When the season fails, cover pays.
- We count
- Farms covered on their season record; losses paid where the field confirmed them, and losses the index missed.
- From
- The insurer’s decision, logged against the record.
Impact
you can check.
Parcel by parcel, season by season: what changes for farmers and their land, counted from the records a lender or an insurer has to trust, by a party that sells the farmer nothing and decides nothing about them.
AgroConform WorkspaceImpact / Season 2025-26 / Parcel 12
Parcel 12, durum wheat, 5.2 haSeason 2025-26
Counted this season
Crop coverOn the recordInsurer decision, 3 Dec 2025
Input creditOn the recordLender decision, 20 Nov 2025
Water consumedAgainst its own pastSatellite and weather, three seasons
Soil practiceSeen from spaceResidue kept, recorded 14 Jul 2026
Counted by a party that sells no inputs, buys no harvest and decides no credit or cover.
Who carries the riskThe farms with the least protection
The farms with the least protection
carry the most climate risk.
A dry spring, a late frost or a wave of pests can take a season. Family farms absorb it alone. Their careful work with water, soil and inputs leaves no trace an institution can read, so their risk is priced as a penalty.
- No cover when the season fails.
- Credit priced on collateral, never on practice.
- A price that ignores how the crop was grown.
What we work forFour changes,
Four changes,
each one counted.
Each is a change in a farmer’s season or on the land. Counting starts with the first pilot season: for each change we name what we count and the record it comes from.
Credit that follows how they farm.
- We count
- Farms assessed on the record they chose to share; credit granted on it, its timing against the crop calendar, and repayment.
- From
- The lender’s decision, logged against the record.
Water and soil work that counts.
- We count
- Water the crop consumed per hectare against the parcel’s own past; parcels where soil practice was recorded and seen from space.
- From
- Field records and satellite readings, read by ARI.
Proof that opens a market.
- We count
- Lots accepted with a verified season behind them, and their price against comparable lots.
- From
- The lot’s passport.
The pathwayHow a record
How a record
becomes a change.
- 01
A season, verified.
The farmer records the work once, alone or with an adviser. Satellite and weather data check it parcel by parcel.
- 02
A decision, made on it.
A lender, an insurer or a buyer decides on the record the farmer chose to share, under their own rules. The decision stays theirs.
- 03
Room to act, and advice to act well.
Cover and credit take part of the risk off the farm. Advice on the parcel helps decide where the next investment goes. Which way practice moves is the link we test hardest.
- 04
A change, measured.
The next season, the same parcel shows what changed: water consumed, practice kept, a loss absorbed. We compare it with its own past and with parcels like it.
Parcel 12, sowing
Recorded18 Nov, with a dated photo
Seen from spaceSentinel-2, 24 NovMatches
Weather12 mm of rain that weekMatches
AgroConform WorkspaceParcel 12 / Decisions
Crop coverWritten by the insurer, 3 DecLogged
Input creditApproved by the lender, 20 NovLogged
Shared by the farmer for these two decisions, until 30 June. Each institution decided under its own rules.
Advice for Parcel 12
Irrigate on Tuesday, about 28 mm: the soil reading is low and no rain is forecast this week.
FromAdviser 03
Cover and creditIn place for this seasonOn the record
Being tested: which way practice movesAgroConform WorkspaceParcel 12 / Next season
Water consumed per hectareBelow its own pastMeasured
Residue keptSeen from spaceMeasured
Compared withIts own past, parcels like it
How we countCounted on the parcel.
Counted on the parcel.
Checked against its own past.
Only what is recorded and cross-checked.
Every count traces to a dated field record matched with satellite and weather data. No survey totals, no averages applied to everyone.
Against the parcel’s own past, and parcels like it.
A change is a difference from the same parcel’s earlier seasons and from comparable parcels not yet enrolled, never from a national average.
Decisions, then outcomes.
Lenders lend, insurers cover, buyers pay. We count the decisions made on our evidence apart from what changed on the farm, and claim neither as ours.
Water as the crop consumes it.
Saving water at the pipe can mean more land irrigated and more water used. We call it a saving only when the water the crop consumes goes down.
Carbon as a co-benefit.
In dry farmland soil gains carbon slowly and is costly to measure. We record soil practice for what it does for the soil and the farm, and never lead with carbon.
