How Innatora works, and what it will never do

Patterns come from fixed, transparent rules, not from AI. Every game, activity, check-in and choice is an observation with a source. Innatora then counts how many observations it has for an area, on how many different days, in how many different activities, and from how many kinds of source.

Three things are kept separate, because they often differ:

• Interest: what your child chooses and goes back to on their own.
• Engagement: how they stay with something, including when it gets harder.
• Emerging strength: comfort with increasing complexity across different games, on different days.

Each area then has an evidence state. It describes how much Innatora has seen, never how good a child is:

Not enough exploration yet

Seen once or twice. Try a few more before Innatora looks for a pattern.

Early signal

A first hint. Worth offering again in a different form.

Emerging pattern

Seen on several days and in different activities.

Repeated pattern

Keeps showing up across different activities, sources and days.

Strong repeated pattern

Consistent over weeks, across many activities and more than one source.

Every insight in the app has a "Why am I seeing this?" button that shows exactly this evidence.

Children have good days and tired days, and anything new is exciting the first time. One session can't tell you much, so Innatora waits for things to show up again across different activities and days before calling something a pattern.

Getting better at the same game can simply mean the game is familiar. So games come in families with different versions and themes, and Innatora only treats growing ease as meaningful when it appears across different games and days.

A pattern seen in games, real-world activities and your own observations is stronger than one seen in a single place. Your check-ins are always shown as your observations, never as measurements.

AI never decides what your child is like. It writes plain-language summaries of patterns the rules above have already found, suggests activities on top of a curated library, and answers your questions in the Parent Companion using only your child's recorded history.

• Your child's name is removed before anything is sent to the AI model.
• AI outputs are checked before you see them: no diagnoses, labels, IQ, percentiles or career predictions.
• Questions about medical or developmental conditions get a careful, fixed answer and a suggestion to talk to a qualified professional.
• AI can still make mistakes, so the evidence it was given is always shown beside it.

Innatora needs only a first name or nickname and an age. No birth date, school, address, location, photos or health information is required.

Optional full names and photos used to personalise Discovery Passports or certificates are processed on your device and are not uploaded to or stored by Innatora.

Site visits are counted without cookies or IP addresses, and not at all while a child is in Child Mode. Product analytics are counts only: they never include a child's name, anything a child typed or notes you write.

You can download everything, delete a child's profile, or delete your account at any time from Your data & account. Read the full Privacy Notice.

• Child Mode never shows adult analytics, notes, payments or settings, and leaving it needs a grown-up.
• Children can't message anyone, and there are no ads or links out of Child Mode.
• Activities list materials, supervision and safety notes. You decide what's safe and stay nearby.
• No streaks or pressure: Innatora celebrates exploring, never punishes a missed day.

Privacy Officer: privacy@innatora.com

Anything else: contact us. The research behind the games.