When AI makes materials for free, what a school should pay for

General AI tools make material creation free. Why a tool built for a teacher's work differs from a general assistant, and what you pay for then.

Lauri Stigell, CEO, Bruukki · Updated 12 August 2026
When AI makes materials for free, what a school should pay for

Producing text, exercises and slides with AI is quickly becoming a free basic feature. General-purpose AI assistants such as ChatGPT, Gemini and Claude already do it, and versions aimed at teachers have joined them. When creating a material is free, it is no longer something a school should pay for.

This is a good moment to ask what an institution then actually pays for. The answer is not the speed of making material, because that will soon be available to everyone. The answer lies in what a general-purpose assistant cannot do.

A general assistant does not know a teacher’s work

A general-purpose AI is a horizontal tool: it does a little of everything for everyone. It writes fluent text on any topic, but it does not know the field of education or the natural stages of a teacher’s work. It does not know what your institution’s curriculum emphasises, by what criteria competence is assessed, or how a week moves from planning to material to differentiation.

A tool built for a teacher’s work is vertical: it knows the field and supports the natural processes of teaching. The difference shows immediately in the result. From a general assistant you get an average set of exercises; from a tool built for the field you get a material that is connected to objectives and assessment criteria and that you can get straight away at different skill levels.

So what a school should pay for

Three things do not come with free generation, and they are exactly where a general assistant lets a teacher down.

The first is knowledge of the field and curriculum accuracy. A tool built for the field leans on the right curriculum, knows the logic of differentiation and assessment, and supports the flow of the week. Giving the curriculum as the basis for the work improves accuracy and clearly reduces errors.

The second is knowing where students’ and teachers’ data goes. A free tool may store inputs anywhere and use them to train models. What matters to an institution is where the data physically is, who holds it, and on what basis it might transfer outside the EU.

The third is continuity. A purchase decision is a partnership decision. It is worth asking who owns the tool, who builds it, and what the direction looks like a couple of years from now.

Transparency about the engine

We use Claude as Bruukki’s engine, because in our testing it has produced the highest-quality pedagogical text. The value, however, does not come from a model that is available to everyone. It comes from the teacher layer built around the model: curriculum binding, differentiation, the EU data boundary and the teacher’s decision-making power. The same model as a general-purpose assistant does not do this work. Because the value is in this layer and not in a single model, we can change the engine as needed when a better one becomes available.

The same applies in the institution’s direction. Which language model Bruukki runs on is a workspace setting, and it can be agreed with the institution, so Bruukki is not tied to one model. When the model changes, teachers do not get a new tool to learn. The views, the materials and the ways of working stay the same, and no one has to be retrained. An institution therefore does not have to choose a language model for good when it chooses Bruukki.

How Bruukki fits

Bruukki is built for this spot. The curriculum is included ready to use: the national core curricula for basic education, general upper secondary and vocational education, plus over a thousand local curricula. Any curriculum can be added to the service, so higher education, for example, is not left outside the ready-made list. Students’ and teachers’ data is stored in the EU, and Bruukki does not use your content to train language models. We build the product in Finland and see the everyday life of schools up close.

An honest caveat

We do not claim to know yet which of these decides the purchase in a Finnish school. The EU data boundary is an advantage in Europe, and curriculum accuracy is a genuine need. What ultimately makes a school pay, though, will only become clear in practical conversations with institutions. That is the work we are doing right now.

If you want to see what a tool built for a teacher’s work looks like, try Bruukki for free.

Read also: support for the whole teaching week: what a teacher can do with Bruukki.

Bruukki's AI was used to help write this blog post.

Try it with your own material

The AI suggests, you decide.