Educators brought challenges tied to their daily responsibilities: differentiating instruction, preparing lessons and materials, assessing work and providing feedback, communicating with families, and managing classroom data.

The intent was clear: solve for time, sustainable workflows, and student outcomes. Innovation theater was not the goal.

AI dexterity is knowing which tool, for which problem, in your context. It's the knowledge, practical skill, and professional judgment to decide whether, when, and how to use AI.

One corps member built a tool at a hackathon and reported no further AI use for three months. Another completed a practicum spanning three sessions and continued refining her tool independently. The difference was continuity.

Hackathons generate momentum, but most prototypes sit dormant 30 days later - not because the tool wasn't useful, but because there was no standing meeting to iterate, no peer group to troubleshoot with.

Nine Teach For America regions completed the first year of the AI train-the-trainer model, with average participant satisfaction of 4.75, and are now running their own AI hackathons.

AI dexterity is hard to develop through passive exposure alone. Hackathons, prototype shops, and regional convenings gave participants time to pose real problems, test outputs, encounter failure, and receive feedback.

In FY26, regional AI Leads delivered nine additional hackathons beyond the original plan. Seven of nine regions also reported preparedness to lead their own trainings.

Educators are learning techniques for using AI to reduce preparation burden, where they also retain responsibility for review, adaptation and instructional judgment. Time savings are real - but we need to think bigger about the quality of what reaches our young people.

A differentiated essay scaffold that "would have taken about 30 minutes or more on my own" reportedly took less than five minutes. But does it preserve the intended challenge? These questions matter.