FIGGY AI

🚀 Inspiration

The rapid growth of India’s gig economy—especially delivery workers on platforms like Swiggy, Zomato, Zepto, and Amazon—has created millions of jobs. However, these workers face unstable income due to factors beyond their control such as rain, floods, pollution, or strikes.

We were inspired to solve a critical gap: 👉 Why is there no instant, fair, and automated insurance for gig workers’ income loss? That’s how FIGGY AI was born.

💡 What it does

FIGGY AI is an AI-powered parametric insurance platform that: 📈 Predicts high-demand delivery zones to increase earnings 🌧️ Detects real-world disruptions (rain, pollution, protests) 💰 Automatically triggers instant insurance payouts 🔐 Prevents fraud using AI + blockchain verification

Unlike traditional insurance: ❌ No manual claims ❌ No long approval delays ✅ Fully automated & real-time payouts 🛠️ How we built it

We designed FIGGY as a modular system with AI + blockchain integration: 🔹 Frontend Flutter (Mobile App UI) 🔹 Backend Node.js + Express (REST APIs) 🔹 AI/ML Models Random Forest → worker behavior analysis XGBoost → risk scoring & demand prediction LSTM + Prophet → time-series forecasting NLP Transformers → disruption detection

🔹 Core Systems 📊 Income & Premium Engine 📍 Demand Prediction Engine ⚠️ Disruption Radar AI 🔐 Proof-of-Work Fraud Detection ⛓️ Blockchain Activity Tokens 💸 Automated Claim System 🔹 Infrastructure PostgreSQL + Redis AWS / Google Cloud Ethereum / Polygon (Smart Contracts)

⚔️ Challenges we ran into

🌐 Real-time disruption detection from noisy data (weather, news, social media) 📍 GPS spoofing & fraud prevention ⚖️ Designing a fair premium system for different income groups ⏱️ Ensuring instant payouts with accuracy 🔗 Integrating AI + blockchain seamlessly

🏆 Accomplishments that we're proud of

✅ Built a fully automated insurance workflow ✅ Designed parametric triggers (rain, AQI, strikes) ✅ Created a Proof-of-Work activity verification system ✅ Developed AI-based demand prediction for income growth ✅ Combined income protection + income optimization in one platform

📚 What we learned

💡 Parametric insurance can be game-changing for gig workers 🤖 AI can predict both risk and opportunity 🔐 Fraud prevention needs multi-layer validation (not just GPS) ⚡ Real-time systems require efficient architecture + caching 🌍 Social impact products need both tech + empathy

🔮 What's next for FIGGY AI

📱 Launch a real-world pilot with delivery partners 🤝 Partner with platforms like Swiggy and Zomato 🌍 Expand to other gig sectors (ride-hailing, logistics) 🧠 Improve AI with live behavioral learning models 🏦 Integrate micro-loans & financial services 📊 Build advanced worker analytics dashboards

Built With

• analytics
• anomaly
• api
• app
• apscheduler
• backend
• backend-ml-+-api-python-3.10+-flask-?-rest-api-server-flask-cors-?-cors-support-pytorch-?-lstm-model
• background
• boosting
• client
• code
• cross-platform
• dart
• database
• detection
• development
• endpoints
• flask
• forest
• fraud
• frontend
• gradient
• handling
• http
• in-memory
• integration
• isolation
• language
• learning
• lists
• local
• logging
• machine
• material
• mobile
• model
• module
• mongodb
• monitoring
• optional
• os
• package
• persistent
• probability
• programming
• python
• requests
• rest
• scikit-learn
• sequence
• static
• system
• test
• testing
• trigger
• ui
• weather
• windows
• xgboost