Aegis-docs

End-to-End Workflow

From onboarding to HALTED-zone payout—fully automated, zone/time aware, fraud-guarded.

1) Worker Onboarding (Zero-Type: Simulated Platform OAuth)

To avoid manual KYC friction in the prototype, Aegis simulates platform OAuth onboarding:

  • app registration
  • simulated identity verification (Aadhaar/PAN + face/OTP)
  • UPI/bank linking
  • platform affiliation confirmation (Zepto/Blinkit/BigBasket/Instamart/JioMart)
  • tier selection (Basic / Standard / Premium)

Outcome: a verified rider profile, coverage choice, and assigned operating H3 zone.

Rider Onboarding Flow (Prototype)

Onboarding is simulated via platform OAuth (mock Zepto/Blinkit/BigBasket/Instamart/JioMart) to avoid manual KYC in the prototype.

2) Profiling & Premium Setup

Weekly workflow:

  • ingest historical activity
  • compute Ew baseline (rolling 4-week avg)
  • enrich zone context for the insured time window
  • compute premium quote for the rider
  • worker accepts/enters the policy cycle

3) Real-Time Disruption Monitoring

Cadence (prototype):

  • every 2 minutes: evaluate zone state for active H3 cells
  • every 10 minutes: aggregate zone intelligence snapshots

On HALTED trigger:

  • persist an audit record
  • emit DisruptionEvent for claim processing

4) Automatic Payout (No Claims) + Fraud Guardrails

When a DisruptionEvent arrives:

  1. eligibility check (policy active + rider/zone/time fit)
  2. fraud gate (silent: does not block honest riders)
  3. decision:
    • clear -> auto-approve and pay
    • suspicious -> hold/manual review
    • ring-connected -> freeze and investigate
Fraud Detection & claim automation — decision and payout workflow

5) Risk Pool Feedback Loop

After payouts:

  • update loss ratio and portfolio exposure
  • recalibrate model confidence bands
  • adjust future premium signals with governance controls
Risk Pool Monitoring — internal feedback loop

Risk pool feedback is computed after payouts: loss ratio tracking and model confidence recalibration for the next premium cycle.

Architecture

Modular, event-centric services for zone trigger detection, eligibility gating, fraud safeguards, and payout automation.

Installation & Setup

Local development setup for Aegis: core backend + ML/risk engines + payout integrations.

On this page

1) Worker Onboarding (Zero-Type: Simulated Platform OAuth)2) Profiling & Premium Setup3) Real-Time Disruption Monitoring4) Automatic Payout (No Claims) + Fraud Guardrails5) Risk Pool Feedback Loop