The annual physical exam catches problems once a year. Your blood pressure gets measured. Labs get ordered. Maybe something worrisome turns up, maybe it doesn't. Then you leave, and for the next 364 days, you're essentially on your own.
It's a ritual so embedded in American healthcare that questioning it feels almost heretical. But that's precisely what a San Francisco startup wants to do—not by improving the annual visit, but by making it obsolete.
Prana, a company that emerged from Y Combinator's Winter 2026 cohort, launched in early February with an audacious premise: continuous AI monitoring that watches for what the founders call "clinical drift" in the long stretches between doctor appointments. Blood pressure creeping upward. Glucose trending into worrying territory. The kind of slow-motion deterioration that looks fine at any single snapshot but tells a different story when you connect the dots over weeks.
"Stop Googling your symptoms, start getting real answers," CEO Meer Patel wrote when announcing the launch. It's a pitch that resonates, if only because most of us have been there—lying awake at 2 a.m., searching WebMD for explanations that only make things worse.
What They're Actually Building
The core product isn't particularly exotic, at least not on its surface. Users connect their electronic health records (Prana taps into data from tens of thousands of healthcare organizations through a FHIR API integration), then link wearables—Fitbit, Garmin, Oura Ring, Dexcom, the usual suspects. The AI ingests this combined stream and watches for anomalies.
An elevated resting heart rate that persists for multiple nights. Sleep patterns that shift in concerning ways. Blood pressure numbers that would look unremarkable at a single appointment but show troubling momentum across a month. When something flags, users can escalate to board-certified physicians through the platform. Video consultations are listed at $49, and doctors can write prescriptions or order labs.
It's worth noting what Prana isn't, at least according to its own terms of service: a licensed medical provider. The company emphasizes that it partners with affiliated medical groups for actual diagnosis and treatment. That's a legal distinction that matters, even if the marketing sometimes blurs the lines.
The model bets on two infrastructure shifts that have quietly matured over the past few years. First, standardized health data APIs that give patients control over their records—a regulatory push that's finally bearing fruit. Second, consumer wearables that have evolved beyond glorified pedometers into devices that track clinically relevant vitals around the clock.
Whether those building blocks are actually ready for this use case? That's the experiment.
Timing the Market
Prana's launch coincides with broader industry chatter about reimagining primary care. Earlier this year, Lumeris, a value-based care organization, published a brief arguing that primary care should evolve into "continuous ambient monitoring" rather than episodic visits. Research papers have started exploring how wearable data can integrate with large language models for persistent vital monitoring.
The regulatory backdrop has shifted, too. Recent data from the Office of the National Coordinator for Health IT suggests that roughly nine in ten U.S. hospitals now enable patient API access to health records, with FHIR adoption climbing. The infrastructure, in other words, is finally catching up to the concept.
Still, Prana faces a crowded field. K Health raised $50 million in early July 2024 for its AI-assisted virtual care model. Curai Health operates a chat-first platform with clinicians in the loop. Amazon's One Medical scales membership-based care with virtual access tied to Prime. Carbon Health built remote patient monitoring capabilities through an acquisition a few years back.
What distinguishes Prana's approach—at least in theory—is the emphasis on background monitoring rather than on-demand consultation. The value proposition isn't better triage when you're already sick. It's catching problems before symptoms surface at all.
Four People, Big Ambitions
The founding team has an interesting composition. Patel, the CEO, studied biomedical engineering at Johns Hopkins and deferred a medical school acceptance at Brown. CTO Vishvam Rawal comes from applied mathematics at Berkeley with a stint in quantitative trading. Chief Health Officer Sanjit Menon is a graduating MD.
According to the company's Y Combinator profile, the team numbers four people. They're operating in beta now, with early users testing the platform. The launch announcement explicitly solicits pilot partnerships with self-insured employers—a savvy distribution strategy, given how difficult consumer health products find it to reach critical mass through direct-to-consumer channels alone.
There's a messaging tension here that sophisticated buyers will notice. The Y Combinator profile describes Prana as a "full-stack medical provider" whose AI "manages your prescriptions." The company's terms of service, by contrast, state clearly that Prana is "an AI-powered health assistant—not a licensed medical provider." It's the kind of gap that needs resolving before scaling.
The Questions Nobody's Answering Yet
Several practical details remain frustratingly vague. The $49 consultation price is visible on the website, but what about the economics of continuous monitoring itself? Is there a subscription? What's included beyond that initial visit? How does follow-up care work, and are there additional charges for prescriptions, labs, or specialist referrals?
The physician network details are similarly opaque. The company references partnerships with "board-certified physicians" and "affiliated medical groups" but doesn't specify coverage by state or typical response times beyond the obligatory "24/7" marketing claim. For a service positioning itself as a replacement for traditional primary care, those operational specifics matter considerably.
Insurance integration is notably absent. The site promotes "no insurance required," which could be a selling point or a red flag depending on your perspective. Most digital health companies eventually need payer relationships to achieve sustainable unit economics, even if they launch direct-to-consumer.
Perhaps the most fundamental question is whether continuous monitoring works at this scale. Wearable data is famously noisy. Electronic health records remain fragmented despite API standards. Detecting meaningful clinical drift from this cacophony requires not just technical sophistication but validated clinical protocols that most AI health startups are still developing, often in real time with actual patients.
One reassuring note for enterprise buyers: the privacy policy cites compliance with the HIPAA Reproductive Health Privacy Rule (a regulation that took effect in recent years) and allows users to opt out of having de-identified data used for AI training. Alert-trigger data is retained for at least seven years. These details matter when evaluating data governance.
What Happens Next
Prana is live, at least in beta form. The value proposition is clear. Infrastructure partnerships are in place. The founders bring clinical credibility and technical chops, plus Y Combinator's imprimatur. Market timing looks favorable, both in terms of available technology and industry receptiveness to AI-enabled care models.
Whether it actually works hinges on execution challenges that aren't visible from a launch announcement. How accurate is the drift detection, really? How quickly can the physician network respond when the AI escalates something? Can the economics support continuous monitoring at $49 per consultation, or will the model eventually need subscription revenue? And can a four-person team scale operations fast enough to capture market interest before better-resourced competitors copy the approach?
Those answers will emerge over the coming months as pilot programs produce data and early users provide feedback—or don't. For now, Prana represents one more data point in an ongoing argument about where primary care's center of gravity belongs. Not in the exam room. In the space between visits, where most of health actually happens.
The annual physical isn't going away anytime soon. But its relevance? That's another question entirely.
