Phase: 1 (Ideation & Foundation)
Status: Workflows Implemented, Frontend Pending
Last Updated: March 2026
| Role | Name | Contact |
|---|---|---|
| Product Lead | Yasharth Kesarwani | yasharthkesarwani24@gmail.com |
| Backend Engineer | Yasharth Kesarwani | yasharthkesarwani24@gmail.com |
| Frontend Developer | Vishwajeet Chauhan | javamihisweet100@gmail.com |
| AI/ML Engineer | Sarthak Dharmik | sarthakdharmik10227xix@gmail.com |
GigShield, an AI-based parametric insurance service, helps gig workers, particularly food delivery drivers, protect their income against loss that occurs due to external events such as extreme weather conditions, air pollution, or regulatory restrictions.
GigShield is different than traditional forms of insurance because they use parametric triggers to automatically detect when a disruption has happened and to pay out instantly without the need for a manual claim process.
Gig workers, such as delivery people, deal with:
- 🌧️ Income loss due to weather disruptions
- ⏱️ Work that has strict time requirements and then the weather also affects that work
- 📉 No safety net financially that they can rely on
- 🚫 The process of getting a traditional insurance payout that is very complicated and that takes a long time to get paid
GigShield can provide:
- the whole claim process is automated, no paperwork, etc.
- analyzes the risk of the gig worker and finds out who is a fraud
- payout for lost gig income within two hours
- Micro insurance paid weekly that aligns with gig income
Why did we target this group – food delivery partner?:
- In general, food delivery partners are exposed to a lot of different environmental risks.
- When there are peak earning hours for food delivery (when the conditions are most likely to cause disruptions to earning), food delivery partners are out working.
- Therefore food delivery requires mobility outdoors – real-time.
- Person: Rajesh 28 Mumbai
- Trigger: Rainfall exceeding 50 millimeters sustained for three hours during the 6-10pm hours
- Manner in which or amount of earnings they will lose: Rajesh will lose between ₹800 to ₹1200.
- How he will respond to losing money (claims): 70% automated payout will occur automatically at the time of claim.
- Person: Priya 32 Delhi
- Trigger – Air Quality Index: 400+
- Work lost: Lost four to six hours of work due to poor air quality.
- Claims to receive: Will receive hour amount claimed for lost time to work caused by poor air quality.
- Person: Amit 25 Bangalore
- Trigger: Curfew or travel restrictions in his area.
- How many earnings will he lose: ₹1500 daily
- How he will receive the claim: Cash payout based on area in which events occur.
- Register worker (platform, location, earnings)
- AI-driven risk assessment
- Calculate weekly premium
- Activate policy
- ⏱️ Hourly tracking of:
- Weather
- AQI
- Location accessibility
- Parametric trigger evaluation
- Automatic trigger detection
- Fraud validation (AI + rules)
- Claim approval
- 💸 Instant payout
This diagram illustrates the complete lifecycle from onboarding to automated claim payout.
Inputs:
- Location
- Platform type
- Weekly earnings
- Work hours
Processing:
- Historical weather analysis
- Risk scoring
Output:
- Weekly premium (₹50–150)
The following diagram shows the high-level architecture of GigShield, including frontend, workflows, AI layer, and integrations.
Tracked Parameters:
| Parameter | Threshold |
|---|---|
| Rainfall | >30 mm/hour |
| Temperature | >42°C / <5°C |
| AQI | >300 |
| Wind Speed | >40 km/h |
IF (Weather > Threshold)
AND (Worker in affected zone ±500m)
AND (Active during working hours)
AND (Policy active)
THEN Trigger Claim
Weekly Premium = (Weekly Earnings × Risk Score × Coverage %) / 52
- Earnings: ₹7000
- Risk Score: 0.15
- Coverage: 70%
➡️ Premium ≈ ₹101/week
- 🌍 Geographic Risk (40%)
- 🌦️ Seasonal Risk (30%)
- 🏙️ Zone Risk (20%)
- 👤 Individual Risk (10%)
| Risk Level | Premium | Coverage |
|---|---|---|
| Low | ₹50–75 | 70% |
| Medium | ₹76–125 | 70% |
| High | ₹126–150 | 70% |
To maintain viability in long run, the following must hold true:
- Maximum payout per event= ₹500-700.
