New Series on The Government AI Journey: State Lessons from Readiness to Public Value

In Code for America’s 2026 Government AI Landscape Assessment we made the case of an institutional journey across four stages: Readiness, Piloting, Implementation, and Impact.  I'm grateful to the Code for America team for trusting me to lead the research and analysis on this important work. This new analysis expands on last year's AI readiness rubric to include the lifecycle of AI development, deployment, and measuring impact with the ongoing learning necessary for how quickly AI evolves. 

The public report summarizes this progression, while the full research report adds the state-by-state evaluation showing where each state sits across the four stages. The analysis shows real gains in AI readiness leading to a lot experimenting and piloting in state government. Mirroring (and perhaps bettering) the private sector, fewer AI projects have moved to production with impact gains clearly articulated. 

As the technology rapidly evolves, I would anticipate this to continue to be a bit of a moving target. For example, some states may be strong at getting LLMs into their hands of their employees to support productivity gains, but may not yet be doing much with AI agents.

In this series, I will take a deeper diver into the findings of the report. I'll also take time to explain the rubric for each stage.  Here's the flow of the series:

1. The State AI Maturity Map
Use the full report’s state-by-state evaluation and distribution of ratings to frame the national landscape.

2. Readiness Is the Foundation
Feature Georgia, New York, Colorado, and Minnesota as examples of governance, executive leadership, infrastructure, and workforce capacity.

3. From Task Forces to Test Beds
Feature Connecticut, Utah, North Carolina, Arizona, New Jersey, and Pennsylvania as piloting and sandbox examples.

4. From Pilot to Production
Feature Maryland, Pennsylvania, Vermont, and North Carolina as implementation case studies, especially around benefits access, document processing, chat assistants, and fraud detection.

5: From AI Adoption to Public Value
Feature Vermont, North Carolina, Pennsylvania, Connecticut, New York, and Washington as examples of inventory-driven governance, fiscal impact reporting, and emerging accountability systems.

Check out the full report here.

Check out Part 1 in the Series.