Evergreen is a warm AI voice that calls every morning — today on a simple tablet or browser, built to grow onto the phone she already owns. Iris screens what she hears on her calls for scam patterns and tells the family; reaching a licensed human in an emergency is the partner path ahead. The daily call is why she answers; the safety net is why the family pays. The call, the screening and the family alerts are built and live today.
Evergreen gives an aging parent a daily companion — today on a simple tablet or browser, built to grow onto her own phone — that screens what it hears on her calls for scam patterns and texts the family; reaching a licensed human in an emergency is the partner path ahead. A warm voice — Iris — calls every morning, and everything she notices lands in one family dashboard that fills itself.
It is not a concept. The voice, the safety backstop, the honest first-call introduction, and the dashboard all run in a working prototype today. What's left before launch is a licensed 24/7 monitoring partner, the stored consent record and the legal wrapper — not the product. Pre-revenue; the monitoring partner is not yet signed.
We're a Canadian carrier, so here's the Canadian case — and it's stronger than the American one. Loneliness erodes an aging parent every day, and it's the thing no free app or pendant actually touches.
We are not sizing this on the US 65+ population — we don't own a network there and the moat wouldn't travel. And the fraud problem is real money: CA$137.9M in Canadian 60+ scam losses in 2024 (40% of all reported losses). The global "AI in elderly care" market is ~US$39B (2025) → ~US$90B by 2031 (≈15% CAGR).
Sources: NIA Ageing in Canada Survey · Statistics Canada (2025) · CAFC (2024 60+ fraud losses) · Mordor Intelligence. TAM/SAM/SOM and ARR are illustrative planning assumptions (SOM ≈ $100 Basic × 1% of 8.1M seniors), not a forecast.
Nothing to install. Nothing to learn. Today she talks to Iris on a simple tablet or browser; her own phone and number — even a flip phone — is where the carrier roadmap takes it. Iris does the rest, and everything she notices lands in one place for the family.
Iris says good morning, every day — the ritual she comes to love. It's the whole reason she picks up, and if she doesn't, somebody hears about it.
Today Iris screens what she hears on her calls for scam patterns and tells the family. The roadmap moves it into the network, before her phone rings — where an app can't follow.
No answer, no interaction — Iris calls. Still nothing, and it escalates on the family's standing order — and to a licensed monitoring partner as we sign one. Alongside a medical alert, never instead.
The family dashboard that fills itself. Every signal above feeds it live, so nobody has to update anything.
StoryLine, Good News and games are the Full tier; the formatted keepsake memoir is a one-time order ($100 for ~80k words, +$10 per extra 50k). Companionship is why she answers the phone; the safety net is why the family trusts it. Basic $100 · Full $200 · Unlimited (bring-your-own-AI-key) $125 — per month.
Everything on the product page is running right now: a real voice you can call, a safety net that texts the family on its own, and a dashboard that fills itself. Here's exactly what's built.
| Capability | Status | What it is |
|---|---|---|
| Iris, the morning voice | Live | A real spoken phone-style call in the browser — speech-to-text, Claude, and a warm synthesized voice — with natural interruption and memory that carries from one morning to the next. |
| The safety backstop | Live | Every call is screened for crisis, scam and emergency language and the family is texted independently of the AI. A second filter checks Iris's own words before Margaret hears them. |
| Honest, from the first call | Live | Iris says plainly she's an automated voice on the very first call, and any "no" or "stop" is honoured immediately — calls pause and the family is told. The stored, revocable consent record ships with the phone line. |
| CareCircle dashboard | Live | The family dashboard, filling itself from real check-ins, blocked scams and alerts — nobody updates anything. |
| The daily rhythm | Live | Plan-aware call limits, a gentle wind-down, and a graceful "that's all for today" that still reaches the family if she's lonely and keeps calling. |
| Built to be trusted | Live | Signed sign-in links, encrypted keys, hardened access — the unglamorous parts a family's trust actually depends on. |
| Licensed 24/7 monitoring partner | Next | The regulated desk that can dispatch real help is fully designed and specified — signing it is the first gate, before we put this in front of families. |
Live in a working prototype — try the interactive demo ↗ — with a public overview site. No paying customers yet, and the monitoring partner isn't contracted — but everything the family experiences is real and running.
