Introduction
For twenty years, Sunday afternoon meant the same thing in my house: a blank lesson plan, a cold cup of coffee, and the quiet dread of knowing I had five different classes to prepare for the next morning.
Five classes. Five different levels. Different ages, different energy, different needs. Teaching ESL in South Korea was one of the most rewarding things I’ve ever done — and also one of the most relentlessly demanding in terms of preparation time.
I used to spend three to four hours every single Sunday just on lesson plans. Before marking. Before materials. Before anything else.
That changed about a year ago. Now the same five lessons take me under twenty minutes. Here’s exactly how, and why it works even better than the old way.
The Old Way Was Costing Me More Than Time
The problem with spending three hours on lesson planning every week isn’t just the time — it’s the mental energy. By the time I’d finished planning on Sunday, I was already tired before the week began.
Teachers know this feeling. You’ve thought about the lesson so many times in the planning phase that delivering it almost feels like going through the motions. The creative energy that should go into actually teaching — noticing which students are confused, adjusting on the fly, making those unexpected connections that make a lesson memorable — is already depleted by the time you walk into the room.
I wasn’t looking for a shortcut. I was looking for a way to preserve the energy I needed for the part of teaching that actually matters.
What AI Actually Does (and Doesn’t Do) in Lesson Planning
Before I share my process, I want to be honest about what AI is and isn’t good for here.
AI is excellent at generating structure. Give it a level, a topic, and a time frame, and it will produce a solid lesson framework in under a minute — objectives, warm-up, main activity, speaking practice, wrap-up. The bones of a good lesson.
What it doesn’t do is teach. It doesn’t know that one particular student needs a visual support, or that your Thursday afternoon class loses focus after thirty minutes, or that a cultural reference that works in one context lands completely flat in another. That knowledge is yours. It comes from years in a classroom, and no AI has it.
The workflow that works is this: let AI handle the scaffolding, and put your experience and judgment into everything on top of it. You spend your planning time making decisions, not filling in templates.
My Exact 20-Minute Process
Here is the prompt I use most often — it’s Prompt #2 in my free guide, and it’s the one I reach for every week:
“Create a 45-minute ESL lesson plan for [level] students on the topic ‘[topic]’. Include: learning objectives, warm-up (5 min), main activity (25 min), speaking practice (10 min), and wrap-up/homework (5 min). Include teacher notes.”
I paste that into ChatGPT, Claude, or Grok with my specific class details, and I have a complete lesson plan in about ninety seconds.
Then I spend the remaining time — usually ten to fifteen minutes — reviewing it with my teacher’s eye. Does the warm-up actually suit this class? Is the vocabulary level right? Does the speaking activity work for a group this size? Is there anything I know about these specific students that should change the approach?
That review step is where twenty years of classroom experience actually goes. The AI gave me a solid starting point instead of a blank page. I turned it into something that fits my actual students.
Multiply that by five lessons, and I’m done in under twenty minutes. The rest of Sunday is mine.
The Results After a Year of Using This Approach
The quality of my lessons didn’t go down. If anything, it went up — because I was walking into the classroom with more energy and a clearer head than I’d had in years.
I also noticed something unexpected: having a solid AI-generated framework to react to made me more creative, not less. Instead of building a lesson from scratch, I was editing and improving one. That’s a completely different cognitive task, and it turns out it’s a much lighter one.
The comparison I keep coming back to is cooking. A recipe doesn’t make you a worse cook. It handles the parts that don’t require your judgment so you can focus on the parts that do.
How to Get Started (Without Wasting Time Figuring Out the Prompts)
The biggest obstacle most teachers hit with AI lesson planning is the first one: writing prompts that actually produce useful output.
“Write me an ESL lesson” gives you something generic and borderline useless. “Create a 45-minute lesson plan for intermediate adult ESL students on the topic of job interviews, with a warm-up focused on useful phrases and a main speaking activity using role-play” gives you something you can actually teach.
The difference is specificity — and learning to write specific prompts takes time most teachers don’t have.
That’s why I put together a free guide: 50 ready-to-use AI prompts for ESL teachers, organised by category. Every prompt is already written in the specific format that gets good results. You just fill in your level and topic.
It covers lesson planning, speaking activities, vocabulary and grammar, classroom management, assessment, and customising activities for different levels and ages. Everything I use regularly, in one place.
It’s completely free. No subscription, no upsell — just a practical resource I wanted to exist when I started exploring this.
Conclusion
Twenty years of ESL teaching gave me the experience to know what a good lesson looks like. AI gave me a way to build one without burning three hours of a Sunday to do it.
If you want to try this yourself, grab the free 50 AI Prompts for ESL Teachers guide below. Pick one prompt, try it before your next class, and see what you think.
