Abridge CEO Shiv Rao, MD, took stage in Midtown Manhattan June 11 to describe a series of moves that firmly take the company from a clinician documentation tool to a healthcare intelligence platform.
The event, streamed virtually and with a live audience of about 150 partners and investors, signaled how Abridge is moving closer to new corners of healthcare:
- It partnered with Nvidia to build a first-of-its-kind foundation model for clinical conversations;
- Cigna and Aetna executives signaled readiness to get in the game for both clinical documentation and movement toward real-time claims adjudication;
- and Eli Lilly is making a “strategic investment” in Abridge to support evidence-based care and research access.
Abridge unveiled the collaborations in addition to advancements in its core technology: ambient AI for clinical documentation. New capabilities include a voice agent, expanded clinical decision support, pre-charting and pre-round notes for inpatient and outpatient settings, real-time patient discussion guidance and clinical trial eligibility determinations at the point of care.
“We have an opportunity right now not to sprinkle AI over all of these workflows and hope things are just going to get better naturally,” Dr. Rao said. “In fact, it can get a lot worse — AI versus AI, agents versus agents, would probably race to some dystopic future nobody wants to live in. Our opportunity right now is to use AI to actually refine, redesign the system.”
Abridge, founded in 2018, has grown to more than 300 health system partners, including the newest — enterprisewide Northwestern Medicine. The Chicago-based system is also among the first to expand into Abridge’s new smart-room capability, which the company is powering through integrations with Artisight and hellocare.ai, among others.
“When we think about the potential of Abridge and their innovation roadmap, we know they’re moving beyond all of these different point solutions for clinical documentation, patient context and workflows across the entire care journey,” said Hannah Koczka, vice president of NM Ventures and Innovation at Northwestern Medicine. “It’s about trying to bring it all together into a single unified intelligence layer, and providing better insights before, during and after the encounter.”
More than minutes saved
For years, ambient AI’s story in healthcare has largely been told in minutes saved and eye contact restored. June 11 gave it an even bigger one.
The conversation is broadening to how those saved minutes can compound into something far larger: curbing the administrative spending that drags on American healthcare. And more ways “Abridge” is looking to make good on its name, with dual focus on compressing burdens and connecting healthcare players and moments.
Consider the moments, first: before, during and after patient visits. The “during” part of the visit, where the technology has lived for the last several years, is no easy feat on its own. Eric Lee, MD, medical director of clinical informatics at Commerce, Calif.-based AltaMed Health Services, can confirm from his vantage point at one of the nation’s largest federally qualified health centers. “It’s really a meeting without an agenda going in — and you’re having 20-minute meetings like this all day,” Dr. Lee said of the patient visit.
Abridge’s clinical decision support now suggests discussion topics mid-visit based on the patient’s record and the live conversation, and lets clinicians ask evidence-based questions without leaving the workflow — drawing on a content library anchored by Wolters Kluwer’s UpToDate, The New England Journal of Medicine and JAMA, and now joined by the American Diabetes Association, the American Academy of Family Physicians, Neurology and the Journal of Clinical Oncology. Clinicians will also now be able to claim CME credit for that point-of-care learning.
Before the visit, Abridge generates pre-round notes for inpatient care drawing on emergency department documentation, nursing assessments, labs and imaging. It also surfaces the care gaps and chronic conditions most relevant to each patient, so clinicians address them in the room rather than discovering them in retrospective chart review. Post-visit, the platform drafts flowsheets, billing codes and orders, with an AI agent that lets clinicians tailor outputs in natural language before they flow to the EHR.
The new voice interface, activated by the phrase “Hey, Abridge,” extends that agent further. In the keynote demo, a clinician used it mid-visit to order a confirmatory echocardiogram without breaking the conversation and within one interface.
“The Abridge AI agent can help a cardiologist like me create an op note, get a patient one step closer to a lifesaving clinical trial or prep my charts for next Monday’s clinic — all in under three minutes,” Dr. Rao said. “Without it, this would be hours of paperwork. Instead, it’s a conversation.”
