Trust
Built for clinical trust.
“Trust is the product. In dementia care, an unsafe AI is worse than none.”
Decision support, not diagnosis
ORMAI supports a clinical decision; it never makes one. There is no autonomous clinical action anywhere in the product, and the clinician stays in control of every finding, referral and report.
Independent safety gate
Every AI response passes a mandatory safety filter before it reaches a user. The filter is independent of the model, and it can only be made stricter — never relaxed at runtime, never bypassed by a prompt.
Privacy by design
Protected health information is de-identified and minimised before any external processing. Storage is on-device first, and the product operates offline when there is no connection.
Full IEC 62304 dossier
A complete medical-device design-control dossier has been authored, with verification executed and 49 of 49 core algorithm and privacy unit tests passing. A dossier authored is not an authorisation, and we do not describe it as one.
Crisis support
Distress gets a person, not a chatbot.
The product ships free helpline routing for caregivers in distress. A positive response on PHQ-9 item 9 triggers an immediate escalation path rather than a conversational reply.
How we describe ourselves
The language we hold ourselves to.
| We do not say | We say |
|---|---|
| “Clinically validated platform.” | “Built on validated instruments, with psychometric validation under way.” |
| “IEC 62304 certified.” | “A complete IEC 62304 design-control dossier has been authored, with verification executed.” |
| “Patented.” | “Patent-pending. Indian provisional filed; PCT targeted 2027.” |
| “AI diagnoses and recommends treatment.” | “Decision support, never autonomous diagnosis. The clinician stays in control.” |
| “Monitors your loved one 24/7.” | “Supports longitudinal monitoring between clinical visits.” |
| “Our sensing is deployed in care homes.” | “Tested in controlled conditions with human participants. Integration is under way.” |
