Most delivery partners live day-to-day. If it rains heavily, if AQI spikes, or if a sudden bandh hits the city — their income just drops. No backup. No compensation. No system that actually reacts in real time. That’s the gap we’re trying to address with AegiSync.
The Idea
AegiSync is an AI-powered parametric insurance system designed for gig workers.
Instead of traditional claims:
- No forms
- No manual approvals
- No waiting
If a verified disruption happens → the system detects it → payout is triggered automatically.
The core idea is simple:
If the loss is predictable and measurable, the compensation should be automatic.
What We’ve Built So Far
1. Core Product Flow
We structured the system around a clear lifecycle:
- Worker onboarding
- Zone-based risk scoring
- Dynamic premium calculation
- Real-time disruption monitoring
- Automated claim validation and payout
Getting this flow right early was critical. Most of our effort went into making sure the system behaves like an actual insurance engine—not just a demo app.
2. Risk-Based Pricing
Instead of flat pricing, AegiSync calculates weekly premiums based on zone risk:
- High rainfall zones → higher probability of disruption
- High AQI areas → increased health risk
- Platform outage history → additional weighting
This makes the pricing:
- Dynamic
- Explainable
- Fair (at least in theory)
3. Real-Time Trigger System
We integrated external signals to detect disruptions:
- Weather conditions (rainfall thresholds)
- AQI levels
- Local disruption indicators (bandh/curfew signals)
The challenge wasn’t just fetching data—it was deciding:
When is a disruption “severe enough” to trigger a payout?
That threshold logic took multiple iterations.
4. Automation Over Manual Work
The biggest design decision we made:
Remove human dependency from claims.
Once a trigger is validated:
- Policy status is checked
- Eligibility is verified
- Claim is processed automatically
This reduces:
- Delays
- Fraud vectors
- Operational overhead
5. Early Fraud Considerations
Even in early stages, we started thinking about:
- Location spoofing
- False disruption claims
- Pattern inconsistencies across users
We’re not solving everything yet, but the system is being designed with these constraints in mind.
Tech Approach
We focused on keeping the system:
- Modular (clear separation of services)
- API-driven (external data integrations)
- Simple enough to demo, but structured enough to scale
This balance is harder than it sounds. Overengineering early would’ve killed us.
What Didn’t Work
Not everything went smoothly.
- Our initial trigger logic was too sensitive → too many false positives
- We underestimated how messy real-world data can be
- Integration took longer than expected (as always)
But fixing these gave us a much clearer system.
Where We’re Heading Next
In the next phase, we’re focusing on:
- Stronger fraud detection
- Better system validation
- Cleaner UI/UX for demo clarity
- More robust automation logic
Basically:
Less “it works” → more “it works reliably”
Why This Matters
Gig workers operate in uncertain environments, but the systems around them are still rigid and slow.
AegiSync is an attempt to flip that:
- Real-time signals
- Automated decisions
- Faster financial support
It’s a small step, but it’s in a direction that feels necessary.
Built as part of Guidewire DevTrails 2026
This project is being developed during Guidewire DevTrails 2026, where we’re building, testing, and iterating in a high-pressure, real-world simulation.
We’re not trying to build “another app.” We’re trying to answer a simple question:
What would insurance look like if it actually adapted to how people work today?
Top comments (0)