While paper documentation still lingers in certain areas of health system operations, today’s physicians are largely spared from the manual record retrieval and clunky metal filing cabinets that were commonplace before the rise of the EHR. However, physicians still face significant cognitive burden in the form of clicks and time spent with digital interfaces.

This burden is perhaps most pronounced in primary care, where the average patient visit runs a half hour, but the total time a physician spends in the EHR per visit — both during the clinical interaction and documentation afterward — is 36 minutes, according to recent research published in JAMA Network Open.

While ambient AI solutions have yielded considerable benefits to primary care physicians in the form of reduced documentation, this alone doesn’t mitigate the burden a clinician may experience during a patient visit when they need to collect patient information from disparate sources to make informed and effective clinical decisions. However, many have predicted that AI will one day alleviate this burden as well. That day is getting closer.

Enter Chart Chat

On Aug. 18, Ambience — a leader in the ambient AI space — unveiled an AI copilot designed to integrate directly into Epic. The solution — dubbed Chart Chat — merges patient EHR data with medical content from BMJ Best Practice and OpenAI’s reinforcement fine tuning (RFT) technology. The solution serves as a companion to Ambience’s Patient Recap — which automatically creates previsit chart summaries — and allows clinicians to ask medical questions relevant to particular patients.

“This technology can fundamentally change how physicians interact with the medical record,” said Trevor Satterfield, MD, a family medicine physician at Boise, Idaho-based St. Luke’s Health System. “It lets me ask specific questions about the patient’s history or treatment, questions that could take minutes of searching to answer, and get clear, accurate responses in seconds.”

While health systems have previously worked to create their own in-house AI assistants, Chart Chat — according to Ambience — is the first “production-quality solution” of its kind built into Epic.

The initial rollout of the solution will involve nearly 1,000 clinicians across a dozen health systems through the end of 2025. This phased deployment allows the team to gather feedback, iterate quickly and offer white-glove service to partners.

From documentation to high-stakes decision support

Early wins with AI have largely focused on documentation and back-office tasks. While important, these areas of operations don’t carry the same stakes of direct patient care.

In an interview with Becker’s, Ambience’s Cofounder and Chief Scientist Nikhil Buduma said the rollout will begin in primary care and cardiology before moving on to other specialties. This plan is indicative of the stakes involved in adopting AI tools that can have a direct influence on clinical outcomes.

“The ways physicians across disciplines will use this technology are incredibly different,” Mr. Buduma said. “We’re going to release specialty by specialty. This gives us the opportunity to build the tool to a level where it can reason through the kinds of complex questions that certain specialists will ask with high fidelity.”

Beyond a capacity to navigate clinical complexity, for Chart Chat to be truly effective, it must also be able to assess conflicting information that can often appear in medical documentation. For example, one piece of documentation may include a medication for chronic kidney disease while another states that the patient doesn’t have a chronic illness. Ambience has trained its copilot to reconcile such issues.

“There is an incredible amount of messiness these models have to reason over accurately in order to come to good conclusions,” Mr. Buduma said. “To build trust in these models … we’re rolling [Chart Chat] out to a small number of users, collecting feedback and monitoring it very closely as we roll this out to more and more users.”

Unlocking AI ‘magic’

The announcement of Chart Chat comes on the heels of Ambience landing $243 million in Series C funding, bringing its valuation to $1.25 billion. According to Ambience’s cofounders, the funding will help the company bring “radical change” to an industry with a reputation for lagging on innovation.

What other role could Chart Chat play in healthcare’s AI-driven transformation? According to Mr. Buduma, the EHR-integrated solution will move beyond reactive feedback into anticipatory assistance. For example, if a physician starts discussing GLP-1 medications with a patient, the tool could automatically surface relevant eligibility data, which would reduce clicks and increase face time.

Mr. Buduma described this scenario as essentially having another clinician in the room during patient visits, anticipating where the conversation might go and what resources the physician might need. In this world, when a physician does look back to a digital screen, instead of moving through a series of clicks, they simply find the resources and information they need readily on display.

“I think we’re going to elevate the clinician workflow into a truly magical user experience,” Mr. Buduma said.

The post With Chart Chat, Ambience aims to unlock AI ‘magic’ appeared first on Becker’s Hospital Review | Healthcare News & Analysis.