I've been teaching ESL for over a decade. I used to spend 2-3 hours every Sunday prepping the week's lessons. Now it takes me about 40 minutes.

Here's what changed — three prompting patterns that actually work for ESL contexts:

1. The Level Adapter

Instead of asking AI to "write a reading passage," I prompt: "Rewrite this article at CEFR B1 level. Keep sentences under 15 words. Replace idioms with concrete language. Highlight 5 key vocabulary words with simple definitions."

The specificity is everything. Generic prompts give you generic output that still needs heavy editing.

2. The Differentiation Machine

"Create 3 versions of this comprehension exercise: A1 (true/false only), B1 (short answer), B2 (open-ended analysis). Same source text, different cognitive demands."

One prompt. Three levels. Done.

3. The Grammar Context Builder

"Generate 5 realistic dialogue exchanges between coworkers that naturally use the present perfect for life experiences. Include cultural notes for East Asian learners who may struggle with article usage."

This one alone saved me hours. AI is surprisingly good at generating contextual grammar practice when you tell it exactly what transfer errors to address.

None of these required fancy tools. Just ChatGPT with specific, teacher-informed prompts.

That's what I teach inside AI for ESL Teachers — not how to use AI as a crutch, but how to make it an extension of the pedagogical expertise you already have.

If you're an experienced ESL teacher who's been curious about AI but hasn't found the time to figure it out, this is built for you.