AI & personalisation
AI at the core. Humans in charge.
Current trial: Pondo keeps humans in charge: permissions, consent, security and safety stay with fixed, testable rules, and Pondo is always clear about what it knows and where it isn’t sure yet.
The intelligence loop
- 1
Observe
Notice what the child does — answers, support used, what happens next.
- 2
Infer
Form a careful view of where understanding is, including possible misconceptions.
- 3
Decide
Choose the next task, explanation and the smallest useful support.
- 4
Teach
Deliver the explanation or representation most likely to help.
- 5
Verify
Check independent understanding, retention and transfer.
- 6
Update
Refine the picture of how this child learns — and say when it isn't sure yet.
What AI decides
Learner-state inference
An evolving, evidence-based view of what a child understands.
Misconception hypotheses
Possible reasons an answer went wrong, treated as hypotheses to test.
Next task
What to work on next, within the curriculum.
Explanation choice
Which explanation to offer, and when to try another way.
Scaffold choice
The minimum effective support for this moment.
Representation
Switching between pictures, numbers and steps when it helps.
Retention scheduling
When to revisit earlier learning.
Transfer checks
When to try an idea in a new context.
Adult summaries
Concise, honest summaries for parents and permitted teachers.
What deterministic systems always own
Sign-in
Account access is handled by fixed, secure systems.
Permissions & consent
Who can see what is decided by adults and enforced by rules — never by AI.
Security
Security controls are deterministic.
Curriculum boundaries
Hard limits keep learning within the approved curriculum.
Audit
Important changes are recorded so they can be reviewed.
Safety rules
Fixed safety rules apply regardless of what the AI suggests.
Honest about uncertainty
Begin your one-month Pondo trial.
Take a thoughtful first month with Pondo and see how it supports the learner, not a label.