AgroConform WorkspaceParcel 12 / Water
Water consumed per hectare
2023-242024-252025-26
Parcel 12Parcels like it
| Change | What we count | From which record | Compared with | First reported |
|---|---|---|---|---|
| Protected | Farms covered; losses paid where confirmed; losses the index missed | Insurer decision against the season record | The same farms before enrolment; farms not yet offered cover | After the first season |
| Financed | Farms assessed on the record they shared; credit granted, its timing, repayment | Lender decision against the season record | Applicants with the same score, with and without a verified record | Terms after the first season; repayment after the second |
| Water | Water consumed per hectare | Satellite evapotranspiration with field records | The parcel’s own past and comparable unchanged parcels | A change needs two seasons |
| Soil practice | Parcels with recorded and visible soil practice; inputs per hectare | Book of Crop entries and satellite readings, read by ARI | Before enrolment; parcels like it | After the first season |
| Markets | Lots accepted; price against comparable lots | The lot’s passport | Same season, crop and grade, with and without a passport | After the first season |
| Advice | Alerts acted on; losses on the parcel | Alert log and field records | The same alert sent at different times | Within the first season |
| Fairness | Accuracy of the score for each group | Score calibration | Each group against the others | With each version of the score |
| Control | Consents given, refused and withdrawn | The farmer’s access log | Season to season | Each season |
The evidenceWhat is known,
What is known,
and what is still open.
We build on what field studies have shown and test what they have not. Each link of the pathway, and the strength of the evidence behind it.
Cover and investment
Insured farmers invest a little more in their fields, and they trust cover more when payouts track what happened on their land.
Sources Karlan et al. 2014, Quarterly Journal of Economics; Castaing and Gazeaud 2025, Journal of Development Economics
Credit and the farm
Credit helps some farms far more than others; timed to the crop calendar, it can raise output. When a farmer chooses to share the record, a lender sees how the farm works; we report who is served.
Sources Meager 2019, AEJ: Applied Economics; Beaman et al. 2023, Econometrica; Fink, Jack and Masiye 2020, American Economic Review
Records and access
Data can widen access to credit or narrow it. Whether a verified season record widens it for small farms is what our pilots are built to measure.
Sources Björkegren and Grissen 2020, World Bank Economic Review; de Janvry, McIntosh and Sadoulet 2010, Journal of Development Economics
Security and practice
Less risk leads farmers to invest more, mostly in inputs. For a soil and water practice, training mattered more than money in the most direct test. So cover and credit come with advice on the parcel, and we measure which way practice moves.
Sources Aker and Jack 2025, Review of Economics and Statistics; Emerick et al. 2016, American Economic Review
Water
Efficient irrigation often leaves the water a basin loses to crops unchanged, or raises it, because more land gets irrigated. That is why we count consumption.
Sources Grafton et al. 2018, Science; Pfeiffer and Lin 2014, Journal of Environmental Economics and Management
Soil carbon
In dry cropland, soil carbon builds slowly, varies with depth and is costly to verify. Practice is recorded for what it does for the soil and the farm.
Sources Moinet et al. 2023, Global Change Biology; Smith et al. 2020, Global Change Biology
Proof and markets
A verified record helps a farmer meet a buyer’s requirements. On its own, it does not guarantee a better price.
Sources Meemken 2020, Global Food Security; Oya, Schaefer and Skalidou 2018, World Development
Advice and alerts
Advice matched to the field helps a little on average. We test each alert against what happens on the parcel and keep the ones that help.
Sources Fabregas, Kremer and Schilbach 2019, Science; Cole and Fernando 2021, Economic Journal
FairnessFair by design.
Fair by design.
Fair by measurement.
ConformScore never reads a farmer’s gender, farm size, crop or the way a record was captured. Those attributes serve one purpose: testing that the score is equally accurate for every group. The test is Equal Performance Difference, and it is part of how the score is calibrated.
Recorded alone or with help, read as accurately.
A farmer who records with an adviser or through the cooperative must be scored as accurately as one who records alone.
AgroConform WorkspaceConformScore / Version check
Accuracy gap by groupMeasured at each version
Gender
Women
Men
Farm size
Small
Medium
Larger
Crop system
Rainfed
Irrigated
Trees
How recorded
Alone
With an adviser
Through the institution
Limit
Never read by the scoreRead only to test accuracy
RisksWhat could go wrong,
What could go wrong,
and what we do about it.
| Risk | What we do |
|---|---|
| Left out: no smartphone, no signal, another language. | The app works offline in four languages, and an adviser or a cooperative agent can record with the farmer. |
| A record turned against the farmer. | It stays the farmer’s. Every look needs consent for a stated purpose and is logged. How consent works |
| A measurer with something to sell. | We sell no inputs, buy no harvest and decide no credit or cover. The roles we keep apart |
| Cover that misses the loss. | Field evidence checks the index before a claim is closed. Index and field |
| Water saved on paper only. | We count consumption, not delivery. |
| A score that serves some better than others. | Equal Performance Difference, tested at each version of the score. |
ReportingMethod first.
Method first.
Results as seasons close.
- 01
Now, on this page
What we count, from which record, compared with what.
- 02
Before the first pilot season closes
The comparison design for each change, and what would count as no effect.
- 03
After each harvest
The count for each change, with its definition and period.
- 04
With each version of the score
The accuracy for each group.
- 05
Each year
An impact report with method, limits and what did not work.
Programmes, cooperatives and funders can start a pilot season with us. A date we miss stays on this page, marked as missed. If anything here looks wrong to you, tell us and we will answer. Start a pilot that measures it