Unfamiliar language
Words, spellings and examples were often American and unlike those our children met at school.
Our story
When our children stumbled across generative AI while gaming, we realised keeping them away from it wasn’t the answer. They needed to understand what AI was, how it worked and how to use it safely.
The first problem
We started with ChatGPT and quickly saw the gaps through a child’s eyes.
Words, spellings and examples were often American and unlike those our children met at school.
Responses were long, dense and designed around the way adults read and understand information.
Content was unfiltered, the product was built for adults, and we couldn’t see how the AI was being used.
So we decided to build the AI experience we wanted our own children to have.
Our children loved it—and shared it with their friends.
Then we listened
One classroom can contain an extraordinary range of age, ability, confidence and need. SEND is becoming more prevalent and more complex, while teachers are expected to make one lesson work for everyone.
Even within a single year group, the oldest and youngest children can be almost twelve months apart. Ability can vary enormously, and each pupil arrives with a different experience of the world.
They wanted AI that understood their curriculum and the way they teach—using language children understand and adjusting the level of support on its own.
When we put those ideas into classrooms, the impact was immediate. Better adaptation did not just make the technology easier to use; it made learning more reachable.
Working with schools, multi-academy trusts and local authorities taught us another lesson: too much education technology is bought but barely used. Schools compromise, and some providers disappear once the contract is signed. We chose a different way of working.
That principle still guides every SchoolGPT product.