Here’s a moment I think a lot of teachers will recognize: it’s 6:30 a.m., you’re standing in front of your coffee maker, and you’re already mentally rewriting your third-period lesson because you realized last night that what you planned just isn’t going to land. Twenty minutes. You have twenty minutes to fix it.
That used to mean a frantic Google session and a lot of hoping for the best. Now I open MagicSchool, type a quick prompt, and I have a revised activity, a differentiated version for my students who need more support, and a higher-order thinking extension. All before the coffee finishes brewing.
That’s what AI in education actually looks like for me on a Tuesday morning. Not robots grading papers (though with the right frameworks and guardrails, it can help a lot). Not some dystopian replacement for real teaching. Just a tool that gives me back time I didn’t have.
This series of musings on AI in Education exists because I think a lot of teachers, parents, and administrators are either overwhelmed by the AI conversation or totally checked out of it. I’m not calling anyone out, but we all know that neither of those situations is a great place to be. So let’s talk about it like real people.
What We Mean When We Say “AI in Education”
Let’s get something out of the way first: AI is not one thing. It’s a broad category of technology that includes everything from the recommendation algorithm that serves up your next Netflix binge to the large language models (LLMs) that power tools like Claude, ChatGPT, and MagicSchool.
For our purposes (schools, classrooms, teachers, kids), we’re mostly talking about generative AI: tools that can write, respond, summarize, explain, create, and converse. These are the tools showing up in teacher workflows right now, and they’re genuinely changing how some of us do our jobs.
AI doesn’t replace what makes a great teacher. But it can clear away the noise so great teachers can actually teach.
The key thing to understand is that generative AI is a tool with a very wide range of use cases. It can help a teacher write a quiz in five minutes, help a struggling student get a patient explanation of a confusing concept, or help a principal draft a parent communication at 9 p.m. without sounding like it was written by a committee. All of those are real use cases happening in real schools right now.
The Tools I Actually Use (And Why)
I want to be concrete here, because I think the AI conversation in education stays too abstract for too long. So here are the tools I’m using, what they do, and why they’ve stuck. I also want to add that I am not paid by any of these companies and I am not endorsing them over any other tools. You’ll need to work with the tools that are 1) available and 2) sanctioned for use.
Let me give you one example before we get into the tools. I was teaching a unit on European conquest in the Americas. My students were reading excerpts from The Broken Spears, which is a collection of indigenous Aztec accounts of the Spanish conquest. It’s one of the most powerful primary source texts you can put in front of a high school student.
I asked Flintk12 to help me turn one of those accounts into a “choose your own adventure” activity where students could step into the perspective of an Aztec official and make decisions at key moments in the encounter with Cortés. What would have taken me hours to build became a finished, differentiated activity in about 20 minutes. My students were more engaged than they’d been in weeks. That’s the version of AI in education I want to talk about.
MagicSchool AI
MagicSchool is built specifically for educators, which matters more than it sounds. It has over 60 purpose-built AI tools: lesson plan generators, rubric builders, email drafters, IEP support tools, and more. The fact that it was designed with teachers in mind means the outputs feel like something a teacher would actually write, not a technical manual.
I use it most for differentiation. Quickly generating tiered versions of an assignment so every student has an entry point that works for them. What used to take an extra hour of planning now takes about five minutes.
But honestly, the most valuable thing MagicSchool does is put specific AI tools directly in students’ hands while keeping the teacher in the loop. You can introduce a particular tool for a specific activity or project, and then monitor how each student is actually using it in real time. You’re not just hoping students are on task. You can see it. That changes the dynamic of AI in the classroom entirely: instead of AI being something students sneak, it becomes something you teach them to use, with visibility and accountability built in.
Flintk12
Flintk12 is not new to my workflow, but it’s essential. It’s an AI platform built around student data and school operations. Think smarter ways to understand student progress, flag patterns early, and support teachers with actionable insights rather than just more dashboards to click through.
One specific place where I’ve found it genuinely valuable is DBQs (Document-Based Questions) in my AP World History classes. DBQs are a core part of the AP history exam: students have to analyze multiple primary sources, build an argument, and demonstrate skills like sourcing, contextualization, and complexity. They’re hard to teach and hard to give good feedback on at scale. Flintk12 helps me track which specific skills individual students are struggling with, so instead of writing the same generic comment on 30 papers, I can see that one student needs work on sourcing while another keeps missing the complexity point. That kind of targeted insight makes my feedback actually useful.
Something else worth noting: Flintk12 also lets teachers see in real time how individual students are engaging with the tool during class. That’s genuinely useful for keeping a room focused and on-task without having to hover.
One feature I’d love to see them build: a direct messaging thread from teacher to student inside the platform. There are moments in class when a student needs a quiet nudge, and right now your options are to walk over (which can draw attention) or let it go. A small, low-key chat window where I could send a private note would solve that elegantly.
