Ask health-focused venture capitalists to name a single top AI startup, and one name comes up time and time again: Pittsburgh-based Abridge, founded before OpenAI was a household name and before LLMs were a household term in Silicon Valley.
In 2019, Shiv Rao, a practicing cardiologist, pitched a startup idea to Andy Weissman, general partner at Union Square Ventures, which Rao called a cross between SoundCloud and RapGenius for medicine.
Weissman said he found it a bit funny to compare a startup AI-powered medical note-taking app to music hosting and song lyrics transcription, but the concept resonated with him.
Rao explained that doctors typically spend up to two hours a day outside of regular work hours typing up notes summarizing what they discussed with their patients that day. Over the years, this administrative work has led to physician burnout, with some doctors even leaving the profession altogether. Rao convinced Weissman that the latest innovations in AI could drastically reduce the time doctors spend on their ever-increasing paperwork burden.
This was years before generative AI took the world by storm and captured the imagination of venture capitalists.
“It was a pretty unusual idea. No one had done it before,” Wiseman said.
But Weissman and the other USV partners appreciated that Rao is not only a physician at the University of Pittsburgh Medical Center, but also spends half his time as a corporate venture capitalist for the health system, investing in medical technology startups. Rao's employees and advisers are also graduates and professors at Carnegie Mellon University, one of the nation's top institutions for engineering and AI research.
“[Shiv] “It was a rare combination of talent: an incredibly ambitious entrepreneur with a vision and a really interesting team,” Weissman says. “It felt unique.”
Abridge also had a basic transcription product that physicians could download for free onto their smartphones to use during patient interactions, and this usage became the basis for Abridge's LLM.
A little over five years after USV led a $5 million seed round in Rao's startup, Abridge, the company has become one of the most talked-about and fastest-growing AI-powered healthcare companies.
While most businesses remain very cautious about adopting AI tools, large health systems have been enthusiastic about signing on with Abridge.
The sales cycle is [health systems] “Abridge implementations can take 18 to 24 months,” Rao said. “We knew what was coming when we started the company.” But with a four-year lead on its virtual scribe product trained on thousands of doctor-patient conversations, and with AI growing at a rapid clip, hospitals are suddenly buying Abridge's products at a rapid clip, in stark contrast to their usual long-term buying behavior: The company has been announcing new health system customers almost weekly since the beginning of 2024.
“We turned this potential energy into kinetic energy almost overnight in January,” Rao said. “We have partners at the University of Chicago, Sutter University, Yale University, Lee Health University, Christus University, Emory University, the list goes on and on,” he said.
AbridgeImage credit: Abridge / Abridge
Large hospitals have not only purchased thousands of Abridge seat licenses, but they also often publish glowing reviews about how the medtech software is changing their physicians' lives. Hospital executives and physicians have described Abridge as “life-changing,” “magical,” and “one of the most important paradigm shifts of our careers.”
One of the biggest criticisms of generative AI is that there are still very few substantive business uses for it, but virtual medical note-taking appears to be a valuable application of this emerging technology.
Drowning in paperwork
“I have my own professional PTSD and war stories of writing notes and doing administrative tasks for hours at night after seeing patients, not only distracting me from what matters most – my patients – but also taking away from my personal life,” Rao said.
By having Abridge record in the background, doctors can avoid the hassle of completing certain fields in the medical record during consultations and focus entirely on the patient.
The return on investment for AI-powered medical scribes is pretty easy to measure, says Dr. Lee Schwam, chief digital health officer at Yale New Haven Medical System, an Abridge customer. That's why so many health systems are looking to adopt AI-powered medical scribes, especially from Abridge. “AI-powered medical scribes are one of the hottest products in the AI space right now,” he told TechCrunch.
As with many health tech administrative tasks, Schwam said the most important considerations when selecting a vendor are price and integration with Epic, the electronic medical record used by most large health systems in the U.S. Abridge, which supports 14 foreign languages including Haitian Creole, Brazilian Portuguese and Punjabi, is often the winner when health systems compare it head-to-head with other AI-powered medical record systems, Schwam said.
Earlier this year, Abridge won the rights to be integrated into Epic. Abridge records the session, and when the doctor stops the recording, “there are notes in English sitting there in Epic, ready to be reviewed, edited, adjusted and acted upon as deemed appropriate,” Rao says.
Abridge appears to have an advantage over competitors such as Microsoft-owned Nuance, as well as Ambiance, Nabla and Suki, but Schwamm said he's not sure the company can maintain its lead in the long term.
“The big question is, do you need a specialized medico-legal master's degree to be successful in this field,” he asked, “or will the big underlying models from GPT-4o, Google, Meta, etc. become so good that they can ingest entire corpora of medical notes and perform similarly?”
This research shows that it is still early days not just for virtual medical record creation, but for most generative AI companies. The pace of innovation is fast and furious, and today's winners could easily lose their advantage.
“Abridge is leading by a length, but it's still early in the race,” Schwam said. “He could hurt his knee and stumble, or he could continue to pull ahead.”
So far, most investors TechCrunch spoke to agree that Abridge is leading the race in AI-powered medical transcription, which is why the company is seeing capital pouring in.
In February, Abridge raised $150 million in Series C funding led by Lightspeed Ventures at a valuation of $850 million.