End-to-End Clinical AI Assistants
Caroline Zhang, CEO of Knowtex, on going from AI medical scribe to AI lab
This reveals that the winning product in clinical AI is no longer the note, it is the layer that turns one patient visit into prep work, live capture, billing, orders, and follow up. Knowtex starts before the visit by pulling labs, scans, pathology, and genetics into a visit brief, then records the conversation, drafts the note, suggests codes and orders, and can route trial matching work to research staff within seconds.1 2
- This is a shift away from standalone scribing. Across the category, note capture has become easier to copy, so vendors are moving into adjacent steps like coding, orders, prior auth, and care coordination. That is where budget and defensibility increasingly sit.1 2 3
- Knowtex is especially built for complex specialties. It started in oncology, where clinicians juggle chemotherapy regimens, comorbidities, genomic data, and trial eligibility. That makes the product less like a recorder and more like a workflow copilot for high stakes care.1 2
- The deeper asset is data, not just automation. When the system sees chart context, hears the visit, and then observes what the doctor signs off on, it captures a full clinical reasoning record. That creates training data for specialty models and future decision support.1 2
The market is heading toward full stack clinical assistants that sit on top of the EHR and absorb more of the visit lifecycle. Products that can reliably turn conversation into billable, actionable, and clinically useful output will keep expanding from documentation into revenue cycle, treatment guidance, and research workflows.1 2 3