Shiv Rao, who appears in this year’s AI Power Index, has built one of the most important applications of AI in medicine today. Rao, a practicing cardiologist turned founder, saw firsthand how much the burden of modern medicine hinges on the complexity of paperwork surrounding care. After a long day with patients, often hours spent charting and documenting, burnout is at record levels across the industry. His health tech startup, Abridge, set out to solve this problem by transforming one of healthcare’s richest but underutilized data sources: doctor-patient conversations.
Founded in 2018, Abridge uses ambient artificial intelligence to convert verbal communications between clinicians and patients into accurate, real-time clinical records that are integrated directly into electronic health records. What started as an ambitious experiment in language understanding has become a $5.3 billion enterprise trusted by more than 150 major health systems in the United States. In an industry where most startups position themselves as disruptors to the major EHR players, Abridge has chosen to partner with Epic to embed its technology into existing workflows rather than rebuild them. Health systems using Abridge report more than 80% reduction in after-hours workload and significant increases in clinician satisfaction. Yet Rao insists the real goal is to restore empathy and connection to clinical practice. “It’s not just about cutting costs,” he said, “it’s about bringing humans back into health care. ”
You are a practicing cardiologist who founded an artificial intelligence company that is currently valued at $5.3 billion. What issues do you see in hospitals that Silicon Valley tech entrepreneurs are completely ignoring?
As a physician, there is nothing more heartbreaking than spending the whole day helping patients go home only to be handed hours of paperwork. I knew the key was to unlock what was being said in real time in the exam room: gathering intelligence while you were talking. Now, Abridge’s enterprise-grade contextual AI is trusted by more than 150 of the largest and most complex health systems in the United States, and we are extending the value of clinical conversations across workflows for providers, payers and patients.
Abridge converts doctor-patient conversations into clinical notes in real time. How often does artificial intelligence misunderstand something crucial? How do you handle these errors?
Science is at the heart of Abridge. Each note is linked to a transcript of the conversation so clinicians can verify evidence before signing off. We run an ongoing quality assurance cycle and shipping model only after an on-site assessment by a clinician. The clinician remains the final signatory. Abridge’s groundbreaking model ensures AI artifacts are eliminated from draft documents, resulting in market-leading accuracy that clinicians can trust. Abridged It captured 97% of fictional content or unsubstantiated claims in draft clinical documents, outperforming GPT-4o which only captured 82%.
You work with Epic, not try to replace them. Most healthcare startups hope to disrupt the electronic medical records giant. Why choose integration over disruption?
Our partnership with Epic remains focused on solving the issues that matter most to health systems, clinicians and patients. By focusing on the conversations themselves, we can achieve better workflows, deeper integrations, and meaningful results at scale. We are meeting with clinicians and this is just the beginning of the conversation.
Your platform now serves more than 150 health systems, growing 50% in just four months. What’s driving this rapid adoption—physician burnout or genuine efficiency gains?
It’s both. Leaders feel the urgency to address burnout, but adoption will only be sustained if effectiveness truly exists. Abridge integrates seamlessly with EHR workflows—across care settings, over 55 specialties, and over 28 spoken languages. This year alone, Abridge will support more than 50 million medical conversations with clinicians, helping to reduce burnout by up to 60% to 70%. At Sharp Healthcare, clinicians using Abridge report 83% reduction In the effort of writing notes. exist Lee Health Center of Florida86% of clinicians reported a reduction in their after-hours workload.
Administrative costs account for 30% of health care spending. How much waste could you actually eliminate if AI documentation worked perfectly?
The 30% figure is shocking, but it’s not just about cutting costs, it’s about getting humans back into healthcare. For Abridge, return on investment includes more than just improving relative value units of work, improving operational efficiencies, and reducing clinician recruitment and retention costs. It will also happen when clinicians feel happy again, when patients feel heard, and when health systems eliminate impossible backlogs and unsustainable workflows.
You are expanding from general documentation to pediatrics and urgent care. What makes medical AI different between specialties – isn’t documentation just documentation?
A professional background changes everything. In pediatrics, the “historian” is often a parent or multiple caregivers, so the AI must attribute who said what, capture developmentally appropriate histories and encode family-centered plans. In emergency medicine, speed, conciseness, and structured clinical reasoning are important. Categorization prompts, timestamps, and question-oriented summaries must be accurate. We worked with Children’s Hospital and Emergency Room teams to build professionally tuned models and workflows, so the notes reflect how clinicians think and document, rather than a one-size-fits-all template.
