KeyCare Deploys NVIDIA Nemotron to Power AI-Driven Clinical Intake at Scale

Key Takeaways

• NVIDIA Nemotron powers KeyCare’s AI intake agent.
• Provider efficiency up 15‑30% for urgent‑care visits.
• AI agent summarizes Epic data before patient encounter.
• Integrated Epic model enables safe, scalable AI deployment.
• Expansion planned to routing, operations, patient support.

Summary

KeyCare, the nation’s only Epic‑based virtual care group, has launched NVIDIA’s Nemotron large language model in production to power its AI‑driven patient intake agent. The agent, embedded in KeyCare’s virtual waiting room, captures and summarizes clinically relevant Epic data before the encounter, allowing clinicians to focus on assessment. Early data show a 15‑30% boost in provider efficiency for common urgent‑care visits such as respiratory and urinary‑tract infections. KeyCare plans to extend NVIDIA‑powered AI agents to routing, operations and patient support across its 50‑state network.

Pulse Analysis

The healthcare sector has long grappled with the tension between rapid telehealth adoption and the persistent burden of documentation. KeyCare’s decision to embed NVIDIA’s Nemotron LLM directly into its Epic‑centric workflow marks a pivotal shift, moving AI from experimental pilots to a production‑grade tool that clinicians trust. By leveraging a model known for strong reasoning and consistency, KeyCare ensures that the AI‑driven intake agent can interpret patient‑reported symptoms, cross‑reference them with existing Epic records, and generate concise summaries that are immediately actionable.

Early performance metrics underscore the operational impact. For high‑volume urgent‑care conditions—upper respiratory infections and urinary‑tract infections—the AI intake agent has delivered a 15‑30% improvement in provider efficiency, translating into shorter visit times and reduced cognitive load for clinicians. Physicians report clearer patient context and faster decision‑making, while patients benefit from a smoother virtual experience. This efficiency gain not only enhances revenue per encounter but also frees clinician capacity to see more patients, addressing access challenges that have plagued virtual care platforms.

KeyCare’s vertically integrated model—combining a nationwide clinician network, Epic’s electronic health record, and proprietary AI tools—provides a blueprint for scaling advanced LLMs across health systems. The company’s roadmap includes extending Nemotron‑powered agents to appointment routing, operational optimization, and patient support, further embedding AI into the care continuum. As more providers seek to operationalize AI rather than merely showcase demos, KeyCare’s approach signals a broader industry trend toward tightly coupled, Epic‑native AI solutions that can be deployed safely at scale, reshaping the economics and quality of virtual healthcare.