Tiers: Pilot — ships at launch for the NCCC pilot. Phase 2 — follows after pilot feedback. Future — in the long-term vision, not yet scoped for build.
Onboarding
| Feature | Description | Tier |
| Conversational onboarding | The Coach gathers the learner's goals, background, and communication style through dialogue — no forms, no formal assessment. | Pilot |
| Avatar assignment | Based on onboarding signals (cultural background, age/life stage, communication style), the Coach assigns the most appropriate avatar from the pool. | Pilot |
| Privacy-first welcome | The first screen surfaces the privacy commitment before any interaction begins — "Your progress is only visible to you." | Pilot |
| Module recommendation | The Coach recommends a starting module based on what the learner shared in onboarding. The learner can accept or choose differently. | Pilot |
| Returning learner recognition | On subsequent sessions, the Coach greets the learner by name and references their last session specifically. | Pilot |
| Learner profile refinement | The Coach continues building its understanding of the learner across sessions — adjusting tone, pacing, and content recommendations. | Phase 2 |
AI Coach and Avatar System
| Feature | Description | Tier |
| Avatar pool | A set of diverse AI avatar characters representing the learner population — Aboriginal and Torres Strait Islander, CALD, and a range of ages. | Pilot |
| Avatar voice output | The Coach speaks all feedback, prompts, and session framing aloud. The avatar lip-syncs to audio at key moments. | Pilot |
| Key-moment avatar presence | The avatar appears and speaks at defined interaction moments — not continuously. Between key moments, the module is foregrounded. | Pilot |
| Voice-first learner input | Learners can speak their responses to the Coach rather than typing. The Coach transcribes and responds to spoken input. | Pilot |
| Text fallback input | Learners who prefer not to use voice can type responses. All Coach functions are accessible via text. | Pilot |
| Culturally safe avatar design | All avatar characters are designed in consultation with relevant community representatives. Aboriginal and Torres Strait Islander avatars require dedicated cultural consultation before finalisation. | Pilot |
| Adaptive pacing | The Coach adjusts its speech pace and response timing based on the learner's pattern — pausing longer for learners who need more processing time. | Phase 2 |
| Emotional tone detection | The Coach reads signals of learner anxiety or disengagement and adjusts its approach. | Phase 2 |
| Avatar customisation | Learners can personalise aspects of their assigned avatar over time — name, appearance elements, voice. | Future |
| Additional avatar characters | Pool expanded with more characters to serve a wider range of cultural backgrounds and life stages. | Future |
Numeracy Module
| Feature | Description | Tier |
| Core game loop | A fast-paced, gamified experience for basic numeracy operations. Designed for short sessions on a mobile device. | Pilot |
| Post-session AI analysis | After each game session, the Coach analyses performance and delivers specific, spoken feedback on strengths and the next skill to work on. | Pilot |
| Adaptive difficulty | The game calibrates its difficulty to the learner's current level — never trivially easy, never overwhelming. | Pilot |
| Session continuity | A mid-session exit saves progress exactly. The learner resumes at the same point, not the beginning. | Pilot |
| Contextualised skill framing | The Coach connects numeracy skills to the learner's stated goals — roster calculations, reading a payslip, working out change. | Pilot |
| Real-time in-game coaching | The Coach intervenes during gameplay to offer hints or reframe a question the learner is stuck on. | Phase 2 |
| Word problem module | English-language word problems that apply numeracy skills to workplace and daily life scenarios. | Phase 2 |
| Employer assessment simulation | Practice sessions modelled on the numeracy components of common job application assessments. | Future |
Reading Module
| Feature | Description | Tier |
| Read-aloud with audio feedback | The learner reads a passage aloud; the Coach listens and gives real-time spoken feedback on pronunciation, pace, and word recognition. | Pilot |
| Learner-selected texts | Learners can read from material relevant to them. The Coach can work with a photo of a page or a typed/pasted excerpt. | Pilot |
| Word-level support | When a learner stumbles on a word, the Coach sounds it out — without making it feel like a correction. | Pilot |
| Passage difficulty progression | The Coach gradually introduces more complex texts as confidence grows — framed as variety, not advancement. | Pilot |
| Text-to-speech for the learner | The Coach can read any text aloud to the learner first, modelling correct pronunciation before the learner attempts it. | Pilot |
| Comprehension conversation | After reading, the Coach asks a simple conversational question — not a comprehension test, but a natural "what did you think of that?" | Phase 2 |
| Vocabulary building | The Coach notes words the learner finds difficult across sessions and returns to them — embedding vocabulary growth into the reading flow. | Phase 2 |
| Reading aloud from a book (camera) | The learner holds their phone camera up to any physical book; the Coach reads along and gives feedback in real time. | Future |
Writing Module
| Feature | Description | Tier |
| Paper-and-photo pathway | The learner writes on paper, photographs it, and uploads it. The Coach reads the handwriting and gives spoken feedback. | Pilot |
| On-screen writing pathway | Writing directly on-screen (finger or stylus) with the same AI feedback. | Pilot |
| Specific written feedback | Feedback is always spoken and specific — what worked, and one thing to focus on next. Never a score. | Pilot |
| Open-ended prompts | Low-stakes, personally relevant writing prompts connected to the learner's stated goals and context. | Pilot |
| Sentence construction guidance | The Coach works with the learner on sentence structure — framed as practical tools, not grammar rules. | Pilot |
| Handwriting recognition | The AI accurately reads handwritten text across a range of handwriting styles, ages, and literacy levels. | Pilot |
| Structured writing practice | Progression from short free-writing exercises toward structured pieces — a short letter, a work email, a form response. | Phase 2 |
| Real document practice | The learner practises completing real-world documents — job application forms, benefit forms, community notices. | Phase 2 |
| Cover letter and application builder | A guided workflow for drafting a job application or cover letter, scaffolded by the Coach. | Future |
Progress and Engagement
| Feature | Description | Tier |
| Private progress view | The learner can see their own session history and a simple view of how they've progressed — framed as movement, not scores or levels. | Pilot |
| Session length flexibility | Sessions have no fixed minimum or maximum. The learner can stop at any time without penalty. | Pilot |
| Re-engagement recovery | After any gap, the Coach warm-starts the learner with a brief recap and a gentle warm-up — not a restart from scratch. | Pilot |
| Milestone acknowledgement | The Coach notices and names significant moments without fanfare or score-posting. | Pilot |
| Goal tracking | The learner's stated goal is held by the Coach and referenced periodically — connecting daily practice to what the learner actually cares about. | Phase 2 |
| Push notifications | Optional, opt-in reminders — gentle, personalised, never guilt-inducing. "Ready when you are" not "You haven't logged in for 3 days." | Phase 2 |
| Long-term progress report | A learner-facing summary of growth over a longer period — expressed in plain language, not charts or levels. | Future |
Facilitator and Community Worker Dashboard
| Feature | Description | Tier |
| Learner enrolment | Facilitators at community centres can enrol learners into the platform and link them to their organisation. | Phase 2 |
| Aggregate engagement view | Facilitators can see how many learners in their cohort are active, without seeing individual learner data. | Phase 2 |
| Referral and warm handoff | Facilitators can flag a learner for follow-up, or the Coach can prompt a learner to speak with their facilitator. | Phase 2 |
| Individual progress (with consent) | With explicit learner consent, a facilitator can see a learner's progress summary. | Future |
| Reporting for funding compliance | Aggregate, de-identified reports suitable for NSW DoE and other funding body requirements. | Future |