Problem & Persona
Why q-commerce delivery income breaks—and what riders need instead of traditional insurance.
Core Problem
Delivery income is fragile because it depends on three things at the same time:
- Rider is online
- Orders exist for the rider's zone
- Roads and air conditions are safe enough to complete deliveries
When any one of these fails (rain, AQI spike, bandh/curfew, platform zone halt), weekly earnings can drop abruptly—often within hours.
Why Traditional Insurance Fails Riders
Most insurance is claim-and-document driven (damage/injury paperwork). That model doesn't match the gig reality:
- losses are short-cycle (hours, not weeks)
- evidence is operational and zone-specific
- riders need compensation before the disruption worsens
Target Persona (q-commerce delivery partner)
Zepto, Blinkit, BigBasket, Instamart, and JioMart delivery partners typically share:
- Earnings: INR 5,000 to INR 8,000 per week (plus incentives; no fixed salary)
- Work pattern: 8–12 hours/day, 6–7 days/week (demand-driven)
- Operating area: dense urban micro-clusters (a few kilometers)
- Platform dependency: order flow and zone assignment come from the app
What Triggers Income Loss
Aegis focuses on disruptions that remove earning capacity by breaking mobility or order dispatch:
| Disruption | What happens to income |
|---|---|
| Heavy rain / flooding | unsafe roads; cancellations and rider self-parking |
| Extreme heat / heatwave advisories | health risk; reduced rideability window |
| High AQI / pollution | air safety constraints; platform guidance to stay off roads |
| Bandh / curfew / govt shutdown | zone closure; fewer/no dispatches |
| Platform-declared zone halt | riders stay online, but orders stop |
Design Implications
Aegis is built around three principles:
- Parametric payout: automatic, trigger-based, no claims process
- Zone-aware decisions: micro-zones using H3 (resolution 8)
- Fraud-resilient guardrails: silent checks so honest riders are not penalized
The problem-to-solution alignment is demonstrated by the end-to-end workflow diagrams in the rest of this documentation.