Nearly two-thirds of U.S. hospitals using Epic’s EHR had adopted at least one ambient AI tool as of mid-2025, according to research published in The American Journal of Managed Care. The technology’s effect on clinician burnout has been well documented; its financial return has been harder to pin down.
When Albert Villarin, MD, vice president and chief medical informatics officer at New Hyde Park, N.Y.-based Northwell Health, set out to scale ambient documentation across the system’s northern region, he built the measurement into the deployment before it began.
The groundwork started in 2022. As the health system emerged from the COVID-19 pandemic, Northwell surveyed its physicians about where the administrative burden was heaviest. The answer came back clearly: the EHR. Documentation workflows were inefficient and the EHR itself was a primary friction point. Leadership launched a roadmap to address it, measuring clinician feedback every six months as improvements rolled out. Ambient AI documentation was the next step in a process already years in the making.
“20% of our goals this year were based on our ability to bring this to fruition,” Dr. Villarin said. “So we set the bar very high. We want to reduce one to three minutes off the chart.”
The target was woven directly into annual KPI measurement. The team tracked charting time across every clinician on the platform. Last year, the data showed an average reduction of nearly three minutes per clinician — meeting the top of the range the team had set. The framing around that result was intentional. Dr. Villarin was deliberate about what the recovered time was meant to accomplish.
“Not to see more patients, but to be more efficient in evaluating the patients that they have,” he said.
That efficiency restored what the transition from paper records to electronic health records had steadily eroded: the uninterrupted clinical encounter. With the ambient tool handling documentation in the background, physicians could attend to the patient rather than the screen beside them.
“That respect, that integration, that interaction that we remember what drives us from the emotional aspect and why we want to deliver care for our patients,” Dr. Villarin said.
A second finding emerged as the team analyzed adoption data across sites. Patient satisfaction scores were moving in step with ambient AI usage rates. When the team ran a correlation between adoption percentiles and scores, the relationship held.
“Clinicians that were in the 75 to 85 percentile use of their charting had patient scores that were higher by a large factor,” he said. “So we did an R-squared analysis, and it showed across the board — the higher you use it, the more positive the responses were from the patients.”
The finding pointed to something the field has long understood anecdotally but struggled to quantify: documentation burden changes the nature of the clinical encounter itself. Dr. Villarin attributed the improvement to a shift patients experienced directly.
“Giving the patients the respect they deserved and having a conversation with them rather [than] having a conversation with your keyboard,” he said. “That’s what they have missed for years since we went through the process of going from paper to an EMR.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
