RakshaRide - Real-Time Income Protection for Gig Workers

AI-powered parametric insurance for food delivery workers.

Team OriginX

1. SHRESTHA VERDHAN
2. ARPIT SINGH
3. RAMYA PATHAK
4. ARYABRATA KUNDU

Overview

RakshaRide is a Guidewire-aligned, IRDAI-compliant micro-insurance platform designed for gig workers, enabling automatic compensation when external disruptions materially impact their earning ability.

Unlike traditional parametric insurance systems that rely solely on environmental triggers, RakshaRide combines:

• Environmental signals (weather, AQI)

• Platform activity (delivery engagement)

• Behavioral patterns

To detect actual income disruption and trigger payouts.

The system is built as an event-driven insurance architecture, inspired by Guidewire-style insurance architecture, where policy, claims, and billing flows are modeled in a modular, event-driven system, while microservices handle real-time event ingestion and risk computation

Problem

Delivery partners rely on consistent daily activity for income. External disruptions such as:

• Heavy rainfall and flooding

• Extreme heat

• Hazardous air quality

• Fuel shortages and curfews

This directly reduces their earning potential.

Traditional insurance fails because it:

• Does not cover short-term income loss

• Requires manual claims

• Has slow settlement cycles

• Cannot verify real income disruption

Solution

RakshaRide introduces a micro-duration (weekly) income protection system.

How It Works:

• User subscribes to a weekly policy

• System continuously monitors disruption signals

• Multi-factor engine evaluates:

  1. weather conditions

  2. platform activity

  3. user engagement

• Claims are triggered automatically

• Claim confidence scoring determines:

  1. instant payout

  2. manual review

Target Persona Income Bands

Persona Segment	Monthly Earnings (INR)	Typical Daily Earnings (INR)	Risk Context
Entry-level	15,000 - 22,000	700	Limited buffer against missed work days
Mid-tier	22,000 - 32,000	1,000	Moderate stability with periodic disruption risk
High-activity	32,000 - 45,000	1,400	Higher income-at-risk during disruption windows

Hackathon Showcase Checklist

1) Registration Process

• UI flow: frontend/src/pages/Register.jsx
• API: POST /api/delivery-partners/register
• Stores profile, city, platform, earnings band, risk category

2) Insurance Policy Management

• Subscribe plan (payment-disabled hackathon mode): POST /api/insurance-policies/subscribe
• Policy by ID: GET /api/insurance-policies/:policyId
• Partner policy history: GET /api/insurance-policies/partner/:partnerId
• Cancel policy: PATCH /api/insurance-policies/:policyId/cancel

3) Dynamic Premium Calculation

Implemented in backend/services/weeklyPremiumCalculator.js using:

• location risk multiplier
• platform multiplier
• earnings-band context
• seasonality multiplier
• loss-ratio guardrails

Subscription API returns transparent pricingJustification.

Example plans:

Plan	Weekly Premium (INR)	Coverage (INR)
Basic	25	300
Standard	40	500
Premium	60	700

Insurance Calculation Maths

Let:

• B = base weekly premium of selected plan
• L = location risk multiplier
• P = average platform multiplier
• S = seasonal multiplier
• k = loss-ratio loading factor

Then the contextual weekly premium is:

Premium = round(B × L × P × S × (1 + k))

From current config values:

• k = 0.10
• S = 1.15 (monsoon), 1.10 (summer heat), 1.00 (default)

Example (Standard plan in monsoon):

• B = 40, L = 1.2, P = 1.05, S = 1.15, k = 0.10
• Premium = round(40 × 1.2 × 1.05 × 1.15 × 1.10)
• Premium = round(63.756) = 64

Projected loss ratio shown in pricingJustification is:

Let:

• C = maximum weekly coverage of selected plan
• e = expected payout amount as a fraction of C in a qualifying disruption week (not the probability of a disruption)

Then:

Projected Loss Ratio = (C × e) / Premium

Current config source: backend/config/parametricInsuranceConstants.js (PREMIUM_MODEL_ASSUMPTIONS.EXPECTED_PAYOUT_SEVERITY_RATIO = 0.30).

