Every child learns differently. Now, finally, something
teaches differently.

A personalised AI companion for children aged 5–11 — loved by kids, trusted by parents.

Every parent asks it. Almost nothing answers it.

And the tools we give them feel like chores.

What kids are given

generic worksheet homework

drill apps they’re told to open

one explanation style for everyone

What kids actually want

to be known

to play

to feel clever

No child has ever asked to do ten more minutes of Doodle.

Everyone picked a lane. The lanes don’t overlap.

Parents trust it →

Kids love it ↑

AI companions

(Character.ai) — loved, untrusted

Answer machines

(Photomath et al.) — neither

Classroom AI tutors

(Khanmigo, DfE pilots) — trusted, unloved

Drill apps

(Doodle, Atom)

“A companion children love, that parents trust, that school benefits from. Nobody is here.”

"Khanmigo consumer adoption: 'a non-event'"
— industry commentary, 2026.

Great pedagogy isn't a consumer product.
A relationship is.

The moat isn’t the model. It’s the memory.

A friend who happens to be a brilliant teacher.

Pip

here for Maya

Message Pip…

memory-driven check-in — Pip knew about the tooth and the practice

talk or type — real-time voice; five-year-olds who can’t type yet get the whole friendship by talking

and there it is: multiplication, smuggled inside her actual life

Child picks the avatar, names it, tunes its personality. You don’t abandon a friend you made.

Point the camera at bad homework. Get good learning.

Scanning · Homework

“Coach me through it” Socratic — never gives answers

“Practise first” bespoke exercises, same objective, her way

Every scan-and-solve app gives answers and undermines school. Pip extracts the learning objective and reinforces it. School-positive by design.

The parent finally gets to see — and steer.

Maya · Week 24

On track in 4 of 5 maths strands

Equivalent fractions clicked on Thursday — she explained it back unprompted

Reading age still ~14 months ahead

One thing to try this weekend: cooking = fractions she can eat ›

one brain · two faces

Parent → Pip

“Parents’ evening: Miss Patel says she rushes her writing.”

“Grandad died on Sunday. Be gentle. Don’t raise it unless she does.”

Steering, sensitivities, teacher feedback in. Plain-English progress and one weekend action out. This screen is why parents pay.

What Pip believes about Maya.live — click around

Every night, a distillation pass turns the day into understanding: updates the model, re-scores mastery, drafts the digest, plans tomorrow’s check-in. This is the compounding asset.

Safety isn’t a feature. It’s the product.

The regulatory bar is rising (AADC, Online Safety Act, DfE safety standards). We treat it as a moat: clear it first, wear it as the brand.

Three curves just crossed.

Priced against tutoring, not against apps.

The next twelve months.

Q3 ’26

prototype agent loop · design-partner families recruited

Q4 ’26

MVP live (Years 3–4) · 50 families · safety review passed

Q1 ’27

paid UK launch · digest-driven retention proof

Q2 ’27

first school endorsement pilots · DfE programme entry · efficacy measurement begins

Success metric the only number that matters early: children opening Pip voluntarily ≥4 days/week at week 4.

It grows up with them.

AGE 5
a voice-first playmate teaching phonics through stories.

AGE 9
a homework coach who knows fractions scare her.

AGE 13
a study partner with seven years of context — parent visibility stepping back by agreement.

AGE 17
exam mentor. First-job counsellor, later.

“Every child, everywhere, with a brilliant, patient mentor who has known them their whole life. That has never existed. It’s buildable now.”

Pip. The friend that knows how your child learns.

Raising [£X pre-seed] to ship the MVP, prove voluntary daily use, and clear the safety bar first.

Chris Meehan · cdcmeehan@gmail.com