HemoConnect – Smart AI Platform for Thalassemia and Blood Donation Management

Inspiration

In India, over 100,000 Thalassemia patients require lifelong blood transfusions every 2–3 weeks. Despite efforts, hospitals still struggle with managing donor shortages, missed transfusions, and lack of predictive tools. Inspired by this critical need, we built HemoConnect – a smart, AI-powered system that connects donors, patients, and hospitals in real time to ensure timely, trusted, and efficient blood transfusions.

What it does

HemoConnect is an AI-first platform that:

Predicts donor availability and matching for Thalassemia patients.
Schedules and reminds patients about transfusion appointments.
Uses WhatsApp/Telegram bots for donor outreach and coordination.
Supports voice-based assistants in regional languages for accessibility.
Integrates with e-RaktKosh, Blood Bridge, and Aadhaar for secure, real-time data sync.
Promotes donor retention with trust scores, nudges, and personalized follow-ups.
Works with federated learning to keep health data private and on-device.

How we built it

MCP Server Architecture with modular AI agents.
Donor Matching AI Agent: Uses vector embeddings of donor history + blood type.
Triage & Scheduling Agent: Uses LLM + rules to recommend slots.
Bot Assistant: Built with Botpress integrated with WhatsApp Business API & Telegram Bot API.
Voice Assistant: Deployed Whisper + lightweight TTS in Tamil, Hindi.
Privacy: Used federated learning for sensitive health data.
UI: React + Tailwind for clinics; WhatsApp bot UI for donors/patients.

Challenges we ran into

Donor matching precision using limited, incomplete health data.
Balancing privacy and personalization in AI recommendations.
Integrating across multiple government databases (e-RaktKosh, Aadhaar).
Handling multi-lingual communication across voice/text.
Re-engaging inactive donors and increasing retention.

Accomplishments we're proud of

Real-time AI donor-patient matching using a smart agent architecture.
Live integration with WhatsApp bot and Telegram for outreach.
Built a low-bandwidth-friendly voice assistant in local languages.
Modular design ready for state-wide deployment in blood banks.

What we learned

AI in healthcare is only useful when paired with empathy and trust.
Modular AI agents can scale better and are easier to update independently.
The key to retention is personalized, non-intrusive communication.
Importance of privacy-first design for health tech adoption in India.

What's next for HemoConnect

Pilot with state government blood banks and Thalassemia NGOs.
Expand support for other rare blood disorder patients.
Introduce donor reward systems (digital wallets/coupons) to improve retention.
Add ambulance scheduling and link to nearby transfusion centers.
Collaborate with India's Ministry of Health and integrate with CoWIN-like platforms.