cast Talk to us

Portal event

Learner state

Outcome to prove

Customized version

Tell us who you are, then land on your version.

Use the bottom bar to describe your role and platform. This section will rewrite itself around your education portal and jump you here.

Waiting for your role Waiting for platform context Recommended scene pending
Native situation

The portal moment

We will identify the learner moment where your portal already has intent, data, and friction.

Avatar layer

What the avatar adds

The avatar becomes a visible emotional and feedback layer inside that moment, not a separate assistant floating outside the product.

Pilot signal

What to test first

Start with one measurable behavior: first attempt, retry quality, return-to-task, spoken completion, or hint-to-answer ratio.

View matching avatar scene

Turn the target list into native scenes.

The page should let every P1 category recognize its own learner workflow first. We can keep a shared product engine underneath, but the sales surface should feel specific by niche.

Language & speaking

Read-aloud turns, pronunciation retries, public-speaking confidence, and conversational wait time.

Language learning appsKids English tutorsSpeaking-coach portals

Math & adaptive practice

Wrong-answer loops, step decomposition, attention pointing, and gap repair after partial progress.

Gamified math appsMath tutoring platformsAdaptive AI tutors

Live tutoring & classes

Live checkpoints, quiet learners, class participation, and warm handoff between tutor and portal.

Live class platformsStudy-buddy tutoringInterest-led class marketplaces

Homework & exam prep

Hint without answer reveal, test-prep pressure, concept repair, and return-to-task after errors.

Homework help platformsExam-prep apps

Coding & project learning

Debugging hesitation, project checkpoints, explain-before-run moments, and confidence after errors.

Kids coding academiesProject-based coding labs

The fake avatar is only a contract.

For now, the CSS character proves where the avatar sits, what state it enters, and what product event controls it. Later, the same slot can mount the real on-device avatar runtime.

portalEvent

Inactivity, wrong answer, hint request, speech activity opened, live poll, partial progress, code error, or review page opened.

avatarState

Listening partner, neutral repair, focused guide, confidence coach, exam mentor, coding coach, or adaptive tutor.

cueGrammar

Posture, gaze, expression, gesture, wait timing, voice tone, and turn-taking behavior. No camera-based emotion inference required.

pilotMetric

First attempt, spoken completion, retry quality, self-check before submission, return-to-task, hint-to-answer ratio, or participation rate.

Research is the anchor, not the headline.

The category scenes are product hypotheses grounded in feedback, active engagement, and embodied-agent mechanisms. We should validate each native scene with its own learner group and subject matter.

  1. FEEDBACKHattie & Timperley — The Power of Feedback

    Anchor for feedback that helps learners understand the goal, their current state, and an actionable next step.

  2. ENGAGEMENTChi & Wylie — The ICAP Framework

    Anchor for moving learners toward active, constructive, and interactive engagement.

  3. EMBODIMENTMayer & DaPra — Embodiment Effect with Pedagogical Agents

    Mechanism evidence for gesture, gaze, facial expression, movement, and voice; not direct proof for every K–12 context.

Make one niche feel seen before showing twelve.

The strongest version of this microsite may become segment-specific: send a speaking platform the speaking scene, a math platform the retry scene, a homework platform the hint scene, and a coding platform the debugging scene.

Talk to us about a category demo