The moments aren’t just widening — the care team is, too. Abridge’s nursing capability turns bedside conversations into structured draft documentation — and, more consequentially, feeds what nurses capture into the pre-charted notes preparing the next clinician who sees that patient.
“I see the gap between the nursing ambient piece and the physician ambient piece growing closer and closer together,” said Misti Foust-Cofield, RN, vice president and chief nursing officer of Richmond, Ind.-based Reid Health, which will have every nurse in its system live on the platform by the end of the third quarter. At Reid, the workforce effects have been measurable: a nursing vacancy rate down from 18% to 8.6% with zero contract staff, and incidental overtime cut 70% on units where the technology has rolled out.
“I think we’ll be able to alleviate many of the things that oftentimes feel redundant and duplicative to our patients and come to better outcomes quicker, because that information is coming closer together,” Ms. Foust-Cofield said.
Calling off the bots
Abridge made headlines last summer when it compressed the prior authorization process to roughly the length of a patient visit, in partnership with Pittsburgh-based Highmark Health. That turned out to be one piece of a larger ambition the company laid out June 11: aligning real-time documentation with real-time claims. No formal agreements were announced, but Cigna and Aetna executives, appearing on stage, indicated interest in building toward it.
When a physician sees a patient today, the information from that visit is processed, reviewed and adjudicated over weeks, critical details get lost, and billions of dollars of care goes unrecognized every year. Or, as Joon Lee, MD, CEO of Atlanta-based Emory Healthcare, put it on stage: “As a provider, you’re paid not by what you do or what you achieve, but by what you document.”
By grounding documentation, billing codes and claims in the full clinical picture at the moment care is delivered, Abridge argues, health systems and payers can compress that 12-step cycle into two or three — and dismantle a shadow industry of denial and appeal that both sides now staff with armies of people and, increasingly, AI.
“My nightmare is that we have bots on the Emory Healthcare side and bots on the Cigna side that keep talking to each other, and a whole data center could be occupied just doing that,” Dr. Lee said. “All of that is a drag on the system — it’s not helping us deliver care. It’s taking resources away from the delivery of healthcare.”
Madhu Nutakki, CIO of U.S. employer technology at Cigna Healthcare, brought the argument back to the person in the exam room. “She’s sitting in that chair, but she isn’t just thinking about the clinical aspects of her symptoms,” he said, referring to the patient in Abridge’s keynote demo. “She’s also thinking about: What is this going to cost? How do I pay for it? Do I actually know how much it costs?” He later added that the bill arrives “30, 60 days after — she doesn’t know how much it’s going to be, and the first time she opens it, in most cases, it’s a surprise. You can eliminate all of that by making it predictable.”
The same logic extends to value-based care, the subject of a panel pairing Johns Hopkins Health System CIO Deanna Hanisch with Ben Kornitzer, MD, senior vice president and chief medical officer of Aetna. Payers hold claims, pharmacy and cross-system data that providers can’t see; providers hold the clinical moment payers can’t access in real time. Dr. Kornitzer noted that about 25% of patients with chronic kidney disease don’t know they have the diagnosis, which is the kind of gap that only closes when both data sets meet.
“Rather than create an arms race and industrialize the parts of healthcare that frustrate all of us, [we can] actually stack hands and say: What can we do that’s better for the patients?” Dr. Kornitzer said.
On the revenue-integrity side, Abridge is also partnering with AHIMA to put its coding and clinical-documentation-improvement outputs under the kind of scrutiny the coding profession expects across both fee-for-service and value-based models.
The trial finds the patient
The platform’s other new frontier reaches toward research, and it is where Eli Lilly enters the room. Abridge said life sciences collaborations will be designed to help clinicians and health systems identify potential clinical trial candidates and initiate screening pathways at the point of care — with appropriate governance, consent, security and institutional control.
The keynote demonstrated the capability live: a heart failure patient flagged as a potential study candidate mid-visit, with the physician ordering a confirmatory echocardiogram and looping in the research coordinator by voice.
