What Agentic Learning Actually Looks Like in Schools — And How One Superintendent Is Making It Real

For district leaders trying to figure out how AI fits into their schools, the conversation often gets stuck in the abstract. Is AI safe? Will it replace teachers? How do we even begin?

‍

Dr. Jared Bloom, Superintendent of Franklin Square Schools in New York, has been navigating these questions head-on — and his approach offers something more useful than a framework on a slide: a set of decisions that are already working in real classrooms.

‍

Agentic learning isn't a buzzword. It's a shift in what's possible.

‍

Bloom has been in education for 25 years. He remembers what differentiated instruction used to cost a teacher in time: "If I wanted to have different levels in my classroom, I'd have to research for a week, different passages that I would pull together and jigsaw with my kids." Today, he points out, a teacher can do that in seconds.

‍

That's what makes agentic learning meaningful right now — not the idea of it, but the access. AI tools for reading and learning can now meet students at their actual level, adjust as they grow, and free teachers to do the work that requires a human in the room: small group instruction, relationship-building, intervention, and enrichment.

‍

"We really do have to meet students where they're at and teach them for their tomorrow, not ours," Bloom said on the Champion Every Voice podcast.

‍

For district leaders, that's not a platitude. It's a purchasing and implementation lens. The question isn't whether your district has AI tools — it's whether those tools are actually producing agentic learning, or just digitizing the same one-size-fits-all instruction you already had.

‍

The traffic light system: a practical approach to AI use in classrooms

‍

One of the most concrete things Bloom shared was how Franklin Square is setting expectations for when and how students use AI in their work.

‍

Teachers communicate assignments using a red, yellow, and green system:

‍

- Green: Students can use AI to explore, draft, and create.

- Yellow: AI can support brainstorming and idea generation, but students complete the final work themselves.

- Red: Students work without AI — the goal is deep thinking that the student has to do independently.

‍

The logic behind this isn't restriction for its own sake. It's preparation. As Bloom put it, if you write a book and send it to an editor, the manuscript comes back different — and that's legitimate. AI can serve that editorial function. But students need to understand when to use it and when not to. That judgment is itself a future-ready skill.

‍

For leaders rolling out AI in education, this framework is worth adapting. It gives teachers clear ownership over how AI tools are used in their classrooms, and it teaches students to think about AI as a tool with appropriate uses — not a shortcut or a threat.

‍

Trust is infrastructure, not a talking point

‍

The piece that makes AI integration in schools succeed or fail, in Bloom's experience, is trust — and it has to be built before rollout, not explained after the fact.

‍

That means involving teachers in the decision-making process, not just informing them. It means understanding what their fears are alongside their excitement. It means student voice too: Bloom has a Superintendent's Student Advisory Group of fourth through sixth graders who advise on curriculum, and they stay with him for three years. The students his district is serving are the ones best positioned to tell him what's working and what isn't.

‍

"Top-down does not work," he said plainly. "You have to figure out how to create organic buy-in on so many different levels."

‍

For district leaders, that's a concrete implementation checklist before any AI reading program or adaptive learning tool goes live: Who was in the room when this decision was made? Do teachers feel supported to explore and fail and adjust? Do students and families understand the why?

‍

What Bloom saw at BETT — and why it matters for US districts

‍

Bloom recently attended BETT UK, the largest ed tech conference in the world. What struck him wasn't the scale of the booths or the technology on display — it was how familiar the challenges sounded. Whether in England or the U.S., educators are asking the same questions about AI in education: How do we use it meaningfully? How do we keep students safe? How do we preserve the human side of learning?

‍

He also noted that schools are landing in very different places. Some are scaling back technology use and being much more intentional about when screens are out. Others are going all in on AI. The divergence is real — and it suggests there's no universal answer, only the answer that's right for your students, your community, and your data.

‍

The bottom line for district leaders

‍

Agentic learning and AI tools for educators are not self-implementing. What Bloom's experience shows is that the districts making progress are the ones treating AI integration the same way they'd treat any other instructional initiative: with clear goals, ongoing training, a feedback loop, and a willingness to stop doing something if the data says it isn't moving students.

‍

"If you're not actually closing that achievement gap for individual students who are struggling... what are we doing this for?" Bloom said.

‍

That question applies to every AI reading assistant, every adaptive reading software purchase, and every pilot a district runs. The technology is only as good as the implementation — and the implementation is only as good as the trust you built before you started.

‍

Want to hear the full conversation? Listen to Dr. Jared Bloom on the Champion Every Voice podcast.

🎧 https://rss.com/podcasts/champion-every-voice/

‍