If you run a small medical practice, you already know this number - but it's worth saying out loud: the average primary care physician spends roughly 2 hours in their electronic health record (EHR) for every 1 hour of direct patient contact.
That's not a fringe finding. The American Medical Association has documented this pattern consistently, and it's part of why more than half of physicians report symptoms of burnout in any given year. The paperwork isn't just annoying. It's eating the practice.
For large health systems, the solution has been to hire medical scribes - trained humans who sit in the room, listen to the encounter, and write the note. At a hospital with hundreds of physicians, that's economically viable. For a four-doctor family practice or a solo internist, a full-time scribe at $30,000-$40,000 per year is a different calculation entirely.
AI scribes are now making that calculation a lot simpler.
What an AI Medical Scribe Actually Does
The term "AI scribe" is being used to describe a range of tools, so let's be specific about what the current generation actually does:
-
Listens to the patient encounter. Using a microphone (your phone, a dedicated device, or an in-room tablet), the AI captures the conversation in real time or from a recording.
-
Generates a structured clinical note. The system converts the conversation into a SOAP note (Subjective, Objective, Assessment, Plan) or whatever format your practice uses - formatted for your EHR.
-
Routes the note for your review. The physician reviews, edits, and signs. The AI is doing the first draft, not the final document.
This is the key distinction: AI scribes are draft generators, not autonomous documentation systems. The physician still reads and approves everything. What changes is how much time that review takes compared to writing from scratch.
In documented implementations, practices are reporting a reduction in documentation time of 40-60% per encounter. At 20 patient visits per day, that's 1.5 to 2 hours of administrative time recaptured daily. Over a year, that's the equivalent of weeks.
The Tools That Are Worth Looking At
Three platforms have established enough real-world use in small and independent practices to be worth mentioning:
Nuance DAX Copilot (from Microsoft) integrates directly with Epic and several other major EHR systems. It's been the market leader for enterprise health systems and has been expanding into smaller practices. Pricing is subscription-based and varies by practice size - typically in the range of $300-500 per physician per month.
Suki AI has built its product specifically with physician usability in mind and has strong integration with athenahealth, which is commonly used by independent practices. Suki has published data showing an average documentation time reduction of 72% across its users.
Freed AI is the newest and most affordable entrant, built for small practices and solo physicians. It operates as a mobile app, doesn't require EHR integration (the notes are generated and copied over manually), and runs at roughly $99-149 per month per physician. For a practice not ready to commit to a full enterprise integration, it's a reasonable starting point.
What Small Practices Are Actually Saying
The practice management community has been candid about both the upside and the friction:
What's working: Time savings are real and meaningful. Physicians who were spending evenings catching up on notes report finishing documentation before leaving the office. Staff are noticing that physicians seem less rushed during visits because they're not mentally composing the note while trying to listen to the patient.
What takes adjustment: The AI doesn't always capture clinical nuance accurately on the first pass. Unusual presentations, complex medication histories, and anything that deviates from typical encounter patterns require more careful review. The editing time is still less than writing from scratch, but it's not zero.
What to watch for in the contract: Many of these platforms have data-sharing provisions that deserve attention. Patient audio and conversation data is being used (in varying forms) to train and improve these models. The HIPAA compliance question is real - most major vendors have Business Associate Agreements available, and you should require one before signing. Don't assume.
The Economics, Translated
Let's run the simplest version of the math.
A primary care physician seeing 20 patients per day, billing at an average of $150 per encounter, generates $3,000 per day in gross revenue. If documentation time can be reduced by 90 minutes per day, and that time is used to add even 2 additional patient encounters, that's $300 of additional daily revenue. Over 200 working days per year, that's $60,000 in additional annual gross revenue.
Against a tool that costs $400 per month ($4,800 per year), the ROI math is not complicated.
Even if the time savings are used for something other than additional patients - finishing on time, going home at a human hour, having energy for family - that has real value that doesn't show up in a spreadsheet but matters enormously for practice sustainability.
The HIPAA Question You Need to Answer First
Before trying any of these tools, three questions:
-
Does the vendor offer a signed Business Associate Agreement? If no, stop. If yes, get it signed before processing any patient data.
-
Where is audio stored, and for how long? Some platforms store the full audio for extended periods. Know what you're agreeing to.
-
Is the generated note subject to the same retention requirements as your other clinical documentation? Your state medical board may have guidance on this. When in doubt, treat AI-generated notes the same as any other documentation in your record.
The AAFP has published general guidance on AI documentation tools and recommends starting with a pilot on a subset of encounter types - straightforward follow-up visits, annual physicals - before expanding to complex new patient encounters.
The Bottom Line
Physician burnout is a practice sustainability problem. When a doctor is spending 10-12 hours per day working and 3 of those hours are on documentation alone, that's not sustainable, and no amount of "better work-life balance" advice fixes it without changing the underlying time equation.
AI scribes are the first tool in a long time that actually changes that equation for small and independent practices at a price point that makes sense. They're not perfect. The editing burden is real. The compliance questions need answers. But the direction is clear - and for a solo or small group practice carrying the full weight of modern healthcare documentation requirements, this is worth a serious look.
Sources: American Medical Association physician burnout and EHR burden research (ama-assn.org/practice-management); Suki AI published efficiency data (suki.ai); AAFP guidance on AI tools in primary care (aafp.org).
Dr. Renee Carter covers the business side of running a small medical practice - from technology and billing to staffing and regulatory compliance. She writes for independent physicians, group practices, and healthcare entrepreneurs.