Who are AI’s ideal classroom users?

You are receiving the community edition of today’s newsletter. Become a member to receive exclusive perks and event invites.

The main discussion about AI in schools focuses on children: should they use it, how often, and how it affects their thinking.

On Tuesday, two founders on our panel discussed their experiences with this debate. They both focus on using AI to assist teachers, which is a perspective that isn't often considered.

The teacher is AI’s first customer in a classroom

Carly Shuler, co-founder & CEO of the literacy tutoring company Hoot Reading, said current research points one way: “The most effective AI is a teacher co-pilot.” Her team’s AI shows a tutor where a child’s reading is breaking down and what to try next, and the tutor decides what happens in the lesson.

Rohan Mahimker, Prodigy's co-founder & co-CEO, made the same call. His team uses AI to produce lessons at scale, teachers sign off on them, and the kids rarely interact with the model.

It’s also what gets adopted. More than 10 US states now limit screens in schools, and Carly said her company has been carved out of those limits because there’s a live human on the other side.

Our read is that this part of the debate gets missed. A teacher makes many calls a day about what each kid needs next, and AI helps most when it gives the teacher better information for each one.

If you’re building for a classroom, the question worth answering is what the teacher needs to decide faster or better, and whether your product does that.

Not all screen time is bad

One of the loudest debates in education right now is screen time. Rohan judges it by what the screen is for, using three questions.

Is the kid active, like answering a math question, or passive, like watching a video?
Is there a learning objective?
Has an independent study shown it works?

A screen that passes all three counts as good screen time to him. “It is a disservice not to lean in,” he said.

Active, purposeful, and proven is a fair bar for any product that claims to help someone get better at something, including whatever you’re building with AI inside it.

Everest uses AI to learn faster

The evening opened with 11-year-old Everest Nevraumont, who already learns to that standard.

Four gold medals at the National History Bee. Behind that achievement? Nine months spent figuring out how to learn better. She now uses a combination of direct instruction, spaced repetition, and memory palaces (a technique dating back to ancient Rome) to learn faster and retain more.

AI does two jobs in her day. It picks what she should work on next, and when she finishes an essay, it tells her within moments what to fix, where a teacher’s comments might take a week. The AI removes the waiting, and the work stays hers.

If you’re building for students, or for anyone learning something hard, the slow feedback loop is the blocker worth going after, since the struggle in between is where the learning happens.

💛 We’d love to hear your TechTO story

Twelve years in, the best part of TechTO is still the people.

If this community has played a part in your journey, whether it’s finding your co-founder, landing a job, or meeting your investors, tell us about it in a sentence or two.

Your words might end up in the spotlight as we celebrate on November 17.

Share your story with us