There is a peculiar moment in a virtual doctor’s visit when the technology disappears. The camera stops feeling like a camera. The patient begins describing the cough, the panic, the baby who will not latch. A clinician leans a little closer to the screen. Medicine resumes. The trick is that somebody had to arrange for the right licensed person to be on the other side, in the right state, at the right hour, with the right record and the authority to do something useful.
That somebody is often Amwell Medical Group, or AMG. It is the physician-owned practice tucked behind Amwell’s software: a national network marketed at more than 2,300 clinicians, operating around the clock in all 50 states and Washington, D.C. It covers urgent care, primary care, therapy, psychiatry, nutrition, lactation and a widening shelf of specialties. The patients may arrive through a health plan, a hospital, an employer benefit or Amwell’s consumer service. What they see is a face in a rectangle. What makes the rectangle credible is everything outside it.
The first failure was the room
When pediatrician Peter Antall started Online Care Group in 2012, telemedicine still sounded like two institutional rooms connected by expensive equipment. The hardware had improved, but the assumption survived: remote medicine was a special event, staged in a special place. Antall’s reversal was wonderfully plain. A visit could happen from home or work. A clinician could use a laptop, tablet or phone. The scarce thing was no longer the video connection. It was an organized medical practice.
Online Care Group - later presented as Amwell Medical Group - was built as a practice rather than a casual directory of people with medical licenses. By 2015 it had more than 600 board-certified physicians, dietitians and psychotherapists in 48 states. It ran clinical leadership, credentialing, quality programs, virtual grand rounds and training. Antall, who became American Well’s first chief medical officer, argued that clinicians should devote meaningful working time to telehealth rather than pop in for a few spare consultations.
“Ten years from now it won’t even be called telehealth.”Peter Antall, 2014
He was directionally right. The interesting question is no longer whether a physician can talk through a screen. It is whether an organization can make the screen behave like an accountable front door to care. That requires licensing, credential checks, prescribing rules, documentation, billing, routing and escalation. It also requires an unfashionable skill AMG calls “webside” manner: knowing how to look, listen and reassure when eye contact is really an arrangement between a lens and two displays.
An overflow valve for American healthcare
AMG’s most revealing customer is not the person with a sore throat. It is the hospital that has plenty of clinicians at 2 p.m. and not enough at 2 a.m. Health systems use the group to launch a virtual program, cover evenings and weekends, handle seasonal surges or add a specialty that is locally scarce. Rural and community hospitals may use remote clinicians when there is no local psychiatrist or stroke specialist at all. Health plans use AMG to offer consistent nationwide coverage to members whose zip codes otherwise dictate their options.
This creates a counterintuitive measure of success. AMG can do its job well and become less visible. A client may begin by leaning heavily on its network, then migrate ordinary daytime care to local clinicians and keep AMG as the overflow valve. The practice does not need to replace the hospital. It needs to make the hospital’s staffing graph less jagged.
The menu follows the same logic. Urgent care catches the fever that cannot wait until morning. Psychiatry extends scarce medication-management capacity. Therapy, nutrition and lactation suit the screen because conversation and continuity carry much of the clinical value. Virtual primary care attempts the harder task: turning a string of convenient encounters into an ongoing relationship, with prevention, chronic-condition support and referrals stitched together.
The business is the bench
Amwell sells the platform. AMG sells clinical work on a fee-for-service basis. That separation matters legally and commercially. The professional entities are physician-owned and retain responsibility for medical decisions; Amwell provides administrative machinery such as scheduling, billing and technology through management agreements. In a country where many states restrict the corporate practice of medicine, the organizational chart is part of the product.
For patients, the price is less tidy. It depends on the service, insurance and employer benefits. A consumer may see a fixed self-pay charge, a copay or no charge at all, and the amount is shown in the booking flow. A published 2024 employer rate sheet listed $67 for a general urgent-care consultation and $263 for an initial psychiatry visit, but those figures are examples, not a universal AMG price list. For enterprise buyers, the economics sit in negotiated service contracts.
The parent company now reports AMG as a visible revenue engine. In the second quarter of 2026, Amwell recorded $24.4 million in AMG visit revenue and maintained full-year guidance of 1.32 million to 1.37 million AMG visits. A different metric is more human: Amwell reported an NPS of 81 for care delivered by AMG in 2025, based on 365,968 completed-visit experiences.
What the Apple Watch could not do
The Apple Heart Study supplied a neat demonstration of AMG’s role. More than 419,000 people enrolled to test whether an Apple Watch could identify an irregular pulse suggestive of atrial fibrillation. The watch could notice a pattern. It could not explain that pattern to an anxious person or decide what should happen next. AMG physicians helped provide protocol-driven telehealth follow-up and direct participants toward urgent care when appropriate. The sensor found a signal; the medical group supplied judgment.
That is the useful distinction in a market crowded with video tools, automated questionnaires and condition-specific apps. AMG is not differentiated because it invented the call. Its advantage is the bench behind the call: clinicians across disciplines and state lines, a credentialing operation, clinical oversight and the ability to turn capacity up where a customer’s local network thins out. Competitors such as Teladoc Health, Included Health and MDLive can make similar national claims. Hospitals can also build their own virtual teams. AMG earns its place when buying flexible coverage is faster or more reliable than assembling it.
The parts worth stealing
The copyable lesson is not “put doctors on Zoom.” It is to design the operating system around the encounter. AMG’s method reduces the number of chores a clinician must improvise and the number of gaps a patient must cross alone.
- Start with the capacity hole. Nights, weekends, state coverage and scarce specialties make cleaner entry points than trying to replace every local visit.
- Train for the medium. Virtual observation, camera presence, documentation and escalation are clinical skills, not software tips.
- Centralize the dull work. Credentialing, eligibility, billing and scheduling are precisely the tasks that let clinicians concentrate on care.
- Plan the handoff. A virtual service becomes more useful when it can refer locally, share records and yield to in-person care without pretending the screen can do everything.
- Measure the practice. Wait times, throughput, patient experience and clinical review turn a roster into an accountable group.
The screen has edges. It is a poor substitute for emergency response, procedures, imaging, many hands-on examinations and care that depends on a local relationship but cannot exchange records. Weak broadband, mismatched state licenses and a fragmented referral network can turn convenience into another dead end. AMG works best as a connected layer in a larger system - not as a claim that every room in medicine has become unnecessary.
That boundary may explain the company better than any slogan. Amwell Medical Group did not abolish the hospital. It removed the hallway from a set of encounters that never truly needed one, then spent fourteen years building the clinical scaffolding that keeps those encounters from floating away.