Three minutes. That is roughly how long a physician has, at the end of a 20-patient day, to pick the billing code a complex visit actually supports. The faster code usually wins.
New data shows 22.5% of physicians now spend more than eight hours a week on the EHR after clinic hours, up from 20.9% two years earlier. The hours show up on a calendar. The undercoded chart from three minutes ago does not, and it is usually the more expensive habit.
Notiro, an AI medical scribe built around the full clinical day rather than just the note, starts the ROI conversation there. AI medical scribe ROI is usually framed as a story of well-being. It is actually a revenue story that gets told as a wellbeing one.
The Real Cost of Documentation Burnout
For a practice manager, burnout is not an abstract wellness metric. It is a staffing risk with a dollar figure attached, and the figure is large. Replacing a single physician costs between $500,000 and $1 million, once recruitment, onboarding, and lost productivity are factored in, according to the same organization’s research.
Across the system, burnout-related turnover and reduced clinical hours cost US healthcare roughly $4.6 billion a year, according to an analysis published in the Annals of Internal Medicine. A solo practice cannot absorb a six-figure departure quietly.
An 11-provider group cannot either, even if the math looks smaller per provider on paper.
The Undercoding Leak Nobody Budgets For
Burnout costs money through turnover. It also costs money every single day, through coding.
ICD-10 carries more than 70,000 diagnosis codes. CPT adds another 10,000-plus procedure codes on top of that (CMS). No physician memorizes that catalog, and under time pressure at the end of a long day, most select the code they remember rather than the code the visit actually supports.
That pattern has a name: undercoding. It is one of the most overlooked revenue leaks in small-to-mid practice, and it is rarely intentional. It is simply what happens when a 20-patient day leaves three minutes per chart for a coding decision that deserves real attention.
That holds whether the practice runs family medicine, internal medicine, or a mixed primary care panel.
This is exactly where ICD-10 and CPT coding automation earns its place in the ROI conversation, separate from how fast a note gets written. Ambient transcription only solves the front half of the visit. The back half, where revenue actually gets captured or lost, is billing accuracy.
What Recent Research Says About AI Scribe ROI
The evidence on AI medical scribe return on investment has matured considerably since 2023, when most studies were small, single-site pilots.
A 2025 UCSF study published in JAMA Network Open followed 1,565 physicians across 1.2 million ambulatory visits. AI scribe adopters generated 1.81 more relative value units per week than non-adopters, a 5.8% increase. That translates to approximately $3,044 in additional annual revenue per physician at 2025 Medicare payment rates.
Adopters also logged 0.80 more patient encounters per week, a 2.8% increase, with no rise in claim denials.
A larger five-site study spanning Mass General Brigham, Emory, UCSF, Yale New Haven, and UC Davis followed 8,581 clinicians, including 1,809 AI scribe adopters. Researchers found 13.4 fewer minutes of total EHR time and 16.0 fewer minutes of documentation time per eight scheduled patient hours. Clinicians using the tool for more than half their visits saw nearly double those reductions, though only 32% of adopters reached that usage level.
The added revenue from extra visit volume averaged $167 per clinician per month, which researchers described as real but modest on its own.
Two patterns stand out across both studies. Time savings alone produce a few hundred dollars a month in most cases, not a windfall. The larger return comes from what physicians do with the recovered hours: fewer charting nights, lower turnover risk, and codes that actually match the visit performed.
Calculating the ROI of AI Medical Documentation
A practice manager building a board case can work from a simple formula:
For a solo family physician, the coding and capacity gains already covered above clear most of a subscription cost within the first year, before undercoding correction even enters the picture. For a five-physician group, avoiding a single burnout-driven departure does that same work many times over. Add fewer denials and faster chart closure, and the formula clears breakeven well before the first renewal.
This is also why AI scribe cost savings should never be measured against subscription price in isolation. The honest comparison is the cost of doing nothing: continued pajama time, continued undercoding, and continued exposure to the next resignation letter.
Why Most AI Scribes Only Capture Half the Return
The research above measures ambient note-taking. Most AI scribe ROI studies evaluate tools that listen and write. They do not evaluate coding, because most tools on the market do not touch it.
Freed is the category’s most recognizable example: simple, mobile-first, well-funded, and limited to transcription. It writes the note well. It does not affect ICD-10 or CPT selection, meaning the undercoding leak described above remains exactly where it was before adoption.
Notiro automates the complete clinical day instead: intake before the visit, ambient scribing during it, and ICD-10 and CPT coding suggestions after it. The finished chart syncs to the EHR in one click. Note-writing speed is the commodity every scribe now offers.
The coding layer is where the second half of the ROI has been sitting unclaimed.
How Practice Size Changes the ROI Math
A solo physician’s ROI case centers on personal time and undercoding correction, since there is no staff layer to absorb either loss. A small group of two to ten providers adds per-provider pricing and billing consistency across associates to that same equation.
An eleven-to-fifty provider group answers to a practice administrator who needs a retention argument as much as a productivity one. A single departure at that scale can unsettle scheduling for an entire quarter.
The size of the practice changes, which line in the ROI formula carries the most weight. It does not change whether the formula applies.
The Bottom Line
Documentation burnout drains revenue long before it drains morale, through pajama time, undercoded charts, and physician turnover that costs six figures to replace. Notiro’s ICD-10 and CPT coding automation closes that revenue leak, while the same platform handles the note, covering the full clinical day rather than just half. Start a free trial at notiro, no IT setup required, and no enterprise contract to sign.
