Inspiration
In remote communities, people do not miss care because they do not care. They miss it because systems fail. Long travel distances, limited access to specialists, and complex digital tools make it easy for follow-ups to fall through the cracks, especially for seniors and underserved populations. While researching rural healthcare access and care coordination, we found that missed referrals and poor follow-up tracking are a major cause of avoidable emergency hospitalizations. We were inspired to explore whether a simple, voice-first experience could reduce friction in care continuity.
What it does
Beacon is a voice-first patient care assistant designed for remote communities. Patients can speak naturally to manage appointments, ask about upcoming visits, access basic medical guidance, receive mental health support, and communicate with a nearby hospital. On the provider side, nurses can track referrals, see missed appointments, and coordinate follow-ups through a lightweight dashboard. Beacon demonstrates how care continuity can be handled through conversation instead of complex interfaces.
How we built it
We built Beacon as a React and TypeScript web app with a simulated backend. The patient experience is powered by Gemini for natural language understanding and response generation, allowing users to interact with the system through voice. During design, we researched rural healthcare delivery, closed-loop referral models, and voice-first accessibility patterns. These findings informed our focus on missed appointment tracking, simplified scheduling, and minimizing UI complexity for patients. Data persistence is handled with LocalStorage to simulate a database, and browser speech APIs handle voice input and output.
Challenges we ran into
Designing a voice-first experience that felt natural without overwhelming users was challenging. Our research showed that many seniors struggle with complex digital interfaces, so we had to simplify flows while keeping them functional. We also had to add safety guardrails to avoid unsafe medical advice. Balancing realism with hackathon constraints was another challenge. Building a believable end-to-end flow without real SMS, phone calls, or a backend forced us to simulate many systems while keeping the demo understandable.
Accomplishments that we are proud of
We are proud that Beacon works end-to-end as a usable voice-first healthcare assistant. Patients can actually schedule and reschedule appointments through conversation, receive supportive mental health interactions, and get basic care guidance. We also built a closed-loop referral flow where missed appointments are detected and surfaced to providers, reflecting what we found in research about the importance of care coordination in rural health systems.
What we learned
Through research and building Beacon, we learned that care coordination failures, not just staffing shortages, are a major driver of poor outcomes in rural healthcare. We also learned that voice interfaces can meaningfully lower barriers for seniors and people with low tech literacy. Finally, we learned how important it is to design health-related AI features with clear boundaries, transparency, and safety constraints.
What is next for Beacon
Next, we would move Beacon to a real backend with secure patient records, real scheduling integrations, and reliable data storage. We would add real communication channels like SMS and voice calls for offline access, integrate with hospital systems, and expand language support for Indigenous communities. Long-term, we envision Beacon as a mobile-first, offline-friendly care coordination layer for rural and northern health systems.
Log in or sign up for Devpost to join the conversation.