Last November, the FDA granted RecovryAI Breakthrough Device Designation for our Virtual Care Assistants. We sat on the news. Breakthrough Device Designation accelerates FDA engagement, which means it accelerates everything else: the pace of evidence gathering, trial site onboarding, patient enrollment, quality management infrastructure, team recruiting, pre-submission interactions, and the painstaking work of adjudicating every AI-patient conversation against clinical ground truth. The pace was already frenetic. BDD raised the cadence across the board; keeping our heads down was the only way to keep up.
Today, we're looking up to say out loud what's been underway quietly for about three years.
RecovryAI is developing Virtual Care Assistants: physician-prescribed AI designed to accompany patients through post-procedure recovery at home. The category we're building occupies a space the healthcare system hasn't yet found a way to fill: regulated, patient-facing clinical AI.
The Structural Collision Ahead
The United States is heading toward a structural collision between healthcare demand and the clinical workforce available to meet it. The Association of American Medical Colleges projects a shortage of up to 86,000 physicians by 2036, driven by an aging population, rising chronic disease burden, and a physician pipeline that can't expand commensurately. One in five practicing physicians is already 65 or older. The math compounds: as the physician workforce ages out, the patients they serve are aging in, requiring two to three times the clinical services of younger populations. Training more physicians matters, but it doesn't resolve the imbalance. The system will require a different architecture for how clinical supervision is delivered.
Where the Collision Shows Up
Post-operative recovery is one of the most legible places where that collision plays out, where the pressure between rising demand and a shrinking workforce shows up in quantifiable outcomes. More than 80 percent of surgeries in the United States are now performed on a same-day basis. Patients go home the same day they go under. The most vulnerable interval isn't the surgery; it's what follows. Seventeen percent of post-surgical complications occur within the first 24 hours after discharge; 60 percent occur within the first three days. The first 48 hours carry the highest rates of morbidity and mortality.
That's the window in which patients are home, largely unsupervised, with discharge instructions that are frequently forgotten or misunderstood, and with care teams managing high volumes of routine patient contact that crowds out capacity for higher-acuity cases. We're starting with total joint arthroplasty, one of the highest-volume surgical procedures in the country and one of the most data-rich for measuring what recovery actually looks like, but the supervision problem it exposes isn't specific to orthopedics. In TJA alone, unplanned readmissions and post-acute complications account for the large majority of post-discharge costs within a 30-day episode of care, and 30-day readmission rates for joint replacement patients run between 3 and 6 percent nationally, with 90-day rates approaching double that. These events are predictable, measurable, and in many cases preventable with earlier visibility into how a patient is actually recovering at home.
What We're Building
We're building physician-prescribed AI that extends clinical supervision from the facility into the home — into the interval where patients are most vulnerable and least supervised. Virtual Care Assistants prescribed by a physician, guided by procedure-specific clinical protocols, and designed to accompany patients through recovery: checking in, interpreting what they report, guiding them when recovery is progressing normally, and escalating to their care team with full context when it isn't. Care teams re-engage when their judgment is actually required.
The medical reasoning framework at the heart of our Virtual Care Assistants harnesses the ambient intelligence of generative AI within a clinically constrained logic architecture, and with regulatory authorization, it's the foundation of what we believe can become an infinitely scalable answer to some of healthcare's most intractable problems. Our physician-prescribed AI extends continuous clinical supervision at a scale that human labor can't reach.
The Regulatory Path
We're pursuing FDA authorization through their De Novo process, seeking designation for patient-facing Software as a Medical Device. If authorized, the designation establishes a new device class and the special controls governing future devices in this category. Our pivotal clinical study is underway across multiple sites and will come to a close in Q2 of this year. The study will produce well over 100,000 AI-patient conversations adjudicated by a panel of orthopedic surgeons to determine concordance between human judgment and AI triage. That body of evidence will contribute to the clinical core of our FDA submission, expected later this year.
The Team
Dr. Richard Watson and I first worked together a number of years ago, when I was on the board of a nanotech company where he led science and regulatory. We became friends. When he and Dr. Marty Sellberg were completing the exit of their medical startup, a company they'd founded, built, and sold, we reconnected and began exploring the space RecovryAI now occupies. The three of us co-founded RecovryAI, and Jesse Thomas joined as our fourth co-founder and Chief Product Officer. In addition to his decades of medical practice, Dr. Watson, our Chief Science Officer, brings more than 100 medical device patents and three prior FDA submissions. Dr. Sellberg, our Chief Medical Officer, brings over 30 years of emergency medicine and still practices in the ER. Jesse brings over a decade of healthcare technology product leadership. There's more to report, and their perspectives will illuminate our public record.
This post is the beginning of that public chapter. We're three years in and just getting started. For those watching where regulated AI in patient-facing clinical care is headed, whether you're a physician, a health system leader, an investor, or a policy observer, there will surely be more to report.