What if doctors could stop worrying about taking notes and just focus on their patients?

DeepScribe makes that a reality—turning medical conversations into structured notes automatically, so doctors spend less time on paperwork and more time actually practicing medicine.

We caught up with Katie Christiano, Solutions Consultant at DeepScribe, to hear how their team uses Vellum to create medical notes from patient visits—and keep relieving doctors of the note-taking burden.

Here’s how they made it happen.

The before: DeepScribe’s challenges

Documentation has long been a burden for medical professionals, taking up valuable time and shifting focus away from patient care. The growing administrative load has made it clear that healthcare professionals need smarter solutions to handle documentation efficiently.

That’s where DeepScribe comes in—using ambient AI to listen to patient visits and generate medical notes automatically, easing the cognitive load on practitioners.

But building trust in AI-driven documentation requires continuous improvement.

To keep delivering a seamless experience, DeepScribe needed a way to quickly incorporate feedback—refining its AI while strengthening trust with the clinicians who depend on it daily.

Vellum makes continuous AI improvements easy

In healthcare AI solutions, precision isn’t optional—it’s critical.

Every update must be rigorously tested to ensure accuracy, safety, and compliance.

Today, DeepScribe uses Vellum to quickly and reliably make updates to certain AI features. With Vellum, they can continuously improve their AI medical scribe, serving thousands of clinicians across various specialties.

"With Vellum, we’re in control. We can move fast on feedback, improve the product, and build trust with the clinicians who rely on us every day." Katie explains, sharing how it has helped their team refine and perfect their AI solution for generating SOAP notes.

Today, DeepScribe uses Vellum’s product to:

1/ Enable precise AI iteration and testing. By using Vellum Workflows, DeepScribe is able to systematically test models and refine logic efficiently.

2/ Rapidly experiment without disruption. When it comes to rolling out updates, Vellum facilitates quick pivots and risk-free testing without disrupting the customer experience.

Impact: DeepScribe ships AI updates 40% faster

After more than a year of using Vellum, DeepScribe has seen increased efficiency, greater accuracy, and the customer trust that comes with clinicians seeing quick responses to their requests and needs.

By improving AI iteration and reducing cross-functional dependencies, they’ve significantly improved how they deploy updates.

“The biggest impact of Vellum is increased customer trust, driven by our ability to quickly and effectively act on their feedback. Vellum has streamlined workflows and sped up project delivery overall, anywhere from 20-40% time savings based on task complexity.“ - said Katie.

With Vellum, DeepScribe has:

• Increased customer trust: Rapid and precise updates provide clinicians more accurate documentation.
• Faster iteration cycles: Updates that once took months now roll out in days.
• 20-40% time savings:  Streamlined workflows free up valuable resources.

Incorporate feedback loops with Vellum

Vellum has transformed DeepScribe’s ability to refine and scale AI-driven medical documentation — and we’re proud to support such an amazing team.

If you want to replicate this process in your company to help your teams move faster and provide more frequent improvements in your AI solutions, we can help.

Book a demo with us and learn what you can achieve with Vellum’s suite of products designed to help with developing, evaluating and continuously updating of any AI features that you want to build.