Knowtex

Founded by AI researchers to build this product and nothing else. Notably the founders spent a year working as human scribes alongside physicians while building it, which is domain grounding rather than model work but explains why the output targets notes, codes and orders together rather than notes alone.

Generates diagnostic and billing codes and orders in real time from the conversation, with the chief executive describing order entry as the natural extension of ambient documentation and the focus of the 2026 roadmap. That places Knowtex on the same side of the split as Oracle Health and Sunoh.ai rather than with the vendors that only suggest, and no confidence threshold, order level accuracy rate, abstention behaviour or review gate description was located. The reasoning the chief executive gives publicly, that a system accurate enough to generate a note is accurate enough to propose the orders implied by it, is a claim that should be substantiated with an order level accuracy figure rather than asserted by analogy.

No accuracy figure, model card, named models or evaluation methodology located. Specialty specific tailoring is claimed as an architectural property but not described. The competitive federal evaluation noted under evidence implies performance was measured by an external party, which makes the absence of any published performance figure more conspicuous rather than less.

Unusually strong independent validation for a company of this size, though not clinical outcome evidence. Knowtex placed in the top three of the 2024 Department of Veterans Affairs AI Tech Sprint for Ambient Scribe, a structured competitive evaluation run by the government rather than the vendor, and was subsequently awarded a 15 million dollar VA contract through federal procurement, deploying from October 2025 across the largest integrated health system in the United States. Neither of those is a testimonial and neither can be commissioned. An earlier pilot ran at the University of Rochester Medical Center from February 2023. Held at B rather than A because no published outcome data, accuracy result or peer reviewed study was located; what is evidenced is that competent external parties assessed it favourably, not what it achieved clinically.

Not assessed. No statement on audio or transcript retention, de identification or training use was located in this pass. Worth asking specifically in light of the VA deployment, since federal health data handling obligations are stricter than commercial HIPAA baseline and the answers may already exist in contract form.

Not assessed. No published business associate agreement posture located in this pass.

Not assessed, but with a strong lead worth following. No named or dated attestation and no trust centre was located. However, deploying across the Veterans Health Administration at national scale would ordinarily require a federal authorisation to operate and the associated security assessment, which is a materially higher bar than a commercial SOC 2. This index does not grade a certification it has not seen evidenced, but a buyer should ask directly for the VA authorisation documentation, because it likely exists and would be the strongest security artefact in this category.

Not a regulated medical device and none claimed. Note the direction of travel: order generation moves closer to regulated territory than note generation, and the classification question is worth asking as that roadmap ships.

Nothing located. No fairness statement, no subgroup analysis and no accent or dialect performance disclosure. That gap carries particular weight given the deployment population: the Veterans Health Administration serves a patient group with distinctive demographic, geographic and clinical characteristics, and a documentation model's performance across that population is a question the deployment itself makes answerable.

Positioned as EHR agnostic and deeply integrated into workflow, accessible from mobile, tablet or computer with no dedicated hardware, and designed to fit team based care. The VA deployment is meaningful evidence that integration works at scale in a complex federal record environment. Held at B because no named commercial EHR integrations, certification programme listings or write back architecture were enumerated or verified in this pass, so breadth outside the federal deployment is asserted rather than demonstrated.

Not assessed. No hosting region, residency configuration or sub processor detail was located, though a federal deployment implies a government cloud posture worth confirming.

No rate card, but an unusual accidental disclosure worth using. Because the VA award is a federal procurement, its value is public at 15 million US dollars, which gives buyers something almost nothing else in this category offers: a real, externally verified enterprise scribe contract figure to benchmark against, rather than a vendor quoted range. That is transparency by operation of procurement law rather than by vendor choice, so it says nothing about how this vendor prices commercially, but it is a genuine data point in a category where every enterprise number is otherwise negotiated in private.

Specialty specific by design rather than one model behind templates, with oncology workflows named explicitly, and the VA deployment spans 170 medical centres and 1,193 outpatient clinics covering effectively every specialty and care setting in a national system. Held at B because that breadth currently rests on a single customer relationship rather than a distributed commercial base, and no specialty count or language coverage is published.