Connect English

As a CELTA trainer, I spend a lot of time reviewing lesson plans, helping trainees think more critically about their choices, sequencing, and overall lesson logic. Recently, I decided to try something a little different.

I took one of my trainees’ lesson plans (with their permission, of course) and asked a GenAI tool to review it and suggest improvements.

I wasn’t trying to mark the plan. I was curious: What kind of feedback would AI give? Would it be useful, practical, or just generic?

What came back surprised me. 🤯

The AI didn’t rewrite the lesson from scratch. It didn’t do anything a trained tutor couldn’t do. But it did highlight areas where the lesson could be tighter, more learner-centred, and better scaffolded. In short, it acted like a second set of fast, objective eyes – and it made me reflect on how this kind of tool could support both trainers and trainees on intensive courses like CELTA.

Here are some specific before and after insights based on that plan.

🔄 Before and after: What changed?

1. Lead-in Activity

• Before: A quick warm-up with a general question —
  “What do you usually do on weekends?”
• After (AI suggestion): A more targeted lead-in to activate relevant language and interest —
  “Here are four weekend activities. Choose one you’d most/least like to do and explain why.”
  ➤ This also led more naturally into the lesson topic and generated useful lexis.

2. Staging and flow

• Before: The lesson had all the stages, but they felt a bit disconnected. Vocabulary was introduced, then came a discussion, but the link between them was thin.
• After: AI recommended a brief guided discovery stage for vocabulary, so students would notice patterns and usage before the speaking task.
  ➤ This gave the lesson stronger internal logic and made each stage feel purposeful.

3. Task set-Up

• Before: Instructions were fine but minimal:
  “Discuss these questions in pairs.”
• After: AI suggested using an example, a quick demo, and a visual prompt:
  “Think about your answer first. Then listen to my example. Now discuss in pairs.”
  ➤ This made expectations clearer and improved learner confidence, especially at lower levels.

4. Timing and pacing

• Before: Some activities were likely to overrun, while others were too short to be meaningful.
• After: With AI’s help, I redistributed time to allow more space for student output and reduced time spent on teacher-led sections.
  ➤ The result? A lesson that felt more balanced and student-centred.

5. Feedback and wrap-Up

• Before: A quick error correction stage, then the lesson ended.
• After: AI proposed a short learner reflection task:
  “What’s one thing you learned today? What would you like more practice with?”
  ➤ This added a metacognitive element and gave the lesson a more reflective close.

💬 A quote that stuck with me

At a conference I attended, Louka Perry said something that really resonated with me:

👉 “AI is not going to replace teachers, but it might replace teachers who do not know how to use AI.”

That quote stayed with me, and this experience brought it to life. AI isn’t replacing tutors, and it certainly isn’t replacing teacher development. But it can support it, especially when used as a reflective or co-planning tool.

🎯 Final thoughts

This wasn’t an exercise in letting AI take over. It was an experiment in collaborating with a tool that’s fast, accessible, and getting smarter all the time. As trainers and educators, we have the opportunity – and responsibility – to explore how AI can support our work and enhance how we guide trainee teachers.

For CELTA trainees, it could mean getting instant input when working independently. For tutors, it might offer new ways to model feedback or prompt reflection.

Either way, one thing’s clear: AI isn’t going away. The question is how we choose to engage with it.References
Perry, L. (n.d.). Conference presentation quote on AI and education.