Most school districts are using AI cautiously, usually to save teachers and administrators time. But a small group of districts is doing something different: using it to rebuild how school works.
CRPE’s early 2026 study of AI Early Adopters sorted districts into five categories, from Dabblers to Reimaginers. This brief follows up with four districts from the most ambitious middle tier—System Changers, where AI use is amplifying an existing reform effort aimed at changing what schooling looks like. In Agua Fria, Arizona, custom AI tools help teachers connect academic standards to each student’s career pathway. In Anaheim, California, an AI-powered learning system tracks student progress in skills like collaboration and critical thinking. At ASU Prep, internal teams are building their own tutoring tools and planning aids. And in Elma, Washington, a custom AI tool keeps instruction aligned to what local employers actually need.
Key findings
These four System Changers:
- Lean on a layered network of external partnerships while simultaneously building their own internal technical and instructional capacity and knowledge.
- Act as R&D sites that adapt, customize, and build AI tools rather than just purchasing them.
- Maintain partnerships with employers to keep AI-informed learning aligned with changing workforce needs.
- See early signs that AI is accelerating their existing reform efforts, but so far have little evidence that AI is improving student outcomes.
- Continuously navigate tensions between leaders’ visions for how AI can help create transformational change and educators’ day-to-day experiences with implementation.
The gap between these four districts and the rest of the field isn’t access to AI tools, it’s technical fluency. That fluency lets a district recognize when a tool won’t fit its instructional vision, and build one that will. Right now, these districts are largely building that capacity alone.
Read the full brief for:
- How each district structured its network of instructional and technical partners
- What acting as an R&D looks like in practice, from custom LMS builds to AI tutoring tools
- Where leaders and teachers diverge on what AI adoption feels like day to day
- What these districts say they need from states and intermediaries to go further
These districts are setting a faster pace than the rest of the field, but they’re doing it without the infrastructure to support them. The question for policymakers, funders, and support organizations is whether that infrastructure gets built, or whether the most ambitious districts keep carrying the load alone.
