Credit intelligence for emerging markets
Credit scoring for the 80% traditional lenders can’t see.
AfriScore turns mobile money transactions, airtime patterns, and device data into explainable credit decisions — for the millions of creditworthy people and small businesses that formal credit bureaus can’t score.
Approval Optimiser
Simulate your uplift
Set your lender profile to tailor the estimation to your scenario.
For individuals
See what your own transaction history says.
Run a live assessment using mobile money and airtime behavior — the same signals AfriScore uses to build your credit profile — and see exactly what drives the result.
Try the live demo →Trust & data governance
Built for regulated environments from day one.
ROI Calculator
Measure your uplift
Estimates based on industry benchmarks for alternative-data credit scoring. Actual results may vary based on portfolio composition.
Where things stand
Built in public. Here’s the honest status.
Scoring engine built
Rust backend, Python ML pipeline, Random Forest ensemble with isotonic calibration and SHAP explainability — working end to end.
Tested on synthetic data
Trained and validated on a synthetic dataset built from empirical mobile money distributions across East, South, and Central Africa.
Seeking first pilot partner
In discussions with a Cameroonian credit union serving informal-economy borrowers to calibrate against real repayment outcomes.
Real-data calibration
Once a pilot is live: replace synthetic training data with real, anonymized outcomes and publish updated performance metrics.
Building or lending in an underbanked market?
We’re accepting a limited number of pilot institutions. No cost, no commitment — just a real evaluation against your own portfolio.