Explore Programs
The Global AI Learning Company

AI for Learning.
Agency for Life.

Learning · Research · Innovation · Community

We are not a tutoring centre. We study how children actually learn alongside AI — then we build the programs that come out of that research.

Explore Programs Visit the Lab
What we do

Four things, and we do them in this order.

Research comes first. Everything a family sees started as a question we could not answer.

01 — LEARN

Learn with AI

Children teach the lesson back to an AI student. You do not know a thing until you can explain it.

02 — AGENCY

Build Agency

The goal is not answers. It is a learner who can decide what to study next, and why.

03 — RESEARCH

Research Learning

We run the lab on our own students first, publish what we find, and correct what we get wrong.

04 — SCALE

Scale Innovation

What survives the lab becomes a program, and travels to learning centres worldwide.

Our ecosystem

One loop, from a research question to a child.

CompanyCyber ClassroomThe AI learning company behind all of it.
EngineFounding AI Learning LabWhere the questions get asked, and answered in public.
OutputLearning FrameworkFindings hardened into a method that can be taught.
DeliveryProgramsThe framework, turned into something a child does on a Tuesday.
PurposeStudents & FamiliesWho then tell us what we still have wrong.
Founding AI Learning Lab

We publish what we find — including what we got wrong.

Notes from the lab, written as we go. No conclusions we cannot show you the evidence for.

Most education companies publish results. We publish the method, the misses, and the numbers underneath — because a claim about a child's learning should be checkable by the parent reading it.

Each note below is a finding already in use inside our own programs. The full write-ups are being published one at a time.

Lab Note 003 · July 2026

An AI marker will hunt for a mistake that is not there

Give a grading model a list of "common misconceptions" and one of them is wrong, and it will find that error in a child's work anyway — then take marks off for it.

Lab Note 002 · July 2026

Why students need AI coaches, not AI answers

Show the answer on a wrong attempt and allow unlimited retries, and the score you record stops measuring understanding. It measures persistence.

Lab Note 001 · July 2026

AI-written lessons fail at the top of the age range

We fact-checked a stratified sample of our own primers. Grade 8 carried real errors; grades 1, 3 and 6 carried none. The harder the content, the thinner the model gets.