Using the same example with Standard plan coverage:

• C = 500, e = 0.15, Premium = 64
• Projected Loss Ratio = (500 × 0.15) / 64 = 75 / 64 = 1.17 (rounded to 1.2 decimals)

Note: the resulting loss ratio of 1.2 in this worked example is included only to show the calculation path. In the current model, it is classified as above_sustainable_band and surfaced in pricingJustification.lossRatioAssessment.

AI-assisted risk hint during registration:

• UI calls POST /ai/quick-risk-assess
• city mapped to risk score/category

4) Claims Management

• Submit claim: POST /api/insurance-claims/submit
• Claim detail: GET /api/insurance-claims/:claimId
• Partner claims list: GET /api/insurance-claims/partner/:partnerId
• Flagged queue: GET /api/insurance-claims/flagged
• Review decision: PATCH /api/insurance-claims/:claimId/review

5) Automated Disruption Triggers

Supported triggers:

• heavy_rainfall
• extreme_heat
• hazardous_air_quality
• lpg_shortage
• flooding (mock/event-driven)
• area_curfew (mock/event-driven)

Trigger threshold snapshot:

Event	Trigger
Heavy Rain	Rainfall > 50 mm
Extreme Heat	Temperature > 42 C
Hazardous Air Quality	AQI > 300
LPG Shortage	Severity Index > 70

Trigger APIs:

• Threshold preview: POST /api/disruption-events/check-threshold
• Create disruption event: POST /api/disruption-events
• Auto-trigger claims for event: POST /api/disruption-events/:eventId/trigger-claims
• Weather monitor run-now endpoint: POST /api/admin/trigger-weather-check

6) Zero-Touch Claim UX

• automatic claim trigger for eligible active policies
• automatic fraud scoring
• instant auto-approval path for low-risk claims
• manual review fallback for suspicious claims

Admin Mode

Admin panel exists for judge/demo operations:

• route: /admin
• frontend gate via VITE_ADMIN_ACCESS_KEY
• supports weather checks, disruption event creation, event-level auto-claim trigger, and flagged-claim review

Payment Mode

Premium checkout is intentionally disabled by default.

• ENABLE_PREMIUM_PAYMENT_FLOW=false
• users can subscribe directly via /subscribe and access coverage immediately
• Razorpay endpoints remain available for future enablement

Guidewire-Centric System Design

RakshaRide leverages Guidewire as a coordinated insurance execution engine:

PolicyCenter

• Dynamic underwriting using behavioral + environmental risk inputs
• Weekly policy issuance with configurable product definitions
• Rule-based eligibility validation

ClaimCenter

• Event-driven FNOL (First Notice of Loss)
• Automated claim triage via confidence scoring
• SLA-based claim lifecycle enforcement
• Fraud flags routed to manual adjuster queues

BillingCenter

• Subscription-based billing (weekly cycles)
• Payment retries and policy suspension logic
• Integration with Razorpay for reconciliation

Data Flow

1. User registers -> KYC validation
2. Policy created -> PolicyCenter
3. Premium calculated -> rating engine
4. Payment processed -> BillingCenter
5. Event detected -> ClaimCenter FNOL
6. Fraud scoring -> ML service
7. Payout executed -> payment gateway

IRDAI Compliance Enforcement

RakshaRide enforces compliance at the system level, ensuring auditability and regulatory alignment.

Regulation Area	System Enforcement	Implementation
KYC / AML	Mandatory verification before policy issuance	Aadhaar/PAN API + pre-bind validation
Claim Settlement	Automated SLA timers and escalation workflows	ClaimCenter lifecycle rules
Auditability	Traceable logs for every claim decision	Event logs + audit trails
Pricing Governance	Actuarial justification for premiums	Rating engine with loss ratio tracking
Data Privacy	Secure storage and access control	Encryption + role-based access

Tech Stack

• Frontend: React + Vite
• Backend: Node.js + Express + Mongoose
• AI service: Python Flask
• Data: MongoDB
• External data: Weather + AQI APIs

Local Setup

Prerequisites

• Node.js 18+
• Python 3.10+
• MongoDB (local or Atlas)

1) Backend

cd backend
npm install
cp .env.example .env
npm run dev

2) AI service

cd ai
pip install -r requirements.txt
python app.py

3) Frontend

cd frontend
npm install
cp .env.example .env
npm run dev

Deployment Configuration

Local Development (Same-Machine)

