A human medical scribe costs about $41,000 a year in salary alone, most of them landing between roughly $33,000 and $49,000 before benefits, recruiting, and training. An AI scribe costs about $79 a month, near $948 a year. So the AI runs roughly a fortieth of the salary. That’s the headline gap. But salary is only part of it. Turnover is the bigger bill, because most human scribes are pre-med students who leave for school within a year or two, and you pay to recruit and train the next one every time.
This is a fit-and-cost page, not a “humans are obsolete” pitch. A good scribe does things software won’t. Below: the salary numbers (sourced), the turnover problem nobody puts on the invoice, the per-note math, and when a human scribe is genuinely the better call.
Key takeaways
- A US medical scribe averages about $41,000 a year in salary (Salary.com, June 2026), with most between roughly $33,000 and $49,000, before payroll taxes, benefits, recruiting, and training.
- Practice Copilot starts at $79 per clinician per month (annual Assist), about $948 a year, roughly a fortieth of the salary line.
- Scribing is largely a pre-med stepping-stone job, so practices re-hire and re-train often. That recurring cost never shows on the salary slip.
- A human scribe still wins when you need a person who does more than document. For the note alone, the AI is far cheaper and doesn’t quit.
Average US medical-scribe salary, before benefits and turnover (Salary.com, June 2026)
Our published AI scribe entry price per clinician, annual Assist (~$948/yr)
To review the AI draft note after the visit ends
What does a human medical scribe actually cost?
Start with the number you can see. Per Salary.com’s June 2026 benchmark, the average US medical scribe earns about $41,000 a year, with the middle of the range running roughly $37,000 to $45,000 and the wider band landing between about $33,000 and $49,000. That’s salary. On top of it sit payroll taxes, benefits, workspace, scheduling, and the cost of recruiting in the first place.
Then there’s the part the salary line hides. For most people who do it, scribing is a temporary job, clinical hours on the way to medical or PA school. Scribe-company reviews say so plainly: turnover reads high because the workforce is college students passing through. So you don’t hire a scribe once. You hire, train for weeks while they shadow, get a few productive months, and start over when they leave for school. That training time is real money. It eats your clinician’s attention, drags productivity during ramp, and the recruiting cycle comes around again next year.
We’re not going to put a precise turnover percentage on the page, because the figures floating around aren’t from a source we’d stake a claim on. What we’ll say is verifiable: the workforce is transient by design, and a tool that doesn’t quit removes the re-hiring cycle entirely. If that cycle is wearing on your practice, book a demo and see what the note-only job looks like without it.
AI scribe versus human scribe: the per-clinician cost gap
Put the two side by side for one clinician for one year.
| Human scribe | AI Medical Scribe by Patient Square | |
|---|---|---|
| Annual cost | ~$41,000 salary (range ~$33k-$49k), plus benefits and taxes | ~$948 ($79/mo, annual Assist) |
| Recruiting and training | Recurring, every time one leaves | None |
| Turnover risk | High; mostly transient pre-med staff | None; software doesn’t quit |
| Ramp time | Weeks of shadowing | Same-day; start in a trial |
| What it produces | Notes, plus flexible in-room help | A structured note, ICD-10 suggestions, a prescription draft |
| Scales to a group | Linearly; you hire one per clinician | Per seat, flat, no hiring |
The salary alone is about forty times the AI subscription, and even that undercounts the human cost. It leaves out benefits, taxes, and the re-hiring loop. None of this says the human does nothing. It’s a question of what you’re paying for. If your scribe’s whole job is writing the note, you’re paying $41,000 for documentation you can get for $1,068.
The per-note math, both ways
Cost per visit makes the gap concrete. Say a clinician sees 20 patients a day, 20 days a month, 400 visits.
| Human scribe (~$41k/yr) | AI scribe ($79/mo) | |
|---|---|---|
| Annual cost | ~$41,000 (salary only) | ~$948 |
| Visits per year | ~4,800 | ~4,800 |
| Cost per visit | ~$8.54 | ~$0.20 |
Roughly $8.54 a visit against about 20 cents. Both look small next to the value of a clinician’s evening, which is the comparison that actually matters. Nobody’s saying a human scribe is expensive in the abstract. But for the narrow job of turning a visit into a note, you’re paying about forty times more, and carrying turnover risk for the privilege.
What an AI scribe gives you that a human doesn’t
The cost case is one-sided. Capability isn’t, so let’s be fair about it.
The AI Medical Scribe is one module inside Practice Copilot: it listens during the visit and hands back a structured SOAP note, ICD-10 suggestions, and a prescription draft, ready to review and sign about two minutes after the visit. The prescription draft mirrors the plan you discussed in the visit; it’s a starting point you review and sign, not a decision the software makes. Screening it for interactions and dosing stays your job, same as it is with a human scribe typing at the keyboard. The scribe saves the writing, not the clinical judgment.
It also doesn’t have an off day or call in sick, and it won’t leave for medical school in August. Audio is processed in memory and discarded the moment the note drafts, so there’s no recording sitting in a file. The notes belong to your practice, exportable or deletable anytime. The full posture is on the security page. For the wider field of AI tools, our best AI medical scribes comparison is the honest roundup, and the AI scribe versus transcription service page covers the other human-labor option.
When a human scribe is the better choice (said plainly)
Here’s the verdict, no hedging.
Keep a human scribe if you need a person who does more than document. A scribe who can take a verbal order to staff, chase a pending result, prep a room, or handle the unscripted parts of a visit is doing a job an AI won’t. If “scribe” in your practice means “flexible clinical assistant,” the documentation tool replaces one slice of that. Not the whole role. And a teaching practice that wants to give pre-med students real clinical hours has a second, legitimate reason to keep humans: the turnover you’d otherwise complain about is the point.
Go with the AI if the job is the note. When your scribe’s day is mostly typing the visit into the chart, you’re paying roughly forty times more than you need to, plus a recruiting cycle, for documentation a tool now does in two minutes. That’s a lot of practices, honestly. They hired scribes purely to escape pajama time.
So it comes down to scope. A human can do more. An AI does the one thing far cheaper and never leaves.
How to decide in a week
- Write down what your scribe actually does all day. If it’s 90% documentation, the math above is your math.
- Add the hidden costs to the salary before you compare: benefits, taxes, and the re-hiring cycle. The $41k line is the floor, not the total.
- Match the output. Note plus ICD-10 suggestions plus a checked prescription draft, so you’re comparing the documentation job like for like.
- Settle the audio question. A human scribe keeps no recording, so ask any AI vendor the same. Ours is never stored.
- Trial on real visits. The only honest test is your own clinic day, not a demo.
The price ladder with no asterisks is on the pricing page, and the security receipts are on the security page. Book a short demo to see a note draft itself against your own visit type, then run the 7-day trial on a real clinic week. If your scribe’s job is the note, the AI does it for a fortieth of the cost and doesn’t quit in August.