Product 01 · Varsha · For banks, NBFCs & crop insurers
Know what the monsoon will cost you — before it arrives.
Varsha simulates rainfall scenarios across your crop-loan or crop-insurance portfolio, village by village, and translates them into crop loss, ₹ exposure at risk and stress-test-ready reports.
The questions a credit committee actually asks.
- If rainfall is 30% above normal, which low-lying villages flood and how much soybean is lost?
- If the monsoon ends 13% short — as it did in 2026³ — which branches face the most loan stress?
- Which villages will cross the 33% crop-loss threshold?¹¹
- How much of our KCC book sits in drought-sensitive, unirrigated blocks?
- What do we disclose to the regulator under a severe-drought scenario?
Move the rain. Watch the book.
Illustrative simulation on synthetic data — not a forecast.
From satellite to ₹ in four steps.
Sense.
Crop type, sowing date and health from Sentinel-2 (optical) and Sentinel-1 (radar, which sees through monsoon cloud).
Fuse.
IMD gridded rainfall, soil water capacity, irrigation, elevation, crop calendars, historical yields, your loan or policy book.
Simulate.
Crop-response models turn a rainfall scenario into village-level yield loss, separately for drought and waterlogging.
Decide.
₹ exposure at risk, threshold-crossing villages, branch-level dashboards, regulator-ready PDF and API.
Specification
v0.1 · Target specs — to be validated in pilot
- Users
- Agri-credit risk teams, crop insurers, reinsurers, climate-risk & ESG teams at banks
- Core question
- Rainfall/weather scenario → crop loss → ₹ portfolio exposure
- Scenario inputs
- Rainfall deviation (−50% to +50%), dry-spell length, excess-rain events, temperature anomaly (v2)
- Satellite inputs
- Sentinel-2 MSI (10 m, ~5-day revisit); Sentinel-1 SAR (10 m, all-weather); Landsat 8/9 (30 m); MODIS (daily, 250 m–1 km)
- Ground inputs
- IMD gridded rainfall; soil (water-holding capacity); SRTM elevation (30 m); irrigation; crop calendars; historical yields; mandi prices; client loan/policy book (secure upload)
- Spatial unit
- Field-level analysis aggregated to village / Gram Panchayat / branch / district
- Crops at launch
- Paddy, soybean, cotton, pulses (expand by region)
- Outputs
- Crop-loss % per village; villages above loss thresholds; ₹ exposure at risk; scenario comparison; portfolio heatmap
- Delivery
- Web dashboard · REST API (JSON/GeoJSON) · PDF report · CSV export
- Refresh
- In-season monitoring every ~5 days (optical) / ~6–12 days (radar)
- Scenario run time
- Target: < 60 seconds per district
- Regulatory fit
- Supports climate scenario analysis & disclosure under RBI climate-risk directions²; KCC satellite-monitoring financing⁵
- Validation
- Back-tested against historical yields / crop-cutting data and claims
- Data security
- Client data encrypted at rest and in transit; India-hosted; role-based access
Not another dashboard.
Not just monitoring
Monitoring tools show crop health today. Varsha shows what a scenario does to your book tomorrow.
Not a black-box cat model
Global catastrophe models are district-level and built for reinsurers. Varsha works village by village and plugs into your loan book.
Rupees, not indices
Your credit committee doesn’t read NDVI. Varsha answers in ₹ and in villages.