BloodBridgeAI

Inspiration

We were inspired by the real delays hospitals face during urgent blood requests. In many cases, the challenge is not just finding donors, but quickly identifying the right eligible donor, reaching them fast, confirming availability, and estimating arrival time without relying on manual calls and spreadsheets. BloodBridgeAI was built to make that emergency coordination process faster, smarter, and more reliable.

What it does

BloodBridgeAI is an AI-powered emergency blood donor coordination system for a hospital. When a hospital raises an urgent request, the system finds compatible donors, filters out donors who are not eligible to donate at that moment, ranks the best candidates, and contacts them through an AI voice call assistant. The assistant asks whether they can come immediately, captures ETA, checks if they are fit to donate, optionally requests consent-based location sharing, and sends the hospital map link by SMS after confirmation. The hospital can then track the full request through a live dashboard.

How we built it

We designed BloodBridgeAI as a multi-agent system with three core agents: a Matching Agent, an Eligibility Agent, and a Call Assistant Agent. The frontend is planned as a React-based hospital dashboard, while the backend uses FastAPI for request orchestration, donor matching, webhook handling, and agent services. PostgreSQL is used for structured donor, request, and call-response records.

Challenges we ran into

One major challenge was designing a system that feels intelligent but still remains safe and operationally realistic. We had to think carefully about donor eligibility rules, uncertain call responses, duplicate call events, multiple donors accepting for the same request, and situations where no donor is available within the initial radius. Another challenge was balancing useful features like ETA estimation and location sharing with privacy requirements, making sure location sharing is always optional and consent-based.

Accomplishments that we're proud of

We are proud that BloodBridgeAI is not just a dashboard idea, but a complete end-to-end workflow. It covers hospital request creation, donor matching, eligibility verification, real-time donor calling, ETA capture, optional consent-based location sharing, SMS map delivery, and live dashboard updates. We are also proud of the modular multi-agent design because it makes the system easier to explain, build, and extend in future versions.

What we learned

We learned that in healthcare coordination problems, speed alone is not enough. The system must also be transparent, privacy-aware, and reliable in edge cases. We also learned how to break a real-world workflow into modular AI agents with clearly defined responsibilities, and how supporting tools like voice systems, messaging APIs, dashboards, and backend logic are just as important as the AI itself in building something practical.

What's next for BloodBridgeAI

Next, we want to make BloodBridgeAI more predictive and scalable. Future improvements include predicting which donors are most likely to respond quickly, adding multilingual voice support, improving backup donor standby logic, and integrating with real screening records or blood bank software. Beyond the MVP, the long-term vision is to expand from one hospital to a broader emergency donor coordination network.