The market splits into medical-alert pendants, AI companions, and scam blockers. No competitor bundles a daily companion, scam screening of her calls, a family dashboard and an emergency escalation path — let alone a carrier roadmap that moves the screening into the network itself.
| Capability | Evergreen | TELUS LivingWell | ElliQ | Alexa Together † | Scam apps |
|---|---|---|---|---|---|
| Daily AI companionship, with memory | ● | — | ● | partial | — |
| Scam screening (in-call today · in-network on roadmap) | ● | — | — | — | app only |
| Licensed 24/7 emergency dispatch | planned | ● | — | ● | — |
| Family dashboard that fills itself | ● | app | ● | ● | — |
| No device to buy or charge — her own phone | roadmap | pendant | robot | speaker | ● |
| Canada-first, carrier distribution | ● | ● | — | — | ~ |
| Price / month | $100–200 | ~$35–55 | ~$249 + sub | $19.99 (shut 2023) | ~$5–10 |
† Amazon Alexa Together discontinued 2023. Evergreen's emergency dispatch runs through a licensed monitoring partner (designed; signing is the first gate). "planned" / "roadmap" mark the carrier path Evergreen grows into — today Iris's call runs on a simple tablet or browser. Competitor facts from public sources; TELUS figures are promotional (existing-customer, 1-yr term). ElliQ is a US countertop robot; Papa (not shown) is human-companion, sold via US health plans.
Fair question. Here are the bodies — and what actually killed each one.
Amazon had 500M devices, Prime's billing rails, and real EMS dispatch — at a fifth of our price. It died without an obituary, then relaunched the emergency button alone at $5.99, sold to the senior, not the child.
They didn't have to find customers — they bought 900,000 already-paying subscribers — and still impaired over half of it. They also owned Geek Squad and 1,000 stores full of the exact buyer.
Human companions. 1,200+ complaints, allegations of assault and theft — and Humana, Aetna, Cigna and ~36 others gone in a single renewal cycle after one Bloomberg story.
The best-funded AI companion in the category. Its flagship deployment was bought by a state agency. Harvard wrote the case study; its central question is "should we go back to consumers?"
Papa proved health plans will pay for this — and that you cannot deliver it with gig workers. Iris costs cents a call and cannot assault anyone. That's why this works now and didn't in 2023.
Let's be honest: owning the fibre saves about $2 a month per subscriber. That's not a moat — it's a rounding error a well-funded rival erases on day one. Here's what actually compounds:
Who we call, in what order, at 3 a.m., and what they're allowed to do. Nobody re-enrols their whole family to save ten dollars.
Her cadence, her word-finding, her mood. The value isn't that Iris talks — it's that Iris knows what a good morning sounds like for Margaret, and this one wasn't.
Today the scam shield lives in Iris's calls. Because we own the network, the roadmap moves it into the signalling path — the call dies before her phone rings. An app can never follow us there. This is what owning a telecom is worth.
Papa was worth $1.4B and lost the category in one news cycle. Trust is expensive to build, instant to destroy: a moat if you're careful, a cliff if you're not.
An AI that detects an emergency and has nowhere to send it is not a safer product. So the architecture puts a licensed, insured monitoring partner in the path — signing one is the first gate — and Iris is honest with the parent from the very first call.
No answer, no motion, slurred speech, a fall mentioned, the safe word.
Calm, patient, keeping her talking.
A trained agent at a regulated PERS centre, white-labelled as Evergreen, who decides — and can dispatch help under their own licence and insurance. Signing this partner is the first gate before launch.
Evergreen never dials 911 — that's the licensed partner's job once signed. The family always gets the alert too — those independent family texts are live today — and escalation runs on the family's standing order.
