All core functionality works without internet. No loading spinners between a child and their voice. AI-powered features deferred until connectivity is reliable in typical usage contexts.
One core app with layered innovation. Depth first (kitchen scene + communication loop), then breadth (new scenes, modes, interactions). Single App Store listing where reviews compound.
Work-for-hire IP assignment to AbleNet. Ari publishes app independently to maintain liability separation. AbleNet stays anonymous as publisher to preserve agnostic positioning in the AAC market.
Superseded by D9. Superseded by D9. Originally: narrow focus on 0-36 month population. Replaced with 12-48 month milestone-based framework per 2026-05-18 meeting.
Superseded by D9. Superseded by D18. Originally: storybook illustration style. Replaced with Bright & Bubbly per 2026-05-28 meetings — confirmed by both Adam Wing and SLP reviewer Alyse Royba.
Scene-based navigation is the differentiator (Kitchen, Playground, etc.). Flashcard mode added for SLP-familiar browsing by category. Both coexist — scenes for contextual use, flashcards for clinical drill. Reference: SuperDuper methodology.
Intentionally limited vocabulary (~1,000 words) vs competitors' broader catalogs. Focused on high-frequency, high-impact words for Meadow's target milestone window. Depth over breadth — every word earns its place.
Weekly sync established: Monday at noon. Team of 4 (Adam, Ari, Jen, Joe). Progress tracked via deliverable portal.
Supersedes D4. Drop 0-12 months entirely. Target 12-48 months chronological age — children arriving with missed language milestones identified by pediatrician/SLP referral pathway. Tiers defined by expressive/receptive milestones, not cognitive developmental bands.
Meadow is a speech generating device (E2510). Child must select words themselves for it to qualify as an SGD. Teaching features (Learn mode) are supportive but secondary — they cannot replace or impede direct communication paths.
'I love you' must be accessible from every screen and every context. Not buried in a vocabulary category. Based on family testimonial data — consistently the most meaningful communication milestone for families.
Text labels on vocabulary items are an SLP-configurable setting. Some SLPs require text for literacy development, others prefer image-only. FlexSpeak precedent. Default: off (image-only for pre-literate users).
Initial content scope: kitchen, bathroom, bedroom contexts within the home. School context removed for 12-48 month age group. Daycare acceptable as secondary community reference but not primary focus.
Navigation shifts from location-based ('kitchen') to activity-based ('time to eat'). Routines like getting dressed, bath time, and bedtime replace static room labels. Scenes still exist visually but are anchored to daily routines.
All vocabulary imagery should be animated. 'Everybody's pictures suck' — animation quality is the visual differentiator. Possible tiered approach: animation in funded device version, static in free version to manage tokenization costs.
Three inseparable audiences: SLP leads (primary customer, makes device/app decisions, guides therapy), parent enables (daily facilitator, home practice, emotional stakeholder), child uses (the actual communicator). Design must serve all three — you cannot separate them. AbleNet's funded model keeps the SLP connected throughout, which is the business differentiator over free alternatives.
Meadow does not assess or diagnose. Users arrive with deficits already identified via the referral pipeline: pediatrician spots missed milestone → refers to SLP → SLP connects to AbleNet → benefit check confirms verbalization deficit → device deployed. The app adapts to what the SLP already knows about the child, it does not determine it.
Bright & Bubbly selected as the visual direction for symbols and scenes. Confirmed by both Adam Wing and SLP reviewer Alyse Royba. Easy to maintain consistency with AI generation. Future style packs (e.g., Storybook for scenes) noted for V2.
ASL signing limited to 12-18 month tier, core words only, toggleable setting (on for T1, off for T2/T3 by default). SLP can override. Signing bubble UI remains but content is tier-gated. Per Adam Wing direction.
Celebration style is a parent/SLP setting: Full (default, age-scaled), Subtle (gentle feedback only), Off. Default stays on — at early ages, celebration is the cause-and-effect mechanism that teaches the child the device responds to them. Toggle exists to respect alternative therapeutic philosophies.
SCS follow-up scaffolding is a setting: Full (sentence templates + follow-up bubbles), Light (templates only, no auto follow-ups), Off (direct speech regardless of tier). Gives SLPs control over how much the app scaffolds without removing the conversational capability.
Promoted me, play, cold, hi, bye from T2 to T1 per SLP review (Alyse Royba). T1 core words now 22 (was 17). All five are clinically justified: me (top-10 core word), play (primary activity verb), cold (pairs with hot), hi/bye (foundational social words for caregiver modeling).
Functional kitchen scene on TestFlight within 30 days for 6 internal reviewers selected by Adam Wing. Kitchen scene prioritized as the hardest complexity gate. Apple approval process is the primary bottleneck, not development speed.
