There’s a Sunday night feeling most ESL teachers know well. The week starts in less than twelve hours. You have three classes, four levels, and a blank lesson plan document staring back at you. You love teaching — but the prep is slowly grinding you down.
That was me, not so long ago. And it’s what pushed me to try something I’d been putting off: letting AI handle my lesson planning for an entire month.
Not occasionally. Not as a backup. Every lesson, for 30 days straight.
Here’s the honest account of what happened, the good, the surprising, and the one mistake I had to correct halfway through.
Why I Was Skeptical at First
I want to be upfront: I went into this experiment as a sceptic. I’d tried AI tools before and come away with generic, surface-level output that felt like a lesson plan from a textbook nobody uses. The kind of thing that looks structured on paper but falls flat in an actual classroom.
My concern was simple. Teaching English is deeply human work. You read the room. You adapt. You notice when a student is confused, bored, or quietly having the best breakthrough moment of their week. No algorithm does that.
But lesson planning, the prep work before you walk in the door, that’s a different story. That’s a task that takes time, follows patterns, and often produces a similar result whether it takes you twenty minutes or two hours. That felt like a fair thing to test.
The Experiment: What I Actually Did
For thirty days, every lesson plan I wrote started with an AI prompt. I used ChatGPT for most of it, with some sessions using Claude. The rule was simple: AI generates the framework, I adapt and personalise it before class.
I wasn’t using AI to replace my judgment. I was using it to replace the blank page.
My standard prompt looked like this:
THE CORE PROMPT I USED:
“Create a 60-minute ESL lesson plan for [level] students on the topic of [topic]. Include a 5-minute warm-up, a 15-minute vocabulary section with 8 target words, a 20-minute grammar focus on [grammar point], a 15-minute communicative speaking activity, and a 5-minute wrap-up. Make it practical and engaging for adult learners.”
I’d swap out the level, topic, and grammar point for whatever my classes needed that week. From first prompt to usable plan: under two minutes.
Week One: Uncomfortable but Promising
The first week felt odd. There was a voice in the back of my head saying this was cheating somehow, that a “real” teacher writes their own lessons from scratch. I had to push past that.
What I found in week one was that the AI output was roughly 70% usable immediately, 25% needed adapting for my specific students, and about 5% I threw out entirely. That ratio was actually better than I expected.
The warm-ups were consistently strong. The vocabulary sections were solid. The grammar explanations occasionally defaulted to a level of complexity I had to simplify for my lower-level groups. But the core structure was there every time.
More importantly, I had time back. Sunday evenings went from two hours of planning to about forty minutes total, including the time spent personalising the AI output.
The Discovery That Changed How I Teach
About two weeks in, I started experimenting with a prompt I hadn’t thought to try before: differentiation.
THE DIFFERENTIATION PROMPT:
“Take this activity [paste activity] and create three versions: one for A1-A2 beginners, one for B1-B2 intermediate students, and one for C1 advanced learners. Keep the core task the same but adjust language complexity, scaffolding, and expected output for each level.”
If you teach mixed-ability groups, and most ESL teachers do at some point, this single prompt is worth the whole experiment. In under a minute, you have three-tiered versions of the same activity. No more rushing to simplify something during class while other students wait.
This is now part of my permanent teaching toolkit. I use it every week.
The Mistake I Made in Week Three
I need to be honest about where this went wrong, because it will save you time.
In week three, I got lazy. Instead of adapting the AI output for my students, I started using it almost verbatim. The lessons worked; they were technically sound, but they started to feel flat. My students noticed, even if they couldn’t articulate why.
The issue wasn’t the AI. The issue was me removing myself from the process.
The prompts give you a professional framework. Your job is still to bring the personal touches, the examples that connect to your students’ lives, the activities you know work with this particular group, and the flexibility to pivot when something isn’t landing. AI handles the architecture. You handle the soul of the lesson.
Once I rebalanced, using AI for the skeleton and my own experience for the detail, the quality jumped back up immediately. The experts call this the 30% rule: AI does the first 30%, and you take that architecture and build on it.
What 30 Days Taught Me
By the end of the month, I’d planned over forty lessons using this approach. Here’s what I know now that I didn’t know before:
- Lesson planning time dropped by around 60%. That’s not an estimate; I tracked it.
- Lesson quality stayed the same or improved because I had more mental energy left for actual teaching.
- The blank page problem is completely solved. I never stare at an empty document anymore.
- AI is strongest for structure and weakest for nuance. Always add your own student-specific details.
- The prompts you use matter enormously. Vague prompts get vague results. Specific prompts: level, topic, grammar focus, activity type, time; get professional output.
The Free Resource I Put Together
After this experiment, I compiled the 50 AI prompts I use most regularly into a free PDF guide. It covers lesson planning, grammar teaching, speaking activities, differentiation, student feedback, and more, with a real classroom example for each one.
It’s completely free. No catch. You can grab it below and start using the prompts this week.
FREE DOWNLOAD:
(You might have to check your spam folder at first. You will get a verification email and then the PDF guide in the following email immediately)
Ready to Go Further?
If you want classroom-ready activities that are already built out for you — no prompting required — my Ultimate ESL Activity Bundle includes 14 complete activities covering speaking, grammar, vocabulary, and more. Everything is designed to run straight from the page with minimal setup.
You can find the bundle at teachingenglishisfun.com or via the link below.
Thirty days of trusting AI with my lesson planning taught me one thing clearly: the best teachers aren’t the ones who do everything themselves. They’re the ones who know which tools to use and when to put them down.
Have you used AI for lesson planning? What’s worked for you? Drop a comment below — I’d love to know.
