NASHVILLE, Tennessee—After a six-month pilot period that pitted five major ambient scribes against each other, the Cleveland Clinic selected a scribing company to partner with: Ambience.
The Cleveland Clinic unveiled the exclusive, five-year partnership Wednesday, following a presentation at digital health conference ViVE 2025 that explained the nonprofit hospital system’s testing methodology for evaluating scribe technology. Ambience’s platform uses AI for documentation, clinical documentation integrity and point-of-care coding.
Executives said it's the first head-to-head test of the major AI scribe products for healthcare, and the extensive piloting revealed differences in the quality of the clinical documentation generated by the vendors.
“We realized that the underlying technology of some of them was significantly better than the others, and that was a surprise,” Beth Meese, executive director of digital health at the Cleveland Clinic, said in an interview.
The technology will be rolled out to ambulatory Cleveland Clinic providers in the U.S. in phases over the next several months. The AI documentation software that records patient appointments and generates comprehensive medical notes. The technology also generates customized after-visit summaries for patients.
The Cleveland Clinic has been ranked the No. 2 hospital in the world by Newsweek for six straight years. It’s a key healthcare research center that spearheaded major medical breakthroughs in the last century. It has a presence in London, Abu Dhabi, Florida, Toronto, Las Vegas and Ohio.
The clinic is also a founding member of the international AI Alliance—launched by IBM and Meta—which convenes researchers, developers and organizational leaders to improve AI safety and security.
Many health systems are testing out multiple medical AI scribe technologies, because choosing a long-term partner is a major investment.
The Cleveland Clinic tested the five AI scribe solutions for documentation quality, product features, provider satisfaction, ease of implementation and return on investment. It analyzed data from Epic, provider survey results, patient feedback and technical evaluations.
Meese emphasized that the clinic’s experience with all five vendors was positive.
The Cleveland Clinic involved over 300 clinicians in the pilot over a six-month period. Each AI scribe solution was tested for three months. The clinic rotated physicians who expressed varying comfort levels with technology and practiced across 20 specialty areas.
“We wanted some [providers] that were perhaps a little less interested in adopting new technology, so that as we reached the satisfaction scores, and as we looked at the time in the chart, we had a variety of users across different specialties, across different types of users,” Meese said.
The clinic started the scribe evaluations on the basis of partnership and physician satisfaction. It used the NASA Task Load Index and the Mini Z test to measure provider burnout.
The organization heavily prioritized the ongoing collaboration with the engineering teams. A collaborative relationship with the vendor, robust engineering support and long-term growth of the company were top priorities for Cleveland Clinic when evaluating an ambient scribe partner.
Ambience was particularly responsive to the clinic’s feedback during the pilot period, Meese said, and it had local team members that assisted clinic providers to implement and understand the tool.
“I really thought that these products were going to be commodity [sic], and that as long as we proved that one of them worked, that we could pick the one in the market that we thought was best,” Meese said. “So I think one of the biggest things that I learned is that they are not necessarily just a commodity product. Each of them had something different, but really the output itself was surprisingly different across some of the different vendors.”
Meese said that she has shared the clinic’s findings with other interested health systems that reached out about the results.
The Cleveland Clinic only publicly released the results of the Ambience pilot and did not release data it gathered from the other four vendors.
The health system's providers racked up 25,000 encounters across 20 specialties using Ambience. In 80% of visits, providers opted to use the clinical documentation generated by Ambience during the encounter, and 67% of clinicians reported reduced cognitive burden. The clinic found a 7% increase in same-day chart closure rates, a 32% increase in face time with patients, 49.6% decrease in pajama time and 25% decrease in note creation time.
According to data from the pilot, the highest adoption was by nephrology providers (99%) and the lowest adoption was by urologists (55%).
The slide deck Meese presented at ViVE noted the 80% adoption figure was two to three times higher than other ambient scribes it tested, based on anecdotal evidence of the comparative pilot results.
When adoption of the solution by the system's cardiologists was low, Ambience worked with the clinic to upgrade product features to better align with the needs of cardiologists, and utilization jumped from 50% to 71% among those specialists.
Meese urged other health systems to also undertake extensive pilots.
“It seems more daunting than it actually was to move forward with five different vendors … there's some opportunity to create some efficiency, even in the piloting space itself, by doing multiple one after another,” Meese said.
Cleveland Clinic intends to conduct post mortem reviews with the other vendors, she noted.
“Every one of the pilots went very well, so there was not a bad experience,” Meese explained. “Some of them were smaller or less integrated. Some of them, the output wasn't as great. Some of them, the technology was just harder in the beginning to get going. So our tech team was less satisfied with working with them.”