AI Didn't Save Me Any Time. And That Was the Point

If you've sat through any CPD on AI in the last couple of years, you'll know the usual pitch. Save hours on lesson planning. Generate worksheets in seconds. Knock up a starter in thirty seconds.

That story is not completely wrong. But I do think it points teachers towards the shallowest use of AI.

Last week I used AI to plan a lesson. It took me longer than doing it myself.

And I think it helped me build the best resource I've made all year.

The problem wasn't planning faster. It was teaching it better.

I teach A Level Economics, and I was planning a lesson on the inverse relationship between bond prices and interest rates.

If you don't teach economics, that does not really matter. The important point is this: I've taught the topic before, and every time, a chunk of the class do not quite get it.

The textbook explanation works well enough for most students. You go through the formula, model a couple of examples, and they seem fine. But there is always a group who can repeat the rule without really understanding why it is true. They can tell you that when interest rates rise, bond prices fall, but if you ask them to explain the logic, they stall.

That is not really a planning problem.

I did not need a worksheet generated more quickly. I needed a better explanation. I needed a way into the topic that would actually reach the students my usual approach was missing.

What actually happened

I started where a lot of teachers probably would. I asked AI for a simple way to show students the inverse relationship between bond prices and interest rates.

It gave me an interactive visualisation. A slider. Move the interest rate up and down and watch the bond price change in real time.

It was neat. It was accurate. And it was not what I needed.

I looked at it and thought: this shows the relationship, but it does not really teach it. A student could play with that slider for ten minutes and still not understand why the price has to move when the rate changes. It showed the what, but not the why.

So I pushed back.

I said what I actually needed was a step-by-step explanation of why the inverse relationship exists in the first place.

That changed the direction completely.

The AI came back with a chain of reasoning built around a very simple question:

If the bank is offering 10% interest on savings, why would anyone pay £100 for a bond that only gives them £5 a year?

That was the key.

Once students see that no sensible investor would do that, the rest starts to click. The bond becomes unattractive at its old price, so the price has to fall until the return is competitive again. The fixed coupon is the whole point. That is the bit they need to grasp.

From there, I asked AI to turn that logic into a worksheet I could use properly in class, step by step.

So in the end, it took three iterations. Not because the AI got magically better, but because I got clearer about what my students actually needed.

That, to me, is the interesting part.

The AI did not come up with the pedagogy. I did.

But it gave me something to react to, push against, reshape, and argue with. And in doing that, it helped me think more clearly about the actual teaching problem.

This isn't really a productivity story

The most useful thing AI did for me in that session was not save time.

It made me slow down and think harder about a problem I had been only half-solving for years.

That is why I think the usual productivity framing misses the point.

Yes, AI can save time. It can produce a decent quiz in seconds. It can draft some retrieval questions. It can turn out a worksheet quickly enough.

That is useful.

But the more interesting use is when it helps you think properly about why students are getting stuck.

When I kept saying, "No, that's not it," I had to become much clearer about what was missing. Why do these students not get this? Where exactly does their understanding break down? What sort of explanation would actually help?

That process was more valuable than the final worksheet.

And I think that matters even more if, like me, you are in a one-person department.

I do not have another economics teacher next door to ask, "How do you teach this bit?" I cannot just wander into someone else's classroom and borrow a better explanation. A lot of the time, if I want to improve something, I have to work it out on my own.

AI is not the same as a colleague. Obviously.

But it does give you something to think with. Something you can test ideas against. Something that will keep going when you say, "No, that's still not quite right."

For a solo teacher, that can make a real difference.

The question I think we should be asking

If you judge AI purely by whether it saves you twenty minutes making a PowerPoint, you will probably decide it is mildly useful and leave it there.

But I think that misses the bigger opportunity.

The better question is not, "How can this help me plan faster?"

It is, "Can this help me think more clearly about how to teach this better?"

That is where I have found the real value.

Not in pressing a button and getting a finished resource.

In the back-and-forth. In the friction. In being forced to pin down what the problem actually is for the students who are still not getting it.

That is not really about saving time.

It is about having something to think with while you work your way towards a better explanation.

---

Gary Roebuck is an A Level Economics teacher and founder of TeachEdge.ai.