Ambient AI Moves Beyond the Exam Room as Hospitals Scale Documentation Tools

For several years ambient AI in healthcare meant one familiar scene: a doctor and patient talking in a quiet exam room while a phone captured the conversation and produced a draft note. That picture is outdated.

Ambient tools are now moving into nursing workflows, inpatient units, and emergency departments. In May 2026, Abridge made its nursing product available to nurses across more than 250 health-system clients. Co-developed with nurses, the tool integrates directly into EHR flowsheets. Systems already using it include Johns Hopkins Medicine, Emory Healthcare, Corewell Health, and Mayo Clinic. Source

Nursing charting is structurally different from physician notes. A single nursing shift can require 600 to 800 discrete data points rather than one continuous narrative. Tools therefore have to turn spoken assessments into structured entries. That forces nurses to narrate findings aloud again, a practice many left behind years ago.

Healthcare Expansion of Ambient Clinical Documentation Tools

The expansion is not limited to one vendor. After an early pilot at Cleveland Clinic, Ambience Healthcare released an inpatient nursing suite with conversational charting and ambient flowsheet support designed to cover an entire shift. Epic’s Chart with Art tool reached bedside nurses at Mount Sinai Medical Center in Miami Beach by late July. Source

Penn Medicine is further along. Roughly 1,300 ambulatory providers and more than 600 nurses, including about 550 on inpatient units, already use ambient tools there. Full expansion across ambulatory, inpatient, emergency, and home care is planned for this fall, according to Anna Schoenbaum, vice president and chief digital applications officer. Source

One of the largest published evaluations to date comes from a Spanish hospital network. Researchers tracked 12 months of ambient scribe use across 48 hospitals and five emergency specialties, covering roughly 2.27 million visits. The tool was used in more than one million consultations. Average consultation time fell by 21.8 percent, transcription accuracy held near 94 percent, and audits rated the generated reports higher on quality. Source

Privacy and Operational Risks of Ambient Clinical AI

For crowded emergency departments, a 21.8 percent reduction in consultation time could change how clinicians divide attention between charting and patients.

The rapid rollout also creates data-quality and privacy concerns. Clinical records feed risk models, quality measures, and decision-support tools. If a medication detail or assessment is missed or summarized differently by the model, the gap can affect later care decisions. Health systems are still deciding how long to keep raw audio, transcripts, and draft outputs. Whether these tools reduce charting burden without stripping out the structured clinical detail that the rest of the care team depends on remains the real test for ambient AI in nursing.