Knowtex as Healthcare Data Flywheel
Caroline Zhang, CEO of Knowtex, on going from AI medical scribe to AI lab
This is a claim about control of workflow, not just note taking. If Knowtex sits in chart prep, documentation, coding, orders, and follow up, it touches the same clinical moments where the EHR collects and moves data. That makes it harder to remove during an EHR switch, because the hospital is relying on it not only to write a note, but to keep visits, orders, and downstream documentation moving across systems.1 2 3 4 5
- Knowtex is already positioning itself as broader clinical AI software, covering chart preparation, documentation, coding, orders, and clinical insights. That matters because a scribe that only outputs text is easier for Epic, Oracle, or Microsoft to copy or bundle than a system writing structured data back into the care workflow.1 2 3
- The VA contract makes the point concrete. The enterprise ambient scribe contract carried a shared ceiling of about $775.7M and expected API level integration with both VistA or CPRS and Oracle Health, including free text and structured write back. That means the buyer was procuring workflow infrastructure, not just a transcription tool.4 5 6
- This is also where the strongest competitors are heading. Abridge expanded from ambient notes into orders, coding, prior auth, and other Epic based workflows, while Ambience emphasizes writing into discrete EHR fields instead of handing clinicians a text blob. In this market, the winning product becomes part of how the record gets created, checked, and monetized.7 8 9 10
The next phase is a fight to become the layer that survives EHR transitions and native bundling. Companies that can keep structured integrations working across Epic, Oracle, and legacy systems, while turning conversations into usable orders, codes, and clinical actions, will look less like point tools and more like permanent operating software inside the hospital.2 3 4 7 9