologies, with nearly all currently running pilot projects aimed at harnessing these systems. Some of these pilot programs are now on the path toward broader implementation across various agencies. However, transitioning to agentic AI is a leap that demands higher sophistication and, crucially, a greater trust from both government employees and the public.
Despite enthusiasm surrounding these advancements, the current level of adoption remains relatively low. Glasscock noted that only about eight states have initiated projects utilizing agentic AI, indicating that the broader landscape is still in its infancy compared to the private sector, which has advanced further in this realm. States are taking a cautious approach, particularly when it comes to AI systems that could potentially interact with sensitive citizen data or influence government decision-making processes.
“It’s common for states to first test systems internally before deploying them externally for public services,” Glasscock said, emphasizing the vital importance of public trust and data security. The careful consideration stems from a desire to ensure that any automation serves to enhance citizen services without compromising data integrity.
As states continue to develop their technological frameworks, there exists the potential for remarkable improvements in the efficiency of high-volume administrative tasks characterized by extensive paperwork. AI agents could automate aspects of applying for state services, such as assisting citizens with form completion and document submission. This development may significantly accelerate response times from state agencies, allowing citizens to receive the assistance they need more swiftly. For instance, AI systems could review applications for completeness, identifying areas where documentation is lacking.
Among the states piloting agentic AI initiatives are Alaska, Tennessee, and Virginia. In Alaska, officials are considering the addition of agentic AI capabilities to the myAlaska portal. This integration would enable the autonomous completion of multi-step government transactions on behalf of users. Tennessee is exploring advanced enterprise resource planning systems featuring agentic AI, while Virginia is testing a tool designed to streamline the review process for regulatory legislation.
Particularly within government agencies dealing with substantial citizen interactions, there is an expectation that adoption of agentic AI will gain momentum. Human services agencies, followed closely by transportation and public health services, are anticipated to be early adopters of this technology.
Moreover, states are eyeing agentic AI as a strategic avenue to mitigate workforce shortages and address administrative backlogs plaguing various agencies. Glasscock pointed out, “A lot of work remains in state government that simply does not have the workforce to support it. Productivity improvements are one of the most significant advantages.”
Despite the promising potential for agentic AI to usher in increased productivity and efficiency, it simultaneously introduces a new set of security and governance challenges for technology teams. Given that agentic AI systems may interact with multiple state databases, any error or breach could have far-reaching implications. Glasscock warned, “There’s an additional layer of concern if hackers are able to gain control of these government agents and manipulate them.”
To navigate these complexities, she recommends that states begin with the National Institute of Standards and Technology’s (NIST) AI Risk Management Framework. Glasscock asserts that states are already considering this framework in the context of generative AI, emphasizing the importance of implementing robust governance structures. “AI is evolving rapidly, and if they don’t have governance in place already, it’s time to prioritize that process,” she stressed.
While agentic AI remains largely experimental within state government structures, there is a consensus among experts, including NASCIO, that dialogues surrounding its adoption will swiftly evolve. Looking ahead, Glasscock anticipates further developments. “It will be fascinating to see what progress we make in a year,” she remarked. “Expect to hear more about agentic AI in the coming months and years.”
In summary, while the path toward widespread implementation of agentic AI may be still emerging, the implications for state government efficiency and citizen engagement are promising and warrant keen attention as advancements unfold.