Before 2025, no randomized trial had tested whether ambient AI scribes save clinicians time. In late 2025, NEJM AI published 2 such trials in the same issue, and the results were more modest, and more useful, than most product pages suggest.
What an ambient scribe does
An ambient scribe listens to the visit conversation and drafts a note from it. The clinician then reviews, edits and signs that draft.
The pitch is simple: less typing during the visit and fewer notes left for the evening. The trials below tested how much of that holds up.
Keep one limit in mind as you read them. All 3 studies took place in health systems rather than small independent practices, and each followed clinicians for weeks or months, not years. Results in a 2-provider clinic could look different.
The UCLA randomized trial
In a trial published in NEJM AI, 238 outpatient physicians in 14 specialties were randomly assigned to 1 of 2 scribe products or to usual care. The trial ran from November 2024 to January 2025.
One product cut time spent writing notes by 9.5% compared with usual care. The other showed no significant change (−1.7%).
Both scribe groups showed improvements on burnout, task load and work exhaustion scores. The authors said those secondary findings need confirming in larger, multicenter trials.
Uptake is worth noting. Physicians in the scribe groups used them in about a third of their visits (33.5% and 29.5%), not all of them.
The Wisconsin randomized trial
A second NEJM AI trial followed 66 practitioners in ambulatory clinics across 2 states for 24 weeks. Of 71,487 notes written, 38% were drafted with ambient AI.
Work exhaustion and interpersonal disengagement fell by 0.44 points on a 5-point scale. Professional fulfillment rose slightly, but the change was not statistically significant.
Time spent on notes dropped by 0.36 hours a day, a little over 20 minutes. A drop in after-hours work did not hold up once the most extreme 3% of days were excluded.
Note quality scores ranged from 3.97 to 4.99 out of 5 across domains, and diagnostic billing codes improved. The team detected no drift in the software's performance over the 24 weeks.
A larger look at burnout
A 2025 study in JAMA Network Open surveyed 263 clinicians at 6 health systems before and after 30 days with a scribe. The share reporting burnout fell from 51.9% to 38.8%.
The biggest self-reported gains were in note-related mental workload (2.64 points on a 10-point scale) and in the ability to give patients undivided attention (2.05 points).
Read that result with care. It was a quality improvement study with no control group, participation was voluntary, and several authors reported ties to the scribe vendor.
Accuracy still needs a clinician
In the UCLA trial, physicians said clinically significant inaccuracies happened "occasionally" with both products. One mild adverse event was reported.
The authors concluded that these occasional errors require ongoing vigilance. In practice, that means every draft stays a draft until the clinician has read it, fixed it and signed it.
What this means for you
Taken together, the trials suggest a scribe can trim note time and ease some of the strain of documentation. They do not show hours handed back each day, and the gains varied by product.
Physicians in the UCLA trial rated the 2 products similarly, even though only 1 produced a significant time saving. A short pilot with a few clinicians, measured against their own baseline, will tell you more than any vendor comparison chart.
Questions before you start
- How is patient consent requested and recorded, and can a patient decline?
- Is the audio stored, and if so, where and for how long?
- Will the vendor sign a business associate agreement?
- Are unsigned drafts clearly marked in the chart?
- Can you measure your own note time before and after you start?
Ask your compliance advisor which recording and consent rules apply in your state before your first scribed visit.
EMR MD's AI scribe add-on follows this pattern: it asks for consent at the start of each scribed visit and keeps the note as a draft until a clinician reviews and signs it.
Sources
- Ambient AI Scribes in Clinical Practice: A Randomized Trial (Lukac et al.) NEJM AI, 2025
- A Pragmatic Randomized Controlled Trial of Ambient Artificial Intelligence to Improve Health Practitioner Well-Being (Afshar et al.) NEJM AI, 2025
- Use of Ambient AI Scribes to Reduce Administrative Burden and Professional Burnout (Olson et al.) JAMA Network Open, 2025
This article is general information, not legal or medical advice. Rules change; check the source and your compliance advisor before acting.