- Maximum weekly payout= ₹1500.
- Target loss ratio= 60-70%
| Trigger Type | Threshold | Payout |
|---|---|---|
| Heavy Rain | >30 mm/hr | ₹150/hr |
| Extreme Heat | >42°C | ₹150/hr |
| Pollution | AQI >300 | ₹150/hr |
| High Winds | >40 km/h | ₹150/hr |
| Zone Closure | Boolean | ₹1200/day |
- Evaluating by AI:
- Historical weather data
- Location specific risks
- Provide custom premium estimates
Detects:
- 📍 Discrepancies in location
- 🔁 Duplicate submissions
- ⏰ Incorrect timeframes
- 📊 Deviating patterns of behaviours
Reasoning:
- If the fraud risk score (≥80) = report as fraud.
- If the fraud risk score (50 to 80) = manual verification.
- If the fraud risk score (≤50) = approved.
- Provide:
- Insight
- Identify anomalies
- All key determinations will continue to be rule-based for dependability.
- Predict future risks.
- Recommend:
- Early start times.
- Change zones (customer works).
- Temporarily suspension of policies.
Reasons for Web-first?
- Installation not required.
- Faster delivery.
- Lower cost.
- Suitable for lower-end devices.
Principles:
- Access off-line.
- Employment of push notifications.
- Mobile friendly.
- React.js + TypeScript
- Tailwind CSS + shadcn/ui
- Workflow Automation Engine
- RESTful API Architecture
- OpenAI GPT (via LangChain)
- Document/Relational Database (MVP)
- PostgreSQL (production)
- Weather: OpenWeatherMap
- AQI: IQAir / CPCB
- Razorpay
- Twilio / Firebase
- Vercel
- GitHub Actions
- Sentry
- Event-driven workflows
- Stateless API layer
- Redis caching for external APIs
- Modular AI agents
- Workflow documentation.
- AI installation/integration.
- Parametric triggers.
- Database design.
- React front-end.
- User dashboard.
- API integration.
- Testing.
- Optimization.
- Documentation.
- Demonstrate.
├── README.md
├── workflows/
│ ├── onboarding-workflow.yaml
│ ├── monitoring-workflow.yaml
│ └── claims-workflow.yaml
├── backend/
├── frontend/
├── docs/
└── tests/
- Users = 1000
- Average Premium = ₹100/week = ₹1,00,000
- Anticipated Claims = ₹60,000
- Profit = ₹40,000
- ⚡ Instant payouts
- 🤖 AI-assisted system
- 📅 Weekly micro-payments
- 🔒 Fraud-resistant
- 📊 Transparent triggers
- Workflow will be fully automated at 100%
- Utilization of dynamic pricing
- Five or more different trigger conditions will be available
- Claims will be processed in less than 2 hours on average
- Claim accuracy will be greater than 95%
- Onboarding of new policyholders will take no longer than 5 minutes to complete
| Risk | Mitigation |
|---|---|
| API limits | Cache client responses and use multiple APIs |
| False-based triggers | Implement validation layers to avoid false triggers |
| Fraud | Artificial intelligence (AI) and rule checks will help to reduce fraud |
| Low adoption | Provide user incentives to drive adoption |
| Data errors | Use multiple sources for verification prior to loading data into the platform. |
- Will be compliant with IRDAI guidelines
- May consider use of a regulatory sandbox or partnership with insurers in lieu of receiving licensing
- Integrate into multi-platforms (Zomato/Swiggy)
- Consider expansion into:
- Ride-hailing drivers
- Construction workers
- Use of Predictive Analytics
- Development of mobile applications
- Pilot program with 50 initial users
- Validate Financials
- Consult with appropriate Regulatory Agencies
- Develop partnerships with potential platform partners
Project Name: GigShield
Type of Project: AI, FinTech and InsurTech
Stage: MVP Development