ImpactHub.ai
AI Innovation Hub
Donoradvisor.ai 
The State AI Maturity Map: Who Is Leading, Who Is Learning, and What Comes Next
Across the country, state governments are no longer tiptoeing into AI adoption. They are leaning in much more aggressively to learn how best to govern it, what use cases are best to test it, which pilots lend themselves to scale it, and how to prove that it creates public value. This post shares a high level review of the Code for America’s 2026 Government AI Landscape Assessment, which Humanservices.ai was contracted to lead the researcha and analysis.
Methodology
The 20206 AI Landscape Assessment evaluates each state across four stages: Readiness, Piloting, Implementation, and Impact. The public report describes these as a progression: readiness builds the foundation, piloting demonstrates what is possible, implementation delivers results, and impact ensures accountability and improvement.
We developed this rubric to extend well beyond last year's assessment that focused solely on Readiness within the domains of leadership, infrastructure, and staff capacity building. So many states advanced enough AI initiatives in the last year that it made sense to capture the entire journey from readiness to making an impact with AI. The full research report turns that framework into a state-level evaluation. Within each stage states are evaluated on the following spectrum:
Early - Initial steps and exploration
Developing - Building core capabilities
Established - Systematic approaches through much of the state
Advanced - Sophisticated and innovative practices
The research consisted of publicly documented sources which includes: legislative session minutes, policies and announcements from state agencies and executive offices, and media reporting of state work. Code for America also provided the opportunity for each state to review the results and submit additional information that might have been missed in the landscape scan.
Summary of Results
The report demonstrated a tremendous amount of progress on readiness. Some states are advanced in readiness but still developing in implementation. Others are experimenting actively but lack strong public evidence of impact. A small group—Maryland, New Jersey, North Carolina, Pennsylvania, Texas, Utah, and Vermont—stand out as leaders because they are not merely adopting tools; they are building the institutional capabilities to deploy AI as a long-term public-sector asset.
That distributions tell the story:
The majority of states are now established or advanced in their AI readiness and only a few are still in the early stages of the journey.
States are moving fast on pilots and experimentation with nearly half at established or advanced stages.
Like what we are seeing in the private sector, states have moved more cautiously in moving from pilot to production deployments of AI.
Very few state can articulate impact of their AI projects and have yet to complete the learning cycles needed for continuous evolution with this fast moving technology.
State examples: different paths through the journey
A few examples show how varied the journey can be.
Maryland is rated Advanced in Readiness and Established in Piloting, Implementation, and Impact. Its strength is that governance, infrastructure, and benefits-focused implementation are beginning to reinforce each other. The public report highlights Maryland’s generative AI-powered benefits navigation and eligibility guidance agent, developed with AI providers including Anthropic, to help residents understand programs such as food assistance, Medicaid, and housing support.
New Jersey is Advanced in Readiness and Established across later stages. The state combines executive-level leadership, employee AI tools, and structured workforce training. The public report points to New Jersey’s use of employee-facing AI assistants and a broader AI task force approach as part of its leadership profile.
North Carolina is Advanced in Readiness and Established in Piloting, Implementation, and Impact. Its Government Data Analytics Center gives it a long-standing data and analytics foundation, which the report identifies as a key reason some states are better positioned to operationalize AI.
Pennsylvania is Advanced in Readiness and Established in the remaining stages. Its intelligent document processing work in the COMPASS benefits application system is one of the clearest benefits-access case studies in the public report, with AI checking uploaded documents for blurriness, quality, and relevance. This AI solution reduced illegible or incorrect documents by 80 percent and saved county assistance office staff more than 700 hours.
Utah is Advanced in Readiness and Piloting, and Established in Implementation and Impact. Its Office of Artificial Intelligence Policy and AI Learning Lab show how a state can build a regulated experimentation model for high-stakes use cases. The public report highlights Utah’s supervised AI medical prescription renewal pilot as an example of sandboxed oversight.
What separates leaders from the pack?
The full report identifies six lessons from leading states:
Leadership matters. States with engaged executive and administrative leadership are capturing value from AI.
Data and technical infrastructure is a major factor in scalability.
Workforce readiness is essential. States making training investments are seeing efficiency and productivity gains.
Government has a lower risk tolerance because AI can affect access to benefits and other essential services. Falling fast and failing forward is still a new set of practices for many governments, but with AI that's largely whats needed with pilots to get the learning for full scale implementation.
Structured experimentation accelerates learning, but it may not get agencies to full deployment. More emphasis needs to be placed on how to scale what works.
Measurement remains the next frontier.
The takeaway is not that every state needs to copy the same model. The takeaway is that states need to know where they are in the journey and what capability they need next.
Next in the series: Why Readiness is the most important stage—and how states like Georgia, New York, Colorado, and Minnesota are turning AI governance into an operating model.