Fill-in-the-blank carrier phrase scaffolding (e.g., 'I want ___ please') available in Tier 2 only, toggleable on/off. Removed in Tier 3 to prevent dependency on pre-built phrases. SLP concern: families over-rely on carrier phrases, children don't learn to combine core words independently.
Apple TTS with kid voice is the required speech output for child communication. ElevenLabs dropped for child speech — mixing AI-generated and Apple TTS for custom words would be inconsistent. ElevenLabs retained only for companion character. V2 plan: server-side custom voice pack generation.
The current 'M with a leaf' logo reads as a different product category and no longer matches the app. Brand direction: logo, name, and app must feel like one coherent product for a pediatric audience (ages 1-4), immediately legible to SLPs and families as a child communication tool. The 'Meadow' name itself is flagged for possible revisiting. The 2026-05-28 brand identity (M1-008) remains accepted as delivered; this is a forward-looking refresh, not a defect.
The app is designed for the SLP first, not only the child — 'the journey is for the SLP; you're taking them on a journey.' Success rests on a three-legged stool of SLP, child, and family buy-in, with the SLP as the priority leg, because a comfortable clinician carries the family and the child along. Grounded in AbleNet's updated 38,000-SLP survey, which found no correlation between the specific app and child-success outcomes — so the win is earning clinician confidence, not a 'better app' in the abstract.
Design rule: ease beats delight, echoing AbleNet's guiding principle that least possible equals best outcomes — make the app simple enough to need no manual, and give the user more control rather than adding surface. Concrete mechanic: rank the vocabulary so density and 'swell' (tile size) scale with relevance, so when a scene is chosen (e.g. snack, breakfast, lunch) the most-relevant words grow and the rest recede, with a standard fallback always available. Equal-sized tiles are a default to avoid. Builds on the 2026-06-11 busyness-reduction and per-tier density work.
Beta SLPs suggested aligning core-word symbols to school-familiar sets for consistency. Decision (Adam): no — the field is already split between licensed symbol sets (TD Snap/PCS, Proloquo2Go/SymbolStix), so there is no single standard to align to, and Meadow intentionally stays out of symbol licensing and the debates that come with it. Meadow's custom art remains everywhere, core words included; the reasoning will be shared with the clinical reviewers.
If Apple voices are not satisfactory and a custom child voice is not feasible, Ari uses his judgment on the best alternative as soon as possible. Voice selection must not block any other work.
Readiness-to-progress signals from usage data are endorsed as a direction. Design driver is family and SLP value — helping families see when a child is ready for the next step. The insurance/reimbursement angle is explicitly not a design driver; that path is more complicated than it appears.
Both beta SLPs independently identified parent coaching inside the app as the missing piece; Adam agrees parent comfort is the real adoption hurdle. A lightweight version is worth shipping now, with the fuller coaching experience to follow.
Directive (Adam): move the pronoun/people section to the LEFT side of the screen so on-screen position reinforces left-to-right English sentence building (subject first). More broadly, selection placement should follow the order words are activated in a sentence — left-to-right — so the board's spatial layout teaches sentence structure. Note for implementation: pronouns currently live on the right edge and the left edge is the routine/scene navigation rail, so this reopens the left-rail role; must reconcile with the scene-navigation axis and preserve a stable motor map (positions are promises). Geometry changes flow through the Placement Tool, not hand-authored in Swift.
Adam's call: the first beta goes to roughly three SLPs in utilization, not forty, starting with Alyse Royba and Nicole Betts. The trade is explicit — a small group who commit dedicated hours and return structured feedback, rather than a large group who install the app and forget it. Every new tester is added by hand, so the team emails Ari with names before access can be granted.
Best case is roughly two weeks to finish the remaining scenes plus about two weeks in Apple review. The earlier hope of landing at the beginning of August is retired. Stated plainly rather than discovered at the deadline.
Both beta clinicians independently named the current onboarding flow as too daunting — for parents, caregivers and SLPs alike. This is the clearest single change the reviewers asked for, and it gates whether families reach the app at all. Related to D33 (in-app parent coaching) but distinct: this is about the cost of getting in the door, not the support offered once inside.
The beta clinicians flagged that a non-grid interface reads as unconventional to practising AAC professionals. A grid toggle has been built in response. Alongside it, the centred-path model is dropped and the scene is removed from tiers 2 and 3, which now use a subject-verb-object layout running left to right (people, then verb, then object) — the direct expression of D34. The governing requirement is that the app serve both audiences: long-term AAC users who expect a grid, and the early-intervention/milestone community for whom the contextual, non-grid view is the point. Toggle, not replacement. Under review with the beta clinicians.