Sensi.ai, an always-on audio monitor for seniors aging at home, has raised $100 million and now operates inside 80% of the largest home care networks in North America, according to the company. The white box sits under a table or chair, records continuously, and flags falls, coughs, distress calls, and what Sensi describes as deviations from a senior's normal routine. The company claims a 90% accuracy rate, with edge cases reviewed by a human in the loop.
The device does not announce itself to the person being monitored. In one case documented by WIRED, Sensi captured a transcript of an 86-year-old man in North Seattle telling a guest he had nearly fallen — a conversation his son, 5,000 miles away in Austria, read months later. The same household has cycled through 8 caregivers in the past year. Sensi has remained constant.
Sensi's algorithm is built on what cofounder and CEO Romi Gubes describes as 1,000 years of audio data. Asked how the underlying models were trained, Gubes said only that they use anonymized datasets stripped of personally identifiable information, and declined to detail the sources. The company markets the device to families as a safety net and to agencies as a revenue tool — one testimonial on Sensi's site claims 88% client growth and an 85% increase in billable hours after deployment.
Key facts
- 01Sensi.ai has raised $100M and says its always-on audio monitor is used by 80% of the largest home care networks in North America.
- 02The company claims a 90% accuracy rate on fall and distress detection, with edge cases routed to a human reviewer.
- 03Sensi has not received FDA clearance for its cognitive decline claims; CEO Romi Gubes says the process has been initiated.
- 04A private nursing home room averages $108,000 a year, roughly double the median income for households over 65.
- 05An estimated 9 million caregiver positions will need to be filled in the US before 2031.
Sensi sits in a crowded field. Earzz and Ally Cares monitor care home residents for coughs, falls, and atypical movement. Cherish Serenity uses radar to detect falls and slumping and bundles with AT&T for emergency response. Unlike Alexa, none of these devices wait for a wake word — they begin recording on triggers like thuds, screams, or unusual movement patterns.
The clinical claims are where the technology gets contested. Sensi markets itself as a tool for tracking cognitive decline by identifying shifts in speech, tone, and activity. Ihab Hajjar, a neurologist who studies AI-based dementia detection, said he has seen clinical models flag 60 to 70% of patients as cognitively impaired when the true prevalence is closer to 10 to 15%.
“I have not seen good evidence from any protocol of analysis that would make me feel comfortable as a clinician to recommend [devices like Sensi] to my patients”— Ihab Hajjar, Neurologist studying AI-based dementia detection
Sensi has not received clearance from the US Food and Drug Administration for those cognitive claims, though Gubes said the company has initiated the process. The gap between marketing language and regulatory standing is a familiar pattern in health-adjacent AI, where consumer deployment routinely runs ahead of clinical validation.
The economic backdrop explains the demand. A private room in a US nursing facility averages over $108,000 a year — roughly double the median income for households over 65 — and nearly 1 in 6 residents deplete their savings entirely, spending down assets until they qualify for Medicaid. Staying home is, for most families, the only affordable option, and the caregiver shortage makes it harder every year. One estimate puts the gap at 9 million caregiver positions to be filled before 2031.
Steve Kamau, a coordinator at Husky Senior Care, the agency working with the Seattle family, described mixed results. In one case, Sensi caught the sound of a senior falling on the way to the toilet and his cries for help; Kamau called 911 and the man was helped off the floor. In another, an early cough flag may have caught an illness sooner. The same system has also mistaken a dropped remote for a fallen senior.
“Your dad is his own decisionmaker, right?”— Steve Kamau, Coordinator at Husky Senior Care
Sensi's investor pitch is more direct than its consumer one. The real customers are the agencies, and the company sells them on retention and billable-hour growth — metrics that translate cleanly to recurring revenue regardless of whether the resident understands they are being recorded. The 86-year-old father in the WIRED story did not remember being told the device was transcribing his conversations; told after the fact, he called it weird but accepted it.
The dignity-of-risk question — first articulated by disability rights activist Robert Perske — is whether smothering someone in safety strips them of personhood. Devices like Sensi are pitched as the resolution: security without physical constraints, surveillance without institutional walls. Whether that's the resolution or just a different trade-off depends on who is reading the transcripts and what the agency does with the billable-hour uplift.
Elder-care AI is on track to become a default layer of home health, not a premium add-on, and Sensi's 80% network penetration is the leading indicator. The market opportunity is real — the caregiver shortage is structural, the cost of institutional care is punishing, and families are desperate for any tool that lets a parent stay home. The unresolved question for the category is whether companies like Sensi can produce the clinical evidence to back the cognitive-decline claims before regulators or a high-profile false negative force the issue. A 90% accuracy rate sounds strong until it's the 10% that matters, and the FDA process Gubes says is underway will determine whether Sensi gets to keep selling the cognitive story or has to retreat to the narrower, defensible pitch of fall detection.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
