HaemoLog: Turning Connections into Lifelines

Inspiration

Every 2–3 weeks, Thalassemia patients in India need life-saving blood transfusions. Despite efforts by organizations like Blood Warriors, finding compatible, repeat donors remains a race against time — often managed manually via WhatsApp groups and calls.

We were inspired by the power of community-driven health systems and saw an opportunity to enhance Blood Warriors’ existing solution — Blood Bridge — by evolving it into a smart, scalable, Agentic AI-first ecosystem that improves outcomes, builds trust, and saves time.

What it does

HaemoLog is an AI-powered upgrade to the Blood Bridge system that automates and enhances the blood donation journey using a WhatsApp/Telegram chatbot and Azure-powered backend.

Key Features:

Donor-Recipient Matching Engine: Matches based on blood group, proximity, last donation date, and AI-predicted availability.
Smart Retention System: Reminds eligible donors every 2–3 months, tracks milestones, and encourages repeat donations.
Event Automation: Blood Warriors can create and promote donation drives via the chatbot.
Health-Aware Matching: Prioritizes same/similar blood type donors, beneficial for Bone Marrow Transplants.
Consent-Based Sharing: Donor info shared only with approval, ensuring complete data privacy.
Built for All: Supports donors with 12.5+ g/dl haemoglobin (including carriers), and multilingual engagement.

How we built it

We divided the development into four main components:

Conversational Chatbot (WhatsApp/Telegram)
- Developed using WhatsApp Cloud API and Telegram Bot API
- Integrated with donor registration, event posting, and donor-patient linking flows
Backend Matching Logic
- Built using Python FastAPI
- Stores data in Azure PostgreSQL
- Uses Redis for fast real-time filtering
Predictive AI Layer
- Trained an XGBoost model on synthetic donor behavior data
- Predicts likely availability based on past donation habits and time patterns
Azure Deployment
- Hosted backend on Azure App Service
- Used Azure ML Studio for model deployment
- Enabled scalability for future donor regions

Challenges we ran into

Data Simulation: With no direct access to Blood Warriors’ real data, we had to simulate donor and patient datasets.
WhatsApp Approval Workflow: Template-based messaging slowed early testing phases.
Health-Aware Logic: Implementing checks like haemoglobin thresholds and donation eligibility windows added complexity.
Balancing Automation & Consent: Ensuring privacy while still automating donor-patient matching was a tightrope walk.
Scaling Logic: Regional donation demands and different blood group distributions made matching logic dynamic and complex.

Accomplishments that we're proud of

Successfully built a smart donor matching engine integrated into WhatsApp.
Enabled health-aware, consent-driven donor selection logic.
Developed a predictive model to support donor availability forecasting.
Designed a self-sustaining ecosystem that reduces Blood Warriors’ manual workload.
Ensured alignment with real-world transfusion constraints, including carrier eligibility and haemoglobin checks.
Deployed core infrastructure on Azure, as required by the hackathon guidelines.

What we learned

Building for trust, empathy, and privacy is as critical as building for speed.
Existing solutions like Blood Bridge provide valuable foundations that should be evolved — not replaced.
Predictive AI can improve donor response rates if used with thoughtful UX and ethical consent practices.
A familiar platform like WhatsApp drastically increases usability and reach, especially in low-tech environments.

What's next for HaemoLog

Full Integration with Blood Warriors’ Donor & Patient Databases
Role-Based Admin Dashboard for Blood Warriors to view and manage regional data
Donor Feedback Loop – to track experience and refine matching quality
Voice/IVR Bot Layer – for rural donors with no internet or low literacy
Blood Supply Forecasting Engine using regional trends and AI
Gamified Donor Community – badges, certificates, and patient stories to build a lasting ecosystem of empathy and action

“HaemoLog isn’t just an app — it’s a movement toward a smarter, faster, and more compassionate blood donation network powered by the people and guided by AI.”