Open Methodology · ATI v1

Score only what we
can actually observe.

A Pakistani smallholder has no credit file. Most scoring models solve that by guessing. The Agri Credit Score doesn't — it scores what satellites and registries can verify on day one, leaves the rest at zero weight until it's genuinely observed, and tells the lender which is which.

We publish the approach openly. The scoreboard is free; the score is the product.

The problem with agri credit models

Ask most agri-scoring tools what a score is based on, and a large share of the weight rests on things that don't exist for a first-time borrower — transaction history, repayment record, platform behaviour. For a farmer who has never borrowed formally, those fields are empty, so the model quietly imputes them as "average". A bank's model-risk team spots that immediately, and rightly discounts the whole score.

Our earlier internal version had exactly this flaw: 42% of its weight sat on behaviour we cannot observe before our marketplace is live. ATI v1 inverts the architecture rather than papering over it.

The unlock: a new farmer is not a thin file

Sentinel satellite archives are retroactive. The moment we capture a plot boundary, we can compute roughly three years of NDVI and radar history for that land — immediately, without waiting.

A brand-new borrower has no track record. Their land does. Three seasons of crop performance, drought recovery and cropping intensity are already sitting in the archive on the day they register. That is the difference between an unscoreable farmer and a scoreable one.

Five gates before any score exists

Gates are pass/fail, not points. Fail one and the file is referred for human review — it does not receive a weaker score. This mirrors EU checks-by-monitoring and Brazil's CPR central-registry anti-double-pledge check.

G1
Identity
CNIC verified biometrically — NADRA e-KYC.
G2
Land linkage
The named person is genuinely linked to the parcel, as owner or verified tenant — LRMIS / provincial records / attested tenancy.
G3
Boundary integrity
Mapped boundary versus claimed area — a shortfall above 30% fails.
G4
Plot liveness
At least six effective satellite observations in the past twelve months. Land that isn't actually being farmed cannot be scored as if it were.
G5
No double pledge
The plot isn't already pledged elsewhere in the registry — the failure mode that has historically broken agri lending.

What we can see on day one

The day-one score — ATI-0 — is roughly 45% satellite evidence and 55% verified registry and declared data. Every input is obtainable for a Pakistani farmer this quarter: NADRA, land records, a boundary walk, and the Sentinel archive. Nothing in it is aspirational.

Satellite evidence 45% Registry & declared 55%
Current crop health + 3-year plot resilienceLand, tenure, crop economics, regional risk, references

Behavioural signals — marketplace conduct and repayment history — sit at zero weight. Not estimated. Not imputed. Zero, until they are actually observed.

The score grows up — and says so

Every score carries a maturity tier, so a lender always knows which regime produced it. As real observations accumulate, dormant signals absorb weight from the proxies. At origination this follows the MYbank pattern (satellite and registry can underwrite a farmer with no financial history — 1.69M borrowers in China demonstrate it); at maturity it follows the FICO pattern (once repayment is observed, it dominates).

ATI-0
Verified + Satellite
Day one. No behavioural history — and we say so.
≤ 30% of satellite-estimated crop value · monitoring mandatory
ATI-1
+ Platform activity
One season of marketplace conduct observed. Behaviour enters at 10%.
≤ 50% of estimated crop value · monitored
ATI-2
+ Repayment observed
At least one completed financing cycle. Repayment enters at 15%.
≤ 70% · standard monitoring alerts
ATI-3
Outcome-fitted
Weights re-fitted on observed defaults across 500–1,000+ cycles. Repayment dominates at 25%.
Bank's own risk appetite · calibrated PD

Missing data is never imputed as neutral. A signal with no usable data is excluded and the remaining weights renormalized, with confidence caps — a thin satellite record cannot reach the top band.

Honesty, priced — not disclaimed

This is the part that makes a day-one score responsibly lendable. The tier doesn't just describe the score; it caps what the score may be used for. Rather than burying "we have no repayment data yet" in a footnote, we convert it into an exposure limit in the API response.

Banks don't need to trust the score. They need to know exactly how much to trust it.
At ATI-0 we suggest financing capped at 30% of satellite-estimated crop value, in kind rather than cash — inputs delivered, synced to harvest. That's the One Acre Fund and Apollo pattern, and it is natively Salam- and Murabaha-shaped.

What's built, and what isn't

We'd rather you knew this before you asked. ATI is a designed model with a live satellite core — it is not a finished production system, and it has never been trained on real defaults.

LIVE
Satellite anomaly engine. Real Copernicus/Sentinel data, crop-and-season-aware expectations, piecewise ramps, unscored-plot handling. Unit-tested. This carries 45% of ATI-0 and is the part competitors can't fake.
LIVE
Validation harness. Scorecard, calibration and fairness tooling exist, so the outcome re-fit at ATI-3 is a planned procedure rather than a hope.
BUILDING
Three-year resilience metric. The Sentinel history pull exists; the stability score derived from it is in development.
BUILDING
Verification gates G1–G5 as an onboarding workflow. Boundary and liveness are already computable from the existing pipeline; NADRA and LRMIS integrations are not yet in place.
PLANNED
Production scoring API. Gates, tier and exposure guidance in the response contract — specified, not yet shipped.
PLANNED
ATI-3 outcome fitting. Requires 500–1,000 completed financing cycles. Until then the weights are expert judgment and probability of default is flagged calibrated: false. No exceptions, no rounding up.
🛰️

See it score a real plot

The interactive sandbox — move a farmer through ATI-0 to ATI-3 and watch the weights shift into repayment as the relationship matures — is available to partner institutions.

Partner access → Request the methodology paper ↗
Full methodology & governance paper shared under NDA · Shariah Board advisory available