The Looping method

1. We start from the curriculum, not from a bank of exercises

Before generating anything, Looping resolves the pupil’s school context: year group, subject, and the exact chapter of the official curriculum. That scope is then imposed on the AI as a hard limit. It is this link to the curriculum that later lets us tell a parent “he is stuck on fractions” rather than “he averages 62%”.

2. A session lasts ten to twenty minutes

A quiz has five questions for the youngest pupils, up to twelve for the oldest. Short on purpose: consistency beats one long Sunday session. A pupil who stops midway finds the quiz waiting under “To finish”.

3. The AI questions, it does not do the work

Targeted question, hint, partial explanation, then the full reasoning — and only after genuine attempts. Every correction gives the rule or the reasoning, not just the right answer. The tone is neutral: no automatic praise.

4. Pupils can challenge a correction

An AI makes mistakes. The pupil explains why they think they are right, the answer is re-examined, and the mark is raised if they are. It is also our best detector of generation errors.

5. Effort becomes visible

Screen time earned, points, badges, daily streaks: concrete markers that give a reason to come back tomorrow. No money rewards. Each badge states exactly what is left to unlock it.

6. Parents stay in control, without hovering

You set the reward rules, the allowed hours and the daily AI usage cap. You see success subject by subject and chapter by chapter, the exchanges with the AI, and even behaviour during quizzes.