Abridge clinches $150M to build out generative AI for medical documentation

Investors continue to pour money into generative AI startups, and Abridge is riding this wave as it rapidly scales its technology across U.S. health systems.

The Pittsburgh-based startup banked a hefty $150 million series C funding round to invest in more research and development to build bedrock foundation models to power both improvements to the company's existing technology and new products, CEO Shivdev Rao, M.D., told Fierce Healthcare.

The latest funding comes just four months after Abridge raised $30 million in a series B round.

Lightspeed Venture Partners and Redpoint Ventures co-led the series C investment, and new and existing investors joined the round including IVP, Spark Capital, Union Square Ventures, Bessemer Venture Partners, Wittington Ventures, Mass General Brigham Artificial Intelligence and Digital Innovation Fund (AIDIF), Kaiser Permanente Ventures and CVS Health Ventures.

The company, founded in 2018, has raised $212.5 million to date.

Abridge declined to disclose its valuation. 

The company, one of Fierce Healthcare's Fierce 15 of 2024 honorees, uses AI to increase the speed and accuracy of medical note-taking, leveraging a proprietary data set derived from more than 1.5 million medical encounters. The company's AI converts a patient-clinician conversation into a structured clinical note draft in real time and integrates it seamlessly into the EMR.

Around 5,000 doctors currently use the company's software.

Abridge was founded on the premise that clinician-patient conversations are at the core of all healthcare, Rao, a cardiologist, noted. His family's own journey through the healthcare system also made him aware of the need for better communication, and he saw an opportunity to use technology to improve the gap between healthcare conversations and what happens next. Motivated by personal and professional experiences, Rao worked together with Florian Metze, Ph.D., and Sandeep Konam to launch Abridge five years ago.

"We're going to invest heavily into R&D with this funding. R&D is a big part of the ethos of the company. With the series B funding, we were really focused on using that capital to successfully scale and demonstrate that we can scale across large health systems. We've absolutely accomplished that in spades," Rao said in an interview. "Now we have a flywheel of clinician and system word of mouth, and on the strength of that growth, this round is going to accelerate R&D to further build out the healthcare AI layer for medical conversations."

The market for generative AI tools to help manage medical documentation has been heating up with Microsoft's Nuance being a dominant player. 

Investors are eager to back generative AI startups in the healthcare space. Ambience Healthcare recently reeled in $70 million for generative AI tools for clinicians. Nabla, an ambient AI assistant for practitioners, closed a $24 million series B round back in January.

Back in late August, Epic and Microsoft laid out plans to integrate a generative-AI-powered tool, called DAX Copilot, into Epic's EHR software to automatically draft clinical notes during patient exams. Epic competitor Cerner, now owned by software giant Oracle, also has integrated generative AI services into its EHR as health IT companies race to harness this cutting-edge tech for providers. 

Google continues to advance its generative AI models designed specifically for healthcare use cases and in December unveiled MedLM, a family of foundation models designed for healthcare industry use cases and available through Google Cloud.

Abridge has worked to build what Rao calls a "world-class AI team" led by its chief technology and science officer Zach Lipton, Ph.D., an associate professor of machine learning at Carnegie Mellon University.

"We also have a 'for clinicians by clinicians' ethos, and when you put those two ingredients together, there's the ability to create that AI layer for healthcare wherein we can go after all the features and functionality that can amount to a magical user experience for the people who matter most, the patients and their care team," Rao noted.

In the last year alone, Abridge delivered an automatic speech recognition engine purpose-built for healthcare applications. The company's speech and language technologies have been rigorously evaluated in 14 languages and functionally support many more—performing well even amid rapid switches between languages and when interpreters are in the room, executives claim.

Abridge is ahead of the market in controlling its full technology pipeline, developing its own purpose-built large language models. 

In past interviews, Rao has pointed to key distinctions in Abridge's capabilities such as the company's "Linked Evidence" feature that maps any highlighted region within a summary to the substantiating evidence in the source transcript. The company is actively developing novel ways to personalize notes, integrate more deeply into the electronic health record and provide clinicians with after-visit insights, executives said.

Abridge also has seen rapid uptake of its AI-powered solution among health systems. The company did not disclose how many hospitals it works with, but Rao said Abridge went live with 10 health systems just in the past three weeks.

Atlanta-based Emory Healthcare, an academic health system in Georgia with over 3,450 clinicians, has signed on to use Abridge's technology, integrated with Epic, enterprisewide across its organization. 

The company also is working with The University of Kansas Health System and UPMC to deploy its technology with Epic workflow integrations. The rollout has the potential to serve more than 1,500 physicians across the system’s more than 140 locations. 

In a major win for the company, Epic integrated the company’s generative AI for clinical documentation into its EHR workflow as part of the tech giant's new third-party vendor program.

Along with the funding round, Abridge also announced an enterprise agreement with Connecticut-based Yale New Haven Health System to give thousands of clinicians access to the technology for clinical documentation. The initial deployment will be focused on reducing the cognitive burden of clinical documentation, the organizations said.

"It's a partnership with Yale where we're aiming to provide this technology to every single clinician there in short order. And that will allow those clinicians to focus on the most important person in front of them, the patient. Over time, we're planning on this partnership with Yale where we can get creative together and build new functionality," Rao said.

"In 2024, the floor for anyone who's trying to build with AI is higher than ever before," Rao noted. "What is becoming clearer in healthcare is that when you've got science at the center, when you know how to integrate deeply and you've got those go-to-market partnerships and when you've got clinicians at every layer of the company, and you also have those Ph.D.s and professors who can make this feel like everything is possible, we cannot see the ceiling on what we can achieve and the impact that we can create."

He added, "What differentiates us is our ability to build technology that can scale across health systems. Being able to build technology that can scale across specialties requires an ability to control your stack, so our ability to control that part of the stack, own it and improve it over time, makes a huge difference. That kind of speed of improvement is something that you can only do when you've got those ingredients, and those ingredients are expensive. Now with this capital, we'll be able to continue to recruit the mission-driven folks who want to change healthcare for the better."

Rao also pointed to Abridge's proprietary data set to train and build AI models for large health systems. "We can be that generative AI partner for everyone in the industry who's looking to completely transform with the care delivery experience can look like," he said.

The company's technology sits at the intersection of two massive trends in healthcare—clinician burnout and accelerated adoption of generative AI.

Abridge is being propelled by a "tornado of tailwinds," Rao noted.

"We have a public health emergency right now, with two out of five doctors saying they don't want to see patients in the next two to three years. The level of burnout, staffing shortages and labor shortages that we continue to see in the marketplace are leading to rural hospitals shutting down. There's such a huge supply-demand mismatch out there in our communities," he said. "We need technology in a way that we haven't recently needed technology in healthcare. We absolutely need something that can assist and augment our clinicians to deliver care to the people who need it. And that's where this technology comes in."

Abridge's technology is live in 55 specialties, he noted. "We are automating well over 92% of the clerical work that clinicians have to do. We're saving people sometimes up to three hours a day. We're live across different sites and settings, not just outpatient clinics, but also urgent care clinics and ERs. And some of our partners are starting to use us in the inpatient setting as well."

A recent survey by Elation Health found that 67% of primary care physicians have not yet tried an AI-powered medical scribe solution and are looking to electronic health record vendors to guide them to the best option that integrates with their system.