Artificial intelligence is not new to medicine. Hospitals and startups for years promoted machine learning as the piece of technology that would finally bring medicine into the modern age. But while AI became commonplace in industries such as retail and finance, its use in clinical practice has lagged behind. Perhaps that is about to change.
Among the most pressing issues doctors grapple with these days has little to do with diagnosing rare diseases or operating with great accuracy. It’s documentation. By most measures, physicians spend nearly double their time with electronic health records as they do with patients. It’s led to burnout, frenetic office visits, and mounting frustration on either side of the exam table.
A new generation of companies is betting that AI can help. Among them is ScribeAI, a startup co-led by Kyle Robertson and Matt Holmes. Their tool uses natural language processing to capture and structure doctor–patient conversations. The system produces clinical notes, assigns billing codes, and generates summaries—all tasks that previously required hours of manual input. The promise is straightforward: let software handle the forms so doctors can focus on care.
For Kyle Robertson, the project builds on years of navigating healthcare’s pain points. He previously founded Cerebral, which brought mental health services online at scale, and later launched Zealthy to expand access to telemedicine. With his incubator, Revolution Venture Studios, he has turned his attention to one of the least glamorous but most consequential aspects of the system: administration. Sources inside the firm say that ScribeAI has already generated an eight-figure business pipeline. More telling is feedback from providers, who say that every week, hours are being freed up by the technology. More hours, they say, are being devoted to better patient visits and more detailed discussions with patients about their health issues.
ScribeAI is part of a broader healthcare AI trend that’s all about human-centric support, rather than replacement. Flatiron Health is helping oncologists to analyze huge datasets related to cancer. Cleerly applies algorithms to cardiovascular images. Enlitic is transforming radiology. Broadly, technology is being positioned as a collaborator—pulling out patterns and insights that would go uninterpreted without it, but still yielding to clinicians’ judgment. Flatiron has even provided proof that large language models can track cancers’ progress at a level of granularity similar to human specialists.
There are reasons to be wary of new technologies. Electronic health record implementation was supposed to simplify workflow, but for most physicians, it brought added frustrations. This background is why many are suspicious of today’s crop of AI solutions. What’s different about ScribeAI and its peers is that they focus on solving small, nuts-and-bolts problems, instead of redoing the system from scratch.
Whether all these tools ultimately result in improved patient outcomes remains to be seen. For now, though, clinicians interacting with Kyle Robertson’s ScribeAI report something novel in medical technology: software that actually streamlines their workloads. In a profession where time is the most desperately scarce of all commodities, that’s no small achievement.