That 30-second sequence stood in for what is today a heavily manual, weeks-long process, with research teams chasing eligible patients chart by chart or patients left to hunt down trial options on their own after the visit.
600 milliseconds
Of June 11’s announcements, the Nvidia partnership goes deepest into Abridge’s core technology.
The two companies are building a foundation model made specifically for clinical conversations, constructed on Nvidia’s Nemotron open model family and trained on Blackwell infrastructure with de-identified data. The goal is a model that reasons clinically from its foundation rather than a general model retrofitted for healthcare.
“Imagine you try to grow a generic AI model into healthcare. It doesn’t understand the clinical language, it doesn’t have the clinical reasoning, and it surely doesn’t have the domain expertise of all the long-running tasks and interconnected work that has to happen for workflows to be completely transformed,” said Kimberly Powell, Nvidia’s vice president of healthcare. “Generic AI is just not going to work.”
The returns are already showing up in benchmarks Dr. Rao shared on stage. Abridge’s in-house model now moves from diagnoses to ICD-10 codes in less than 600 milliseconds — fast enough to live inside the clinical workflow rather than behind it. Seventy percent of Abridge notes already contain sections generated by the company’s in-house models.
Reading the room
Most partners in the audience on June 11 had little sense of what Abridge was about to unveil, and their reactions are a signal of what the company’s customer base values most: simplicity, deep understanding of clinical complexity, and technology that leads with the users rather than the tech.
One physician and innovation executive with an academic medical center said after Dr. Rao’s presentation that Abridge’s advancements further cement the EHR as a “data repository,” with Abridge becoming a clinical intelligence partner that allows him to “take care of my patients the way I want to” with medical knowledge, reasoning and context.
Or as Craig Norquist, MD, CMIO of Scottsdale, Ariz.-based HonorHealth — who noted his views are his own — put it: “The last 10 to 15 years of health tech has been about how do we make the EHR work because once it works, it’s going to be glorious. Really, it’s just a tool, like a scale. If you thought your scale was going to make you lose 30 pounds and give you a six-pack, your scale is not the right instrument. The EHR should never have been held to the expectation that it would do more than house information.”
What captured Dr. Norquist most was pre-charting: walking into a room already knowing the important points, so the note is complete the first time and the after-hours back-and-forth with documentation improvement, coding and billing teams never starts.
That practicality ran through the room. Providers at the event were the most pragmatic voices in it — appreciative of the Eli Lilly investment and the payers’ presence on stage, but with a mood best summarized as “more to be seen.” Cigna and Aetna executives said they share the same end goal as providers — health and good outcomes for patients — but many health system leaders said they want more candor.
‘The first step is admitting you have a problem,’ as the adage about recovery goes — and June 11 contained a bit of that, with academic medical center and payer executives joining Dr. Rao in denouncing the arms race of bots tangling over claims. But at least one health system leader craved more accountability in the conversation, especially from the payers.
“It’s kind of fresh for insurance companies to say, ‘We want to work on improving prior auth.’ Well — you created prior auth, and you now recognize what a nightmare it is,” Dr. Norquist said. “Part of me feels like we’re letting them off the hook by having a conversation about how we work together to fix the hurdles they put in place, when the conversation should be: We understand why this was in place, but can we acknowledge it doesn’t work — and how can we do something better?”
He gave the payers this much: “It’s great that they’re willing to sit on the same stage and say things we all like to hear. Hopefully they actually start to deliver on them.”
What Dr. Rao asked the room to hear was bigger than any one announcement. “Imagine every clinician rounding in the hospital right now feeling like there’s a team of residents, a team of assistants, a team of agents figuring out all the things they want to know — helping them ask the right question, leverage the right evidence, create the most clinically useful and compliant note,” he said. “And imagine it works at every single moment of care, across all the different care settings.”
Dr. Rao, who is still a practicing cardiologist, seems to understand the terms of that ask. Gesturing on stage at a map of Abridge’s 300-plus health system deployments, he offered his own measure of the company: “Those aren’t installations. Those aren’t integrations. That’s a map of trust — and in healthcare, the only currency that matters is trust.”
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.