In MGMA's August 4, 2026 Stat poll of 189 medical groups, 83.0% said AI plays some role in patient visits, up from 71.0% a year earlier. The other 17.0% report no AI use in visits.
Use has also deepened. Groups that use AI in more than a quarter of their visits rose from 24.0% in 2025 to 45.0% in 2026.
How many groups use AI in patient visits
MGMA asked the same question in two August polls. The 2025 poll, fielded August 5, 2025, drew 244 applicable responses. The 2026 poll drew 189. All shares below use all respondents as the base.
| Measure | August 2025 | August 2026 |
|---|---|---|
| Any AI use in patient visits | 71.0% | 83.0% |
| No AI use in visits | 29.0% | 17.0% |
| AI in 25% of visits or fewer | 47.0% | 38.0% |
| AI in more than 25% of visits | 24.0% | 45.0% |
| AI in 76% of visits or more | 7.0% | 17.0% |
Source: MGMA Stat, August 6, 2025 (n=244) and August 5, 2026 (n=189). The 2025 figure for more than 25% of visits comes from MGMA's 2026 article.
Both polls collect answers from practice leaders who choose to respond. The respondents differ between years, and the sample fell from 244 to 189, so neither poll is a census of practices.
Any use rose 12.0 points in one year. Heavy use grew faster: the 76% or more group went from 7.0% to 17.0%, about 2.4 times its 2025 size.
An earlier MGMA Stat poll points the same way. On September 30, 2025, 351 groups answered a question about the year behind them. Of those, 68.0% had added or expanded AI tools in 2025, 31.0% had not, and 2.0% were unsure. MGMA noted that 53.0% had not added or expanded AI the year before. In an April 2025 poll, 19.0% of groups used chatbots or virtual assistants.
How deep the use goes
The low end of the scale shrank by 9.0 points. In 2025, 47.0% of groups used AI in 25% of visits or fewer. In 2026 that figure was 38.0%.
The middle and top grew by 21.0 points together. In the 2025 poll, about 10.0% of groups used AI in 26% to 50% of visits, about 7.0% in 51% to 75%, and about 7.0% in 76% or more. MGMA's 2026 article reports the combined share above 25% as 45.0%, against 24.0% a year before.
Among groups that use AI at all, about one in five now use it in at least three quarters of visits. MGMA puts the 2025 figure at about one in ten.
A group with 12,000 visits a year and AI in 25% of them has AI in at most 3,000 visits. At the 76% line the same group has AI in at least 9,120 visits. That is arithmetic on a round number, not a survey result, but it shows how large the gap between the bands is.
Does AI cut staff workload
MGMA's 2025 poll asked groups that use AI in visits about workload. About 44.0% said AI had not reduced it, 39.0% said it had, and 17.0% were unsure. MGMA did not state the base size for that question.
The 2026 poll coded open-ended replies from leaders whose groups use AI. MGMA did not state the sample size for this question either. The results:
- 30.0% reported no real change in workload.
- 27.0% said work shifted to other tasks.
- 19.0% said workload went down.
- 14.0% said it was too early to judge.
- 10.0% described both a reduction and a shift.
Counting every mention, 37.0% described some shift in work and 29.0% described some reduction. The 2025 and 2026 questions differ in format, so the 39.0% and 19.0% figures are not a trend line.
What physicians report
The American Medical Association's 2026 Physician Survey on Augmented Intelligence, reported on March 12, 2026, reached nearly 1,700 physicians. Of those, 81.0% use AI professionally. The AMA measures use by physician, not by group, so the figure does not compare directly with MGMA's.
The AMA listed these uses among physicians:
- 39.0% summarize medical research and standards of care.
- 30.0% draft discharge instructions, care plans, or progress notes.
- 28.0% document billing codes, charts, or visit notes.
- 28.0% summarize charts.
- 19.0% draft replies to patient portal messages.
- 18.0% use translation services.
- 17.0% use AI for assistive diagnosis.
In the same AMA survey, 70.0% see AI as a way to automate tasks that add to burnout. A share of 85.0% want to be consulted or directly involved in AI adoption decisions.
athenahealth's fifth annual Physician Sentiment Survey, reported by Medical Economics on May 28, 2026, found 54.0% of physicians comfortable using AI. The same survey found 73.0% still struggle to exchange data across EHR platforms and 63.0% say patient chart data overwhelms them. The article does not state the sample size.
What to measure in your own practice
Three counts place a practice against the MGMA bands: total visits, visits where an AI tool produced a note or message, and the share of those two. A practice below 38.0% sits with the low-use group. A practice above 45.0% sits with the heavy-use group.
The workload question needs its own baseline. In MGMA's 2026 poll, 30.0% of AI-using leaders saw no change and 14.0% could not yet tell. Minutes of note time per visit, taken before and after a tool goes live, give a practice a number to hold against those replies.
MGMA's full 2026 analysis, including research on productivity and governance, is open only to MGMA members.
Sources
- MGMA Stat, As AI use expands, medical practice leaders shift attention to measurement, August 5, 2026
- MGMA Stat, Most practices use some form of AI, but is it actually reducing staff workloads?, August 6, 2025
- MGMA Stat, Document, schedule, communicate: Where ambulatory care has added or expanded AI in 2025, October 1, 2025
- American Medical Association, More than 80% of physicians use AI professionally: AMA survey, March 12, 2026
- Medical Economics, Why doctors are worried about the patient affordability crisis, May 28, 2026
Talk to us
Fifteen minutes about your practice and the systems you run.