Granola
Granola is my AI meeting note-taker. It runs quietly in the background during meetings (department check-ins, parent conferences, PLC sessions) and produces clear, organized notes automatically. I can actually be present in a conversation instead of frantically scribbling while also trying to listen.
For teachers who sit in a lot of meetings — which is all of us — this one is a quiet game-changer. You end the meeting with a record of what was actually decided, not what you vaguely remember being said as you were walking back to your classroom.
Claude Cowork and Claude Code
These two are from Anthropic, the company behind Claude, one of the leading large language models. Claude Cowork is a desktop tool that lets you automate file and task management without needing to know how to code. Think of it as having an AI that can help you organize files, process documents, create materials, and manage workflows from a simple conversation.
Claude Code is the more technical sibling. It’s a command-line tool for developers, which means it’s more relevant for tech-savvy educators or school IT staff who want to automate deeper workflows or build custom tools for their school.
What I use Claude Cowork for, practically, is generating content like this, creating structured documents quickly, and handling tasks that would otherwise eat up my prep time. But I’m also getting more comfortable with the coding side of it. For instance, I’m currently turning a Build-Up to WWI activity into an interactive game that students will be able to play online in class. I’m adding follow-up questions and SAQs from the activity to tie everything together. We’ll see how it turns out, but so far, I’m super excited.
What This Means for Parents
If you’re a parent reading this, you might be wondering: Is AI helping my kid, or is it just helping them skip the thinking?
That’s the right question to ask. And the honest answer is: it depends entirely on how it’s used.
Used well, AI can give your student something teachers rarely have time to offer: personalized, patient, on-demand explanations. A student who doesn’t understand a concept can ask an AI tutor to explain it five different ways until one sticks, at 10 p.m., without waking up a parent or waiting until the next school day.
Used poorly, AI becomes a shortcut that skips the learning entirely. A student who uses AI to generate an essay they never thought through hasn’t learned anything. They’ve just gotten a grade they didn’t earn.
Schools are actively working through how to handle this. Most of us aren’t trying to ban AI. We’re trying to teach students how to use it responsibly, the same way we teach them to evaluate sources online or cite their research. Digital literacy has always been part of the job. AI literacy is just the newest chapter of that work.
What This Means for Administrators
For administrators, the AI conversation is happening whether you lead it or not. Students are already using it. Many of your teachers are already using it, and the ones who aren’t are going to have questions.
The schools that are going to come out ahead on this aren’t the ones that locked everything down, and they’re not the ones that let it be the Wild West either. They’re the ones where leadership got curious, built some guardrails, and created space for teachers to actually learn and experiment with these tools in a supported way.
That means professional development that doesn’t feel like a compliance checkbox. It means giving teachers access to tools like MagicSchool or Flintk12 and letting them tell you what’s working. It means developing a clear AI use policy that students and parents actually understand — one that talks about responsible use, not just consequences.
The schools winning with AI aren’t the ones with the most tools. They’re the ones with the most thoughtful people asking the right questions.
And honestly, it means being willing to learn alongside your staff. The administrators who earn the most trust on this issue are the ones who show up to the AI professional development session and actually participate — not just introduce the speaker and slip out the back.
Try It This Week
Every issue includes one practical move for two kinds of people: those who are just getting started with AI, and those who are already using it and want to go a little further.
Start Here: Sign up for MagicSchool AI and run your first prompt. Go to magicschool.ai and create a free account. Then open the Lesson Plan Generator, type in your subject, grade level, and a topic you’re teaching this week, and see what it produces. You don’t have to use it — just get a feel for how it works. The whole thing takes five minutes, and you’ll immediately understand what the AI conversation is actually about.
Level Up: Use Granola to capture your next meeting and share the notes. If you’re already comfortable with AI tools, try adding Granola to your next department or PLC meeting. Let it run quietly in the background while you stay focused on the conversation. After the meeting, review the notes it produced and share them with your team. It’s a low-risk way to see how AI-assisted workflows actually function in a real school setting.
Important: Let people know that you’re using an AI note-taker. It does record the meeting, so being upfront with everyone involved will save some potentially difficult conversations in the future.
The Big Picture
AI in education is not a trend that’s going to pass. It’s also not the apocalypse. It’s a tool — a genuinely powerful one — that is going to keep changing what teaching and learning can look like.
What I care about, and what I think most educators care about, is whether that change makes things better for kids. So far, in my own classroom and in conversations with colleagues, the answer is: it can. If we’re intentional about it.
That’s what this is for. Every week I’ll try to share what I’m learning, what tools I’m using, what’s working, and what’s not. No hype. No panic. Just a real conversation about something that matters.
See you next Monday.
— David
Tools mentioned this issue:
– MagicSchool AI — AI tools built for educators (lesson planning, differentiation, rubrics)
– Flintk12 — Student data and school operations platform
– Granola — AI meeting notes that run quietly in the background
– Claude Cowork — AI desktop tool for file and task automation (no coding required)
– Claude Code — Developer-focused AI tool for deeper workflow automation
