AI scribes cut documentation time by about 16 minutes per 8 hours of scheduled patient care, a 10.0% relative drop, according to a 2026 JAMA study of 8,581 ambulatory clinicians. Total EHR time fell by 13 minutes, or 3.0%, so the saving is real but small for most users.
Practice leaders weighing a scribe contract need the measured number, not the vendor's. This post sets out what two peer-reviewed studies found and what they mean for a practice budget.
What the JAMA multisite study measured
The paper, titled "Changes in Clinician Time Expenditure and Visit Quantity With Adoption of Artificial Intelligence-Powered Scribes: A Multisite Study," appeared in JAMA in April 2026. Mass General Brigham and the University of California, San Francisco co-led it, with data from five U.S. academic health systems. Medical Economics reported that it compared 1,809 clinicians who adopted an AI scribe with 6,772 clinicians at the same institutions who did not.
Mass General Brigham summarized the results for adopters against that comparison group:
- Documentation time fell by 16 minutes per 8 hours of scheduled patient care, a relative decrease of 10.0%.
- Total EHR time fell by 13 minutes, a relative decrease of 3.0%.
- Productivity rose by 0.5 patient visits per week.
Mass General Brigham also reported that the added visits brought a statistically significant but small revenue gain of about $167 per month for each clinician who adopted a scribe. Over 12 months that is roughly $2,004 per clinician, before the cost of the tool.
Why most users saw less than the headline
Use of the tool varied widely. Mass General Brigham found that only 32.0% of adopters used the scribe in more than half of their visits. That group saw twice the reduction in total EHR time and three times the reduction in documentation time compared with the average adopter.
The study followed ambient documentation use for more than two years. Even so, time spent in the EHR outside work hours did not differ significantly between adopters and the comparison group. Documentation time moved down during the workday, but evening and weekend charting did not shrink in a way the study could measure.
That matters for staffing math. A 16-minute saving per 8-hour block is about 3.3% of the block (16 divided by 480). A practice that expects a scribe to add a full patient slot to every clinic day will not get it from these numbers. The 0.5 extra visits per week is the figure the study actually observed.
What a small-office simulation found
A second study, in JAMIA Open in July 2026, tested a scribe in a controlled setting. Medical Economics reported that a University of Toronto team recruited 9 primary care physicians licensed in Ontario. Each saw 4 standardized patients (2 simple and 2 complex) in a simulated office at Women's College Hospital. Each physician charted two visits by typing and two with an AI scribe, then reviewed and edited the scribe's note.
The simulation measured a much larger effect than the JAMA study, because it timed only the charting steps inside a visit.
| Measure | Without scribe | With scribe | Source |
|---|---|---|---|
| Share of encounter spent documenting | 36.3% | 11.2% | JAMIA Open 2026, via Medical Economics |
| Average documentation time per visit | About 5.5 minutes | Under 2 minutes | JAMIA Open 2026, via Medical Economics |
| Post-visit charting share of encounter | 16.2% | 23.2% | JAMIA Open 2026, via Medical Economics |
| Scrolling share of visit time | 1.2% | 4.6% | JAMIA Open 2026, via Medical Economics |
Overall, documentation time fell by 69.1% in the simulation. Visits with a scribe ran about 1 minute shorter, which was not statistically significant. The post-visit difference narrowly missed statistical significance, and the authors did not assess note quality in this paper.
What this means for a practice budget
The two studies measure different things. The JAMA study counted real clinicians over months, and the average effect was 10.0% of documentation time. The JAMIA Open study counted 9 physicians in 36 scripted visits, and the effect was 69.1%. The first number is the better guide to what a whole practice will see.
Clinicians also report mixed results. Medical Economics described a flash poll of 103 physicians at an April webinar. In that poll, 42.0% said a scribe saved zero minutes on a typical clinic day, and 20.0% said it saved more than an hour. The poll was informal and covered a small group, so it shows spread more than an average.
Accuracy needs a check as well. Medical Economics cited a January 2025 JMIR study that found errors in 70.0% of notes from commercial scribes in simulated visits, with omissions the most common type. The same article cited earlier Mass General Brigham research in which clinician burnout among scribe users fell from 52.6% to 30.7% after 84 days.
A practice can use these figures to set a test. Track documentation minutes per 8 hours of scheduled care for 30 days before and after a pilot. Compare against the 16-minute and 13-minute marks from JAMA. Log weekly visit counts against the 0.5-visit mark, and review a sample of notes for omissions.
Track evening charting as its own line. Mass General Brigham found no significant change in after-hours EHR time among adopters over more than two years of use, so a scribe pilot that shows shorter in-clinic notes can still leave the evening workload where it was.
Sources
- Mass General Brigham, AI scribes linked to modest reductions in EHR documentation time, 2026
- Medical Economics, AI scribes cut documentation time by 69% in simulated primary care visits, study finds, September 16, 2026
Talk to us
Fifteen minutes about your practice and the systems you run.