Finding a primary care doctor has become a test of patience (and phone stamina) for many people, even in places known for having a strong healthcare system. In Massachusetts, the situation is reaching a particularly fragile point: the primary care workforce is shrinking faster than in most states, and patients are feeling it firsthand—sometimes hearing something as surreal as “we can book you in a year and a half… or in two years.”
Against that backdrop, Mass General Brigham (MGB), the state’s largest hospital network, has started leaning on artificial intelligence with a program called Care Connect. The goal is to offer fast access to telemedicine appointments and relieve a primary care system that has been under strain for years. Is this the kind of upgrade the system needed, or a stopgap that covers cracks without fixing the foundation? Do you see now why major companies are starting to put AI data centers in space?
What Care Connect is, and how AI-assisted care works
Care Connect is an MGB program launched in September that combines an initial interaction with an AI agent and—when appropriate—follow-up care with a doctor via video visit. The typical experience starts in an app: the patient requests an appointment and spends a few minutes chatting with the AI to explain what’s going on and why they want to see a clinician. The tool then produces a summary of that conversation and sends it to the doctor, along with a suggested diagnosis and a potential treatment plan, so the clinician can make decisions with more context from minute one.
One of the clearest examples is Tammy MacDonald, a director at an adult education center near Boston. After her doctor died suddenly, she found herself in a situation that affects about 17% of adults in the U.S.: not having an assigned primary care provider. Needing to renew blood pressure medication and arrange follow-up after a breast cancer scare, she called ten nearby practices with no luck; no one was taking new patients, and some offered waits of a year and a half or two years. When she received a letter from MGB saying there were no in-person openings available, she noticed the Care Connect link, booked a telemedicine appointment, and after about a ten-minute conversation with the AI, she got a video visit with a doctor within a day or two—an impossible contrast to ignore.

MGB’s pitch is straightforward: 24/7 availability for common issues, with doctors working remotely from across the United States. The program currently has 12 doctors logging in to see patients, and the idea is for the tool to act as a fast entry point—similar to an urgent-care service for minor issues—when the real failure is the system’s inability to offer primary care appointments within a reasonable timeframe.
What it’s for (and where it falls short)
MGB describes Care Connect as useful for common concerns: colds, nausea, rashes, sprains, and other typical “something’s wrong and I need guidance now” situations, as well as mild to moderate mental health concerns and some matters related to chronic conditions. According to the program lead at MGB, Dr. Helen Ireland, it isn’t designed for emergencies and it doesn’t replace a physical exam: if tests, labs, or imaging are needed, the patient is referred to clinics or laboratories within the network.
The nuance here is what often separates tech promise from cautious use. For Dr. Steven Lin, head of primary care at Stanford and founder of an applied AI health research group, the safest use today is for immediate, straightforward issues: upper respiratory infections, urinary tract infections, musculoskeletal injuries, rashes. By contrast, for people with multiple chronic conditions—say, hypertension and diabetes—or with particularly serious illnesses such as heart disease or cancer, the real differentiator remains a clinician who sees you regularly and builds a sustained relationship, with context and meaningful follow-up.
That’s where the limits of a chatbot show up, no matter how “smart” it seems. Critics point out that AI can miss important details when conditions overlap, and that it has no real way to understand very human barriers: whether a patient can afford follow-up, whether they can travel, or what family and caregiving dynamics shape treatment. That kind of information—the kind that doesn’t fit neatly into a form—is precisely what traditional primary care often uncovers over time.
Still, the program fits a theme that has echoed through health tech for years: a safe, supervised alternative is better than pushing people toward hospital emergency departments simply because they can’t access care. Lin puts it bluntly: if the care can be delivered safely, it’s better than getting none at all.

The primary care crisis and the debate: bridge or Band-Aid?
Care Connect didn’t appear out of nowhere—it’s a response to a familiar mix: staff shortages, administrative burden, and burnout. In primary care, many doctors—pediatricians, internists, and other continuity-focused specialties—earn on average 30% to 50% less than specialists such as surgeons, cardiologists, or anesthesiologists, while their workload has been rising for at least two decades. The pattern is the same: complex visits during the day, then charts, records, messages, and bureaucracy in the late afternoon and evening. If that sounds like “permanent multitasking mode,” it’s not by accident.
Within MGB, when MacDonald signed up in the app, there were 15,000 patients without a primary care doctor in the system, and the number has grown as clinicians have moved to competing networks. Dr. Madhuri Rao, a primary care physician at an MGB practice, has voiced frustration with system leadership: she believes specialties are prioritized while primary care is left behind, even though it’s the foundation that supports everything else. At the same time, part of the workforce is involved in unionization efforts and is calling for structural changes, including better pay.
MGB has promised to invest $400 million over five years into primary care services, a figure that includes the multi-year contract with Care Connect. MGB’s chief operating officer, Dr. Ron Walls, frames the program as one piece of a broader strategy: retain doctors, recruit new ones, strengthen support, and roll out other AI tools. Among them is one that can transcribe doctor–patient conversations during visits, though not all clinicians are adopting it: Rao, for instance, raises concerns about privacy, potential data leaks, and whether those conversations could be used to train future generations of models.
The underlying worry, shared by some doctors, is that a solution designed to fill a gap could end up normalizing less in-person access. Dr. Michael Barnett, an MGB internist also tied to the union push, describes it as a “Band-Aid” for a broken system: useful for covering things up, questionable as a long-term fix. The irony is that, as in so many automation stories, the technology arrives to relieve pressure… while simultaneously opening a negotiation over what kind of care will be considered “standard” tomorrow.
In the meantime, the program is growing. By mid-December, each Care Connect doctor was seeing 40 to 50 patients a day, and MGB plans to make it available by February to insured residents of Massachusetts and New Hampshire, expanding staffing as needed. Patients can use it as an occasional service or even choose one of the remote doctors as a permanent provider, in a model that bets on always-on availability for those who prefer it.
The technology platform behind Care Connect comes from K Health, whose CEO, Allon Bloch, argues that solving the U.S. issues of cost, quality, and access has to start with primary care and be supported by technology and AI. K Health also works with networks such as Mayo Clinic and Cedars-Sinai Medical Center. In a small study funded by the company, Cedars-Sinai researchers compared hundreds of AI vs. physician diagnostic and treatment recommendations: the AI performed slightly better at spotting critical “red flags” and aligning with clinical guidelines, while doctors stood out in tailoring treatment as the conversation progressed.
For patients stuck on impossible waitlists, what matters isn’t who “wins” in a paper—it’s being able to speak to a clinician in time. MacDonald has, in fact, used Care Connect several more times: in some cases she ended up speaking with a remote doctor; in another, for travel-related vaccines, she only interacted with the chatbot before going to a travel clinic. The practical takeaway is simple: if you can’t get an in-person appointment, a telehealth route with AI-assisted triage can give you a plan—and a bit of peace of mind—at least until that primary care doctor finally appears, the one we all keep searching for as if they were a device sold out on launch day.
Source: NPR / WBUR / KFF Health News

