Most monitoring products are judged on whether they detect the event. The harder problem is what happens in the ninety seconds afterwards — who is told, what happens when nobody answers, and how you stop four relatives all driving over for the same fall. Trigger one below and watch it run.
Pick a scenario. The phone on the left is the caregiver's screen; the timeline on the right is the escalation ladder that runs behind it. Tap “I'm on it” at any point and the whole chain stops for everyone.
These are the events NeuraSense is built to catch.
The alarm is the visible half. The other half is a quiet daily record of walking, sleeping and activity, measured against the resident's own prior pattern — never against a population average. This is a sample fourteen-day view.
Two measures have moved away from her own baseline, so they are flagged for review. A flag is a prompt to look — not a diagnosis, and not a prediction. The same figures export as a one-page PDF to hand to a clinician, with the data-completeness note on it, because a trend computed over three days means something different from one computed over fourteen.
Radar can resolve heart and breathing rate only while the chest is the dominant motion in the room — which means while she is lying still. These are therefore measured across the night, inside the time actually spent in bed, and are meaningless as a daytime average.
These do not appear in the family view and never raise an alert. They are shown to a clinician, labelled as exploratory, and nothing in the product acts on them. Contactless vital signs are unvalidated in this population and have no established threshold here.
The built-in assistant answers questions about how the product works. When a question is medical, it does not answer — it hands you your own figures to take to the clinician instead. These are real responses from the shipping system.
NeuraSense