From Scribe To AI Overlay

Diving deeper into

Caroline Zhang, CEO of Knowtex, on going from AI medical scribe to AI lab

Interview
You used to have 10-year innovation cycles in healthcare
Analyzed 4 sources

The main shift is that AI is turning healthcare software from a slow replacement market into a fast overlay market. Instead of waiting years for an EHR vendor to rebuild core workflows, startups can plug into the record system and ship new products in months. In this category, the winning product is no longer just a note taker. It is the fastest tool that can turn a live visit into notes, codes, orders, and next steps inside the clinician workflow.1 2 3

  • Knowtex was built around this overlay model from the start. It pulls chart context before the visit, records the conversation during the visit, and generates codes, orders, care coordination tasks, and trial matching after the visit. That is a much shorter product loop than a full EHR release cycle.1 2
  • The market has already rewarded faster AI native vendors. Abridge used Epic and athenahealth partnerships to expand from about 8,000 to more than 60,000 clinicians in 18 months, reaching an estimated $100M ARR by May 2025. Ambience reached an estimated $30M ARR by May 2025. Both expanded beyond documentation into adjacent workflow and billing tasks.3 4
  • This also explains why standalone scribing is fading so quickly. Once the conversation is captured, the real economic value comes from everything downstream, billing codes, orders, prior auth support, and clinical decisions. The note is becoming the entry point, not the end product.1 3

The next phase is that EHRs look more like systems of record, while AI layers become the systems clinicians actually work inside. Companies that combine deep integration with rapid shipping will keep moving outward from documentation into revenue cycle, decision support, and research workflows, and that is where the biggest healthcare software value will accrue.1 3