Primary care does not lack technology.
Doctors have electronic health records, patient portals, automated reminders, telehealth platforms, digital intake forms, remote monitoring tools, and a rapidly expanding catalog of artificial intelligence (AI) products. AI can answer the phone, draft a medical note, summarize a chart, suggest a refill, and route a patient message.
Yet a patient with a new symptom may still wait weeks to see a doctor.
In AMN Healthcare's 2025 survey of 1,391 physician offices across 15 large metropolitan areas, the average wait for a new-patient appointment reached 31 days. Family medicine averaged 23.5 days. Those figures were 19% higher overall than in 2022.[1]
This creates an awkward possibility: healthcare may become much better at processing patients without becoming much better at caring for them when they need help.
The more important breakthrough may be simpler. Let patients reach a responsible clinician sooner.
The Problem Is Not Only What Happens During the Visit
Most conversations about healthcare AI begin after access has already been granted. The patient has secured an appointment. The doctor has opened the chart. The microphone is running. The algorithm can now summarize the conversation or retrieve a relevant piece of history.
Those tools can make the encounter better. They do not solve the three weeks before it.
A patient who wakes up with painful urination does not primarily need a more elegant visit note. A parent whose child has a 103-degree fever does not need a better chatbot handoff. A patient who becomes dizzy after a blood-pressure medication change needs to know whether to stop, adjust, continue, schedule an examination, or seek urgent care.
The first question is not whether the practice uses AI. It is whether someone with clinical responsibility can respond in time.
This distinction matters because access is part of care, not an administrative feature sitting outside it. A same-day answer can determine whether a problem stays in primary care, moves to urgent care, or becomes an emergency-department visit. A three-week delay can make even an excellent physician irrelevant to the decision the patient has to make today.
Our earlier explanation of why primary care visits feel so short focused on what high patient volumes and administrative work do inside the appointment. The access problem begins before the clock starts.
Baker Health Built the Practice Around the Wait
Zeyad Baker, MD, made this argument directly during an August 2026 episode of the Lifers podcast. Baker is a pediatrician and the founder of Baker Health, a hybrid membership practice with locations across New York and New Jersey.[2]
His position was blunt: if you want to improve healthcare now, start by getting patients in immediately. Technology should support that goal instead of becoming the goal itself.[2]
Baker Health's public membership materials promise 24/7 doctor chat, same-day appointments seven days a week, and in-person access 365 days a year. The practice charges a $200 annual membership fee while continuing to bill insurance for covered medical care.[3] Its website also markets offices without traditional waiting rooms and appointments that begin on time.[3]
Those are company-stated service promises, not independent evidence that the model improves outcomes. Baker also made several utilization and operational claims during a sponsor-affiliated interview. They should be treated as founder-reported until supported by independent data.[2]
The useful part of the example is the operating choice. Baker Health did not begin with a diagnostic algorithm. It built the practice around answering messages, keeping appointment capacity available, and making the physical office work differently.
Patients comparing practices in the New York City metro can evaluate those access promises against other concierge, direct primary care (DPC), and traditional practices. The relevant questions are concrete: Who answers after hours? Is the response from a physician or a shared team? Does same-day access mean an in-person examination, a video visit, or a text exchange? What happens when the patient's usual doctor is away?
AI Is Most Valuable When It Creates Clinical Capacity
This is not an argument against AI in primary care. It is an argument for judging AI by what it gives back.
Ambient AI scribes offer a good example. They listen during a visit and create a draft note for the clinician to review. A 2026 study of clinicians using one ambient system found modest reductions in time spent documenting during appointments and after hours. The researchers cautioned that the improvements were moderate and that the observational design could not establish every effect as causal.[4]
That is still valuable. If a scribe returns ten or twenty minutes to a doctor's day, the practice can use that time in several ways:
Add an urgent appointment slot.
Return patient messages sooner.
Review outside records before a complicated visit.
Call a patient after an emergency-department discharge.
Finish notes earlier so the physician is less likely to leave the profession.
The technology matters because of the capacity it may create. If the recovered time simply allows the practice to add more short appointments to an already overloaded schedule, the patient may notice little improvement.
The same test applies elsewhere. AI scheduling software is useful if it finds a real opening sooner. A chart summary is useful if it helps the doctor act safely during a same-day visit. A voice agent is useful if it connects the patient to the right person instead of creating another automated barrier.
As NextMD's guide to AI scribes and medical-visit recordings explains, automation can give a clinician more eye contact and less typing. It cannot perform the physical examination, take responsibility for the diagnosis, or create appointment capacity by itself.
Fast Access Is Not the Same as Good Care
Speed alone is not enough either.
