From AI Scribe to Action Layer

Diving deeper into

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

Interview
They’re always stuck in cycles of EHR upgrades or just getting the system up to parity.
Analyzed 4 sources

This is the core opening for AI scribe startups, hospitals already spend so much time keeping the record system running that the record system rarely becomes the place where new AI shows up first. Knowtex is selling an overlay that sits on top of Epic or Oracle, pulls chart data, listens to the visit, and then drafts notes, codes, orders, and trial matching faster than an EHR vendor can ship those features inside a broad platform upgrade cycle.1 2 3 4

  • Knowtex was built around being EHR agnostic. In the interview, the company describes winning VA work partly on integration, and supporting hospitals even during transitions from legacy systems to Oracle or Epic. That matters because an overlay can stay in the workflow while the underlying record system changes.1 2
  • The market has already shown that EHR access is powerful but not sufficient. Abridge grew quickly by becoming Epic’s closest scribe partner and reaching 60,000 plus clinicians, yet even that model still expands outside plain note taking into coding and revenue cycle tasks. The note alone is no longer enough.3
  • Knowtex is using the same wedge to move beyond documentation. Its product starts before the visit with chart prep, runs during the visit with ambient capture, and finishes after the visit with draft orders, billing codes, and care coordination. That is a much larger workflow surface than an AI note button inside the EHR.1 2

The next phase is that the EHR becomes more like the system of record, while the AI layer becomes the system of action. Companies that can sit across multiple EHRs, ship monthly, and turn visit data into billing, decision support, and research workflows will keep taking share from slower platform vendors.1 3 4