Inspiration Our company delivers virtual healthcare services to elderly patients with chronic illness. As our patient base grows, we've had to continuously hire more nurses and case managers. However, with the ongoing healthcare staffing shortage in the U.S. and increasing burnout among providers, this approach is becoming unsustainable. To address this, we’re leveraging AI to handle time-consuming administrative tasks and involving human professionals only when their expertise is truly needed.
What it does: In this use case, we leverage AI to transcribe telemedicine call recordings into text, generate a concise summary, highlight key health symptoms, identify necessary referrals, and create follow-up tasks for case managers. This AI solution saves at least 20 minutes per patient per day, significantly improving efficiency and care coordination.
How we built it: We used four tools to build our solution: Anthropic (MCP servers), Auth0 (authorization), AWS Transcribe, and n8n (workflow automation). We created a new agent on top of our platform, which sends the call recording to the agent, processes it, and then passes the results back to our platform.
Challenges we ran into: We had a hard time integrating the transcribe into the MCP
Accomplishments that we're proud of: Not only did we transcribe the audio recording, but we also used AI to generate an analysis and identify follow-up tasks or questions for the case manager to address in the next patient call.
What we learned: We can use the MCP Agent Server to create and manage multiple agents, and leverage powerful tools like n8n to visually design workflows and streamline agent integration.
What's next for Lucy - AI Agent for Case Managers in virtual healthcare: Next, we plan to use Apify to scrape insurance websites and retrieve a list of in-network mental health providers. This will enable us to build a referral AI agent that can efficiently identify the most appropriate healthcare provider covered by the patient’s insurance.
Log in or sign up for Devpost to join the conversation.