Blog

Trends

HCP AI Pulse Survey | Q4, 2025

Written by:

Michael Hess

September 25, 2025

8 minute read

Artificial intelligence is quickly becoming part of the clinical toolkit, but the conversation often centers on hype rather than how healthcare professionals themselves are actually using it. To ground the discussion, we fielded an AI in Healthcare Pulse Survey—asking front-line clinicians and leaders how familiar they are with AI, which tools they’re already using, where they see the most value, and what would build their trust going forward.

We heard from 109 healthcare professionals across 35 U.S. states, the vast majority (95%) working in direct patient care. Titles ranged from physician assistants and attending physicians to highly specialized experts such as pediatric surgeons, vascular surgeons, orthopaedic surgeons, fellowship-trained Mohs surgeons, and hematology/oncology physicians. This mix gives us a clear view into how AI is landing in real clinical environments today.

AI in Healthcare Pulse

How familiar are you with AI in healthcare?

Slightly familiar
53%
n=59
Very familiar
34%
n=38
Somewhat familiar
12%
n=13
Not familiar at all
1%
n=1
AI in Healthcare Pulse

Which AI-powered tools (if any) are you aware of or using?

Documentation assistants / automated note‑taking
73%
n=83
Clinical decision support (diagnostics; treatment recommendations)
52%
n=59
Predictive analytics (e.g.; risk stratification; readmission)
20%
n=23
Imaging interpretation (radiology; pathology)
20%
n=23
Patient chatbots / triage tools
19%
n=22
Other
5%
n=6
AI in Healthcare Pulse

How valuable do you think AI is today in improving patient care?

Somewhat valuable
55%
n=61
Very valuable
35%
n=39
Neutral / unsure
7%
n=8
Not valuable yet
3%
n=3
AI in Healthcare Pulse

Which area of healthcare do you believe AI will have the biggest positive impact in the next 3–5 years?

Reducing administrative burden
50.5%
n=56
Improving diagnostic accuracy
21.6%
n=24
Enhancing patient communication and engagement
16.2%
n=18
Predicting health risks and outcomes
10.8%
n=12
Other: don't know
0.9%
n=1
AI in Healthcare Pulse

If you could "hire" an AI assistant for one task tomorrow, which would you choose?

Drafting and organizing patient notes
48.6%
n=54
Managing patient communications and scheduling
18.0%
n=20
Assisting in diagnostic decisions
12.6%
n=14
Identifying high-risk patients early
9.0%
n=10
Summarizing the latest medical research
6.3%
n=7
Other
5.4%
n=6
AI in Healthcare Pulse

What would make you most likely to trust AI in clinical practice?

Proven accuracy in studies
49.5%
n=55
Integration into existing workflows
21.6%
n=24
Transparency in how AI makes decisions
13.5%
n=15
Regulatory approval / FDA clearance
8.1%
n=9
Peer adoption and recommendations
7.2%
n=8
AI in Healthcare Pulse

Overall, do you believe AI will make patient care better or worse in the next decade?

Somewhat better
62.2%
n=69
Much better
31.5%
n=35
No real change
5.4%
n=6
Somewhat worse
0.9%
n=1
AI in Healthcare Pulse

Overall, what excites you most (or worries you most) about AI in healthcare?

What excites clinicians

“Reduction of busywork and improvement in diagnostic accuracy in radiology and other medical technology.”

Attending Physician • Illinois

“I am excited about potentially less provider documentation required to do my job.”

Physician Assistant • Washington

“Creating a workflow that allows providers to spend more time with patients while also creating improved accuracy.”

Oncology Physician Assistant • Pennsylvania

“The ability to improve patient outcomes and reduce administrative burden.”

Resident Physician • New York

What worries clinicians

“Concern that it will miss something critical or life‑threatening.”

Physician Assistant • Florida

“Errors worry me—or 'missing' something because I relied on AI.”

Physician Assistant • New Jersey

“Wrong diagnosis; duplication of record; errors; miscommunication.”

Attending Physician • New York

Never miss an opportunity.

Resources for Researchers

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.