When my partner and I relocated to the Netherlands earlier this year, I expected the cultural shift to be exciting.

What I didn’t fully anticipate was how demanding the professional transition would be.

Joining a new school meant stepping into unfamiliar curriculum content, new assessment expectations and a community of pupils whose learning habits I had yet to understand.

In those first weeks, the challenge wasn’t only preparing lessons; it was working out how to make my planning genuinely land with the learners in front of me.

Uncovering a Gem

It was during this slightly messy period that I turned to artificial intelligence (AI), specifically Google Gemini. Partly out of curiosity and partly because I needed something that could help me organise my thinking. I wanted to see whether AI could be more than a generator of generic plans.

I’d used tools like this in the past and abandoned them quickly. With a new timetable and a new country to settle into, I felt it was worth giving AI a more intentional try.

My starting idea was simple: create a class-specific Gem - Gemini’s version of a custom GPT - for each group I taught. A Gem lets you build a small bank of information and set expectations before the AI produces anything.

By giving it clearer, more grounded information about each class, I hoped it might shift away from the bland, “standard” lesson plans I’d been given by AI before and move towards something shaped by the realities of my pupils.

Building slowly

I added to each Gem steadily. Each week, and sometimes each day, I wrote a short reflection: the group’s reading confidence, what kind of scaffolds they actually used and what typical lessons looked like with them.

Because the school uses the Google education suite, the environment is already set up for data-protection requirements, but I still kept all information general - no names or personal details. Instead, I used descriptors like “needs structured reading guidance” or “benefits from clear modelling”.

Over time, these notes grew into a genuinely useful knowledge base: misconceptions that appeared frequently, how much explanation a class needed before independent work and the balance of structure versus autonomy they handled best.

A starting point

Since I was teaching some subjects outside my specialism, I also used the Gem for quick content refreshers in physics, chemistry or geology. This made lesson preparation feel more grounded and less rushed.

When planning, I dropped in learning objectives or a Microsoft PowerPoint outline. The Gem would then generate a draft sequence - title, starter, objectives, possible activities, checks for understanding and a plenary - and it could produce a worksheet if needed.

None of it was perfect, and I never expected it to be. What mattered was that it gave me a starting point already shaped by the information I’d provided. My job became refining the pitch, adjusting explanations and shaping the lesson for that particular class.

As a time saver, this approach made a real difference. More importantly, though, it helped me understand my pupils more quickly. Having to articulate their needs forced me to think more precisely about my teaching decisions.

Pointing the way

The AI’s interpretations sometimes exposed gaps in my own understanding of a class, which nudged me to reflect more deliberately. In this sense, the AI supported my thinking as well as my planning.

That said, I’m under no illusion that AI can produce impactful lessons on its own.

The human elements - judgement, instinct, pacing - remain central. But as a scaffold that helps teachers move from a blank page to something workable, especially when settling into a new system, it has been valuable.

After a few weeks, I shared my approach with colleagues at the British School of Amsterdam, and the response was encouraging. Several teachers said they had tried AI previously, found the results too generic and given up.

Structured method

Seeing a more structured method - especially the class-specific knowledge base - encouraged them to try again. Some have since told me it has saved them time or given them stronger first drafts during busy weeks.

For me, this experience has reinforced something straightforward: AI isn’t a shortcut; it’s a multiplier. When we provide thoughtful, precise information, it gives us something that genuinely supports our practice.

Used carefully, it has helped me settle into a new country, a new curriculum and a new professional environment.

And as colleagues continue to experiment, the conversation is gradually shifting from scepticism to genuine curiosity - a small but meaningful change in an increasingly complex profession.

Henry Ward is a science teacher at the British School of Amsterdam

You can now get the UK’s most-trusted source of education news in a mobile app. Get Tes magazine on iOS and on Android