The daughter buys it; the mother consents to it. Under the TCPA an AI voice is an "artificial voice," and consent must come from the person being called. Twenty seconds of product; the whole legal foundation.
A buyer who feels the problem every day, a distribution channel no AI competitor has, and an enrolment step that doubles as the legal foundation.
Three cities away and carrying the worry. She buys peace of mind, with nothing for Mom to install or learn — today a simple tablet or browser, her own phone and number on the carrier roadmap.
Warm distribution into an existing telecom base (Fibernetics / Fongo) that no US AI player has — and that the device incumbent (TELUS) can't match on companionship.
Iris says plainly she's automated on the first call and honours any "stop" — live today. The recorded, revocable consent step ships with the phone line: the enrolment step and the TCPA / CASL legal foundation at once.
Start on Basic (daily call + safety net); grow into Full (StoryLine, Good News, games) and the keepsake memoir. Later, a B2B / white-label tier for health plans.
The adult child signs up; Mom just answers Iris's call. Iris is honest about being automated from call one.
She comes to love the call; the family watches it fill the dashboard. Six months of "normal" accrues.
StoryLine, Good News, games, and a one-time formatted memoir order.
Roll out on-net; open a white-label tier for health plans.
Any deck that calls this an 80%-gross-margin business by ignoring inference is lying to itself. Here's the real number — one we now control, because we tune the AI ourselves.
Inference gets cheaper every quarter; the live product already cascades routine turns to a small model, caches hard, and caps the tail — with daily limits built in. Owning the network is worth about $2 a month: worth having, not the business.
Real pricing, a cost to serve we now control, and a Canada-first ramp toward the SOM. The figures below are illustrative planning assumptions, to be replaced with a measured model.
The near-term spend is small and specific: the two-week legal-architecture sprint, integrating and signing the licensed monitoring partner, and a small build + go-to-market team to run the first-families pilot. The product itself is already built.
Illustrative planning assumptions — not a forecast. COGS pending a measured model; SOM/ARR figures sized for order-of-magnitude.
Evergreen isn't starting from zero: it runs on the AI-receptionist platform we already operate, on our own network. The hard, regulated parts are outsourced to specialists by design.
Built on the AI-receptionist platform we already operate, made to distribute over our own carrier network — not a greenfield build and not someone else's cloud economics.
A regulated, white-labelled monitoring desk handles dispatch under its own licence and insurance — we don't reinvent PERS. Evergreen never dials 911 itself.
A small build + go-to-market team runs the first-families pilot. The expensive, risky part — building the product — is already done.
To be named for the board copy — leadership, advisors, and the selected licensed monitoring partner.
Named plainly, with the mitigation for each — most already engineered into the live product.
Mitigation: consent-first onboarding, litigator-reviewed copy, breach-notification readiness.
Mitigation: a licensed, insured partner sits in the path; Evergreen never dials 911 itself.
Mitigation: an independent keyword backstop + a reply guardrail + human escalation — already live.
Mitigation: careful, consented launch; transparency; no gig strangers in the home.
Mitigation: small-model cascade, caching and daily caps — already shipped.
Mitigation: runs on an existing platform + network; first hires are a lean build + GTM team for the pilot.
The hard part — the product — is done. What's left is the legal wrapper, the licensed partner, and a careful launch.
The consent flow, the escalation ladder, the licensed-partner crisis protocol, and a full pass over our marketing copy with a litigator holding the pen. In this category your promise is your liability — Life Alert lost seven figures on its advertising, not its protocol. Cheap now, ruinous to retrofit onto 100,000 subscribers later.
Take the working Basic product — the Morning Call, Scam Shield, the safety net and CareCircle, all already running on our stack — to the first Canadian families. Prove retention and the safety desk.
The daily call is why she picks up; six months of Iris knowing what "a good morning" sounds like for Margaret is the thing a free app can never copy. The scam screening of her calls and the safety net around them — with the licensed partner ahead — are what a family pays for — loneliness is the need, and the relationship is the moat.