VA awards contract to Knowtex for AI scribes

Little-known startup Knowtex has emerged as one of just two primary winners in the Department of Veterans Affairs’ $776 million contract for ambient scribe technology. In a recent conversation, CEO Caroline Zhang discussed the VA deal and how the company stands apart from competitors in the field.

Zhang explained that she and co-founder Jocelyn Kang spent six months embedded in clinical teams working as medical scribes to learn the business. They spent time in oncology and orthopedics, respectively, to understand the end-to-end workflow inside the clinic.

Origin Story and VA Contract

This experience informed the initial platform that Knowtex built, which involves everything from knowing what makes a note great to building specialty-specific AI models and end-to-end workflows. Zhang noted that the VA is an opportunity to lead the way in healthcare transformation with AI technology, being the largest U.S. healthcare system.

When the VA issued a 2024 AI tech sprint, Knowtex saw it as the perfect opportunity to demonstrate their product. They were competing against a field of 150 competitors, and the VA evaluated product capabilities and integration. In July 2025, Knowtex was awarded a $15 million pilot award, along with Abridge, and they were able to deploy in separate regions of the country.

Knowtex has had extraordinary success with their VA product adoption, with their product already in 310 clinics in the VA and 79 VA hospitals. They brought all of this live in just six months, which Zhang described as a historic speed of roll-out.

Product Deployment and Scope

The product Knowtex built commercially services over 200 specialties today, but in the VA, it’s a controlled roll-out. They started in primary care and did 13 specialties in the pilot phase, but the enterprise contract covers all VA clinicians. Zhang noted that the emergency department is an area on the roadmap for the VA that they’ve supported commercially before.

Knowtex has also done demos about nursing workflows, and they know that’s also an area of interest for the VA. Zhang credited their success to working with a good chief medical informatics officer and building a specialty-specific model for each clinical specialty workflow. They have worked within the VA’s area of interest, which includes various clinical settings.

One of Knowtex’s early customers was the University of Rochester Medical Center, an academic medical center with an Epic system. When they integrated, they did an initial pilot, and the results were strong. The team in orthopedic surgery really felt the difference between Knowtex’s tool and more generic AI scribe tools.

Knowtex has released a benchmark tool, called KnowBench, to measure how much clinician effort they can reduce. Zhang explained that KnowBench is about defining benchmark criteria around real clinical adoption and impact, looking at deployment experience nationally and measuring across notes, codes, orders, and decision recommendations.

Knowtex is seeing a high level of acceptance, with 97.99% of clinicians accepting and putting their signature on the output. Zhang noted that the goal is to continue to advance this high bar, and they’d like to hold general-purpose models to this standard as well.

The idea behind KnowBench is that a chief medical informatics officer could use this tool to assess not just Knowtex’s product but a bunch of different clinical AI tools to measure the impact they’re having. Zhang believes that this will bring awareness to the expectations for healthcare adoption of AI for the long-term.

In the VA, Knowtex’s product is being used in a variety of clinical settings, including primary care and specialty care. The company is also exploring the use of their product in other settings, such as the emergency department and nursing workflows.