Here is a fact about American healthcare that is both obvious and, somehow, mostly ignored: when a person has dementia, the most important member of their care team usually has no medical training, no badge, and no salary. It is the spouse. It is the adult child. It is the person who notices, at 7 a.m., that Mom did not eat dinner and seems confused in a new way.
Ceresti Health, a company in Carlsbad, California, is built entirely around that person. The pitch is almost aggressively simple. People living with dementia account for roughly 27% of all Medicare hospitalizations, and a large share of those hospitalizations are avoidable - a urinary tract infection caught late, a medication quietly skipped, a fall that follows a week of subtle decline. The clinical system is not in the house to catch these things. The family caregiver is. So Ceresti equips and coaches the family caregiver to catch them.
The mechanics are deliberately unglamorous. Ceresti analyzes claims data from a health plan to find members with Alzheimer's or another dementia, runs a predictive model to decide who to reach first, and - after getting consent - ships the family caregiver a tablet. Not an app to download. A physical, cell-enabled tablet, no WiFi required, pre-configured to show only the Ceresti program. Then it assigns a human being, a "care navigator," who calls the caregiver every week.
That is the product. A tablet and a coach. What makes it a business rather than a charity is the second half of the sentence: it saves money, and Ceresti can prove it. In a matched-cohort study, the company compared 131 Medicare Advantage members with dementia whose caregivers enrolled against 393 similar members who did not. The enrolled group generated statistically significant savings - Ceresti cites figures in the range of $6,800 to $7,400 in medical costs per patient per year, driven by more than a 50% reduction in avoidable hospitalizations. The outcomes have been independently reviewed by the Validation Institute.
Now, "we can save you money" is the single most-repeated sentence in all of digital health, and it is usually followed by a slide deck rather than a check. Ceresti's twist is that it is willing to put the promise in the contract. The company offers health plans a cost-savings guarantee - a model where, roughly speaking, if the caregivers it coaches do not reduce hospitalizations, the plan does not pay full freight. In an industry that runs on optimistic projections, offering to be measured on results is a genuinely different posture. It is also a good way to force yourself to have real data.
There is a nice piece of timing underneath all this. Ceresti was founded in 2013, by Dirk Soenksen, Mark Wrenn, and Kevin Liang, with about $4 million raised from friends and family. For roughly a decade it worked on an unfashionable problem with no obvious way for Medicare to pay for the solution. Then, in 2024, the Centers for Medicare & Medicaid Services launched the GUIDE model - Guiding an Improved Dementia Experience - an alternative payment model designed to fund exactly the kind of caregiver support Ceresti had already spent ten years building. The company did not have to pivot into the GUIDE model. It had, in effect, been waiting for it.
That is the boring-then-sudden shape a lot of durable companies have. In August 2025, Ceresti raised $11.7 million, part of a planned $15 million round, to scale the GUIDE-powered platform. The money is not being spent to invent the product - the product already works - but to enroll far more caregivers and hire far more coaches. Around the same time, two of the largest home-care networks in the country, Interim HealthCare and Griswold, both signed partnerships to bring Ceresti's program to their patients nationwide. When the companies that are already inside patients' homes come to you for your caregiver platform, that is a reasonably strong market signal.
There is also a quieter, more interesting detail. In 2024 Ceresti received a $478,107 NIH grant to build a Spanish-language version of its program, aimed at Hispanic and Latino families, who carry a disproportionate share of the dementia-caregiving burden. Building carefully for the population that usually gets left out of the first pilot is both the decent thing to do and, not coincidentally, where a lot of the real market is. Ceresti already staffs bilingual coaches; the grant is about doing it properly rather than as an afterthought.
The thing to understand about Ceresti Health is that it is not really a technology company that happens to work in dementia. It is a services company - a "technology-enabled service," in the sector's phrase - that uses just enough technology to make a human relationship scale. The tablet is dumb on purpose. The coaching is the intervention. The claims data and the AI-driven risk models exist to point the coaches at the right families and to prove, afterward, that the whole thing worked. Everything is arranged around a single unfashionable bet: that the most valuable and least utilized asset in dementia care was standing in the kitchen the entire time.