Know Your Harvest Before It Happens
Pre-harvest yield intelligence for agricultural buyers, processors, and grain marketers operating across Zimbabwe and Southern Africa.
Early access — now onboarding agricultural buyers and processors sourcing across Zimbabwe.
Forward contracts are priced on guesswork. Logistics plans are built on hope.
Every season, your procurement team makes million-dollar decisions based on imperfect information. How much will arrive at the buying point? What's the quality going to look like? Which regions are underperforming?
Traditional sources — Agritex circulars, satellite imagery you have to interpret yourself, agronomist field visits — give you fragments. You're left filling in the gaps with experience and instinct.
A 5% improvement in your pre-harvest forecast translates to millions saved in forward contract pricing, storage allocation, and logistics planning.
What We Provide
District-level yield forecasts, delivered 6 to 8 weeks before harvest.
KurimaSense aggregates field-level satellite data, ground truth from in-app farmer logs, weather, and our proprietary yield models — calibrated for Zimbabwe's five Natural Regions and the specific varieties grown locally — to produce buyer-ready forecasts.
How we work together
Engagements scale with the depth of modelling you need — from standard reporting through to bespoke prediction models — not by seat count. We scope and price each one on a short briefing call.
Core
District Forecast
Crop-level yield forecasts per district, refreshed weekly through the season. Built for early procurement planning.
- District-level yield forecasts
- Weekly updates during the growing season
- Confidence band on every forecast
- CSV / Excel export
- Email support
Standard
Zone Forecasts + Briefings
Forecasts broken down to growing zones within districts, plus monthly written briefings on the key risks and supply signals behind the numbers.
- Everything in Core
- Zone-level forecasts within districts
- Monthly written supply briefings
- API access for system integration
- Anomaly alerts on critical regions
- Priority support
Enterprise
Custom Yield Models
Bespoke calibration and deeper modelling for your specific crops and varieties, backtested against your own records.
- Everything in Standard
- Custom crop & variety calibration
- Custom yield prediction models
- Historical forecast backtesting
- Quarterly strategic reviews
- Dedicated success manager
How it works
Coverage definition
We map your sourcing geography — districts, wards, or growing zones — to our satellite-monitored field network. Custom areas of interest are included in the scope.
Multi-index satellite analysis
We process Sentinel-2 optical imagery (NDVI, EVI, NDRE, NDMI, SAVI) and Sentinel-1 SAR backscatter (VV, VH) — cloud-resilient through Zimbabwe's rainy season. No single-index black boxes.
Ground truth integration
Our farmer app, used by smallholders across your sourcing regions, captures planting dates, variety, fertilizer applications, and final harvest yields. This ground truth continuously calibrates our forecasts to local reality.
Yield model with Natural Region weighting
Predictions are computed using crop- and variety-specific yield potentials, weighted by Zimbabwe's Natural Regions (I through V), corrected for water availability, input quality, and observed crop stress.
Confidence-scored delivery
Every forecast comes with a confidence band — high, medium, low — so your team knows when to trust the number and when to add buffer.
What you receive
- Live dashboard accessible to your procurement and operations teams
- District- and zone-level forecasts with confidence intervals
- Weekly automated updates throughout the growing season
- Monthly written briefings (Standard and above)
- Anomaly alerts when critical regions show stress
- CSV/Excel export for integration with your existing tools
- API access for direct system integration (Standard and above)
- Quarterly strategic reviews with our team (Enterprise)
Why Zimbabwe-specific matters
Global agtech platforms are built on American maize and Brazilian soya datasets. They don't know what Compound D is, can't tell Natural Region II from Natural Region IV, and have effectively zero training data for tobacco — a crop most global yield models ignore entirely.
KurimaSense was built in Harare, calibrated on Zimbabwe ground truth, and tested against actual harvest results from across the country's Natural Regions. Our tobacco-specific model is, to our knowledge, the only one of its kind in commercial deployment.
Read our methodology
We publish how KurimaSense computes yield forecasts, what indices we use, how confidence scoring works, and how we validate against ground truth. A 12-page technical brief, free to download.
Download PDFFrequently asked
Schedule a 30-minute demo
We'll show you a live yield forecast for one of your active sourcing districts.
Book a Demo