Documentation burden fuels ambient AI
- Clinicians still spend far more time documenting in the EHR than providing direct patient care, increasing interest in ambient-AI tools. - Studies cite clinicians spending nearly two hours documenting for every hour of direct care, and behavioral-health notes can take 20–30 minutes. - Hospitals are adopting ambient AI specifically to reduce note burden, but questions remain about safety, editing time saved, and true workflow fit ( ).
Hospitals are buying ambient artificial intelligence scribes at a rapid clip as clinicians keep spending more time in the record than with patients. (ajmc.com) Ambient AI listens to a visit, turns the conversation into a draft note, and drops that draft into the electronic health record for a clinician to review. A January 27, 2026 study in *The American Journal of Managed Care* found 62.6% of U.S. hospitals using Epic had adopted one of these tools by June 2025. (ajmc.com) That hospital study covered 2,784 Epic hospitals out of 6,561 U.S. hospitals and found adoption was higher at metropolitan hospitals, nonprofits, larger systems, and hospitals with stronger operating margins. The three tools used by more than 80% of adopting hospitals were DAX Copilot, Abridge, and ThinkAndor, according to Emory University’s summary of the research. (emory.edu, ajmc.com) The sales pitch starts with a math problem clinicians already know. Time-motion studies cited in a 2026 narrative review say physicians spend about two hours on electronic health record and administrative work for every hour of direct patient care. (pmc.ncbi.nlm.nih.gov, pmc.ncbi.nlm.nih.gov) Behavioral health is one of the clearest examples because notes are long, narrative, and packed with risk, history, and billing details. A guest article published April 22, 2026 by *Healthcare IT Today* said behavioral-health notes often take 20 to 30 minutes each, which helps explain why vendors are targeting therapists, psychiatrists, and counselors. (healthcareittoday.com) Early results show time savings, but not a clean escape from documentation work. A multisite *JAMA* study published April 1, 2026 found AI scribe adoption at five academic medical centers was associated with 13.4 fewer minutes of total electronic health record time, 16.0 fewer minutes of documentation time, and 0.49 more visits per week. (jamanetwork.com) A separate Sutter Health study in *JAMA Network Open* found ambient AI was associated with less time in notes per appointment and lower mental, physical, and temporal demand after implementation. That study was a quality-improvement project at one organization, not a randomized trial across many health systems. (jamanetwork.com) The unresolved question is how much editing shifts from typing to checking. A 2026 narrative review found AI scribes often cut documentation burden but also reported frequent omissions and occasional clinically significant hallucinations, especially in specialties where precise details can change decisions. (pmc.ncbi.nlm.nih.gov) The adoption pattern also points to a second divide: the hospitals most able to pay for new tools are getting them first. The AJMC study found lower uptake in nonmetropolitan hospitals and for-profit hospitals, and the authors said uneven diffusion could widen gaps if the tools eventually improve burnout, quality, or costs. (ajmc.com, emory.edu) So the ambient-AI boom is less about replacing clinicians than about reclaiming minutes from the note. Whether those minutes hold up under real-world editing, safety checks, and specialty-specific workflows is what hospitals are testing now. (pmc.ncbi.nlm.nih.gov, jamanetwork.com)