An unfamiliar clinician on a five-minute video call may be available immediately but know nothing about the patient's previous reactions, family history, specialist plan, or usual blood pressure. A retail clinic may treat today's sore throat without noticing that it is the fourth infection in three months. A chatbot can produce a quick answer without accepting responsibility for what happens next.
The better target is timely access plus continuity.
A 2026 observational study followed 100,450 adults across 48 general practices in the Netherlands. Compared with people registered at a practice for five years or less, those with longer relationships had 9% to 21% lower odds of an urgent hospital admission. Longer registration was also associated with lower hospital costs.[5]
The study does not prove that staying with one practice caused the difference. It took place in the Dutch healthcare system, where general practitioners play a different gatekeeping role than many American primary care doctors. The researchers adjusted for age, sex, income, comorbidities, and other factors, but unmeasured differences may remain.[5]
Still, it points toward the combination healthcare should try to preserve: a practice that can respond now and remember what happened before.
This is where some virtual-care products remain incomplete. They can provide fast access to a clinician, but not necessarily the same clinician or a stable team. Traditional practices may provide a long-term relationship, but not timely access. The strongest primary care models try to deliver both.
Membership Medicine Changes the Capacity Equation
Concierge and DPC practices commonly make access their central promise because their economics allow them to carry smaller patient panels.
Based on NextMD directory analysis and category research, traditional primary care physicians often care for 2,000 to 2,500 patients. DPC physicians generally care for up to 800, while concierge physicians often care for fewer than 300. DPC memberships typically cost $50 to $200 per month. Concierge medicine generally costs $3,000 to over $40,000 per year, depending on the service level.
Fewer patients do not automatically make a doctor more skilled. The membership fee does not guarantee a correct diagnosis, a same-day appointment, or a good relationship. It creates the possibility of more available time. Each practice decides how to use that capacity.
The American Academy of Family Physicians' 2024 DPC survey found that 98% of participating DPC practices offered same-day appointments, and the same share offered phone or text consultations and telemedicine.[6] Because this was a self-reported survey of DPC physicians, it describes participating practices rather than proving that every patient always received same-day care.
Patients considering the model should read the benefits and trade-offs of direct primary care, then verify the access policy of the specific practice. "Available 24/7" might mean the personal physician, a rotating doctor, a nurse line, or a portal monitored the next morning.
Five Questions Matter More Than Whether the Practice Uses AI
When comparing primary care practices, ask:
How soon can I be seen for a new problem? Ask separately about office visits, video visits, and messages.
Who responds after hours? Find out whether it is your doctor, another physician, a nurse, or an outside service.
Will I usually see the same physician? Fast access loses part of its value when every encounter starts from zero.
What happens after an emergency or specialist visit? Timely follow-up is one of the places continuity matters most.
How does technology improve my access? Ask whether AI creates more appointment capacity, faster replies, better preparation, or only more automation.
AI belongs in the fifth question, not the first.
The Better Future Is More Human Care, Delivered Faster
Primary care needs better technology. Doctors should not spend evenings reconstructing conversations into billing-ready notes. Staff should not manually retype every referral, hunt through every fax, or answer the same scheduling question hundreds of times.
The best use of AI is to move that work away from the clinician so the clinician can move closer to the patient.
Success should be visible without a product demonstration. The patient gets an answer today. The doctor has read the history. The appointment starts on time. The plan accounts for the patient's other medications and conditions. Someone follows up, and the same practice is still there the next time the patient needs help.
That may sound less futuristic than an AI doctor. It would be a much bigger breakthrough.
Find Primary Care Built Around Access
NextMD helps patients compare physician-led concierge and DPC practices by location, model, pricing, services, and doctor credentials. Use the NextMD search to find practices near you, then ask each one exactly what same-day and after-hours access mean.
This article provides general educational information and is not medical advice. New or severe symptoms may require urgent or emergency evaluation. Call 911 for a medical emergency.
Sources
AMN Healthcare. (2025). 2025 Survey of Physician Appointment Wait Times and Medicare and Medicaid Acceptance Rates. Read the full AMN Healthcare report
Farr, C., and Baker, Z. (2026). How Do You Make Primary Care Cool Again? Lifers. The episode was sponsored by Baker Health; operational and utilization claims are founder-stated. Watch the Zeyad Baker interview
Baker Health. (2024–2026). Membership Policy and Member Services. Read the Baker Health membership policy and review its access commitments
Husa, R. A., et al. (2026). Ambient Artificial Intelligence Use and Clinician Documentation Burden, Productivity, and Efficiency. JAMA Network Open. Read the open-access study
te Winkel, M. T., et al. (2026). Association of General Practice Continuity With Hospital Admissions and Costs: A Retrospective Study. Annals of Family Medicine. Find the study on PubMed
American Academy of Family Physicians. (2024). 2024 Direct Primary Care Data Brief. Read the AAFP data brief