For local development with all services on localhost:

• Frontend: VITE_API_BASE_URL=/api, VITE_AI_BASE_URL=/ai (relative paths via proxy)
• Backend: FRONTEND_URL=http://localhost:5173, AI_SERVICE_URL=http://localhost:5001
• AI Service: FRONTEND_URL=http://localhost:5173

Separate Deployment (Different Hosts)

When frontend, backend, and AI are deployed on different hosts (e.g., Vercel, Netlify, Render):

• Frontend: Set VITE_API_BASE_URL to full backend URL (e.g., https://api.example.com), VITE_AI_BASE_URL to full AI service URL
• Backend: FRONTEND_URL=https://frontend-domain.com (will auto-accept Vercel, Netlify, Render domains)
• AI Service: FRONTEND_URL=https://frontend-domain.com (CORS patterns auto-configured for major platforms)

The backend and AI service automatically accept requests from:

• Configured FRONTEND_URL
• Vercel deployments (*.vercel.app)
• Netlify deployments (*.netlify.app)
• Render deployments (*.onrender.com)

Current Deployments

• Frontend: https://raksha-ride.vercel.app/
• AI Service: https://origin-x-shield-parametric-protecti-five.vercel.app/

Environment Variables

Backend (backend/.env)

• PORT
• NODE_ENV
• MONGODB_URI
• JWT_SECRET_KEY
• ENFORCE_AUTH
• AUTH_DEMO_USERNAME
• AUTH_DEMO_PASSWORD_HASH or AUTH_DEMO_PASSWORD
• FRONTEND_URL - single URL or comma-separated list for CORS allowlist
• AI_SERVICE_URL - URL for internal AI service calls
• WEATHER_API_KEY
• POLLUTION_API_KEY
• WEATHER_API_BASE_URL
• POLLUTION_API_BASE_URL
• PAYOUT_MODE
• ENABLE_PREMIUM_PAYMENT_FLOW
• RAZORPAY_KEY_ID
• RAZORPAY_KEY_SECRET
• RAZORPAY_ACCOUNT_NUMBER

Frontend (frontend/.env)

• VITE_API_BASE_URL - /api for local proxy or full backend URL for separate deployment
• VITE_AI_BASE_URL - /ai for local proxy or full AI service URL for separate deployment
• VITE_RAZORPAY_KEY_ID
• VITE_ADMIN_ACCESS_KEY

AI Service (ai/.env)

• FRONTEND_URL - single URL or comma-separated list of frontend origins
• AI_SERVICE_PUBLIC_BASE_URL - public URL for frontend to reach AI service
• PORT - service port (default 5001)

API Quick Reference

Delivery Partner

• POST /api/delivery-partners/register
• GET /api/delivery-partners/:partnerId

Insurance Policy

• POST /api/insurance-policies/subscribe
• GET /api/insurance-policies/partner/:partnerId
• GET /api/insurance-policies/:policyId
• PATCH /api/insurance-policies/:policyId/cancel

Insurance Claims

• POST /api/insurance-claims/submit
• GET /api/insurance-claims/partner/:partnerId
• GET /api/insurance-claims/:claimId
• GET /api/insurance-claims/flagged
• PATCH /api/insurance-claims/:claimId/review

Disruption Events

• POST /api/disruption-events
• GET /api/disruption-events
• POST /api/disruption-events/check-threshold
• POST /api/disruption-events/:eventId/trigger-claims

Admin Utility

• POST /api/admin/trigger-weather-check

Tests

cd backend
npm test

cd frontend
npm run build

Admin Page Credentials

Mail - guidewire@gmail.com
Password - guide123

Demo Assets

• Website: https://raksha-ride.vercel.app/
• Demo video:https://youtu.be/UCE6xzRXyTQ
• Architecture diagram: https://github.com/user-attachments/assets/9107d99c-d7ef-4b40-bc98-940dc99d9e12
• Workflow diagram: https://github.com/user-attachments/assets/97824b2b-dddb-4aa1-9db9-ee80a4e600ad

Pitch Deck

• Pitch Deck Link : https://drive.google.com/file/d/1D0KQAHumvD5_dgiP9fhFcILheP2JelX2/view?usp=sharing