I’ve been thinking a lot about workflows.
We can assume the models will keep getting better. But how do we best incorporate them into the school day, the teacher workflow, so that teachers and students can make the best use of the powerful capabilities?
In software development, this shift is well underway. We have seen a move from copy/pasting in ChatGPT to embedding in an IDE like Cursor. Now, developers commonly outsource entire features or projects to standalone agents using Codex, Claude Code, or even an OpenClaw agent.
But in education, we’re still stuck. Earlier this week, I wrote:
AI in schools has reached a plateau. The majority of teachers have used AI to plan lessons, give feedback, and differentiate their curriculum, but are still stuck in a cycle of opening tabs, typing in prompts, and copy/pasting the results.
With the proliferation of tools, we have seen this is only getting worse.
I think we should look to the teachers for the solution.
—
Some of the best apps in education are built by teachers, for teachers. I had the pleasure of speaking recently with the co-founder of the popular teacher AI tool Eduaide.
Thomas Thompson was a middle school social studies teacher in Maryland. During COVID, he and a colleague started building Eduaide, now used by over 800,000 educators. Thomas identified the same issue:
The way most of these tools are working — ourselves, MagicSchool, DiffIt, Brisk, SchoolAI — what you do is you have a directory of 120 or so different prompts. You click on it, preloads a prompt, then the teacher fills in some information, that gets injected into the prompt and the prompt goes to the AI. You get a response, typically in like a chat format. And then the teacher will copy that, take it to Google Docs and, you know, put some flourishes on it, make it their own.
His team recently shipped their solution to this problem: a unified workspace that includes all the relevant context for a given lesson. The AI reads your document, understands context, and lets you revise in place. It’s the same shift we’re seeing in developer tools, from ChatGPT to Claude Code to OpenClaw, applied to education.
Here he shows how the new workspace works for teachers:
I talked with Thomas about his approach to product development, how they use evaluations to ensure that the tools do what they should, and how to apply pedagogy with myriad choices in AI tools. He gives us a sneak peek into where the future of AI education tools is going.
Listen to the full conversation here.
Try the tool: Eduaide.ai
