An explainable score built from eight everyday signals, with savings discipline weighed as heavily as mobile money and airtime, designed to widen fair access to credit rather than push instant debt onto a shock. Try the model below, or query the engine directly, every number comes with a full breakdown of what moved it.
Every score returns the factors behind it, in plain language, so every number arrives with a clear reason attached.
Built on eight everyday signals, phone, savings group, utility repayment and farm data among them, the record any household already generates simply by living its daily life. Every household that exists can be scored fairly, and savings behaviour carries the same weight as credit behaviour, keeping the score honest about how much debt a household can actually carry.
Every score requires explicit opt in, and a human reviewer sits above every automated lending decision.
Three steps, start to finish, using only the paperwork a household already has and returning a fully explained decision at the other end.
The household opts in. Mobile money, airtime, group savings, utility repayment and farm records link up, only what they agree to share.
Eight factors combine into a 24-point score in under 400ms, with every factor returned individually alongside the total.
A SACCO, bank, MFI or insurer pulls the score and its breakdown straight into their own underwriting. The model recommends, and a human always makes the final approval, keeping a household's exposure matched precisely to what its own signals can support.
Mobile money, airtime, cooperative savings, utility repayment, production history and market transactions, cross checked against each other before they ever reach the model.
Every response includes all eight factor scores individually alongside the composite, keeping the full calculation visible on both ends.
Retrained quarterly against real repayment outcomes, with dedicated audits for women led and youth led households.
Score on demand, subscribe to score changes, or submit a portfolio for batch underwriting overnight.
Every signal is encrypted in transit and at rest, the same standard formal lenders already trust with their own core systems.
Every connection requires an explicit opt in, and a household can see exactly which signals feed their score at any time.
Signals are scored, kept strictly to that purpose, generating and improving a household's own score alone.
A household can disconnect any signal, pause scoring, or leave the system entirely, with a single request and no penalty.
A simplified, illustrative version of the scoring engine. The live API scores directly from verified data, this just lets you feel how all eight factors add up.
A household this consistent typically qualifies for a standard rate microloan and weather index insurance.
Illustrative only. Factor weighting shown here is simplified for demonstration; the production model scores against verified data directly, and child linked factors never subtract from the total.
"I never had a bank statement to show anyone. Turns out my phone had been keeping the record the whole time. First loan I've ever qualified for."
"We used to reject applications we simply could not assess. Now the score explains itself, so our loan officers spend their time on judgment calls, not guesswork."
"Our chama's savings record finally counts for something outside the group. That is the part nobody had built for us before."
Eight signals today, spread across income, savings, protection and resilience so the score stays fair even when any one signal goes quiet for a season. More are already in the pipeline.
POST /v1/score Authorization: Bearer sk_live_•••••••• Content Type: application/json { "farmer_id": "AK KE-88213", "consent_token": "ctk_9f2...", "data_sources": [ "mobile_money", "airtime", "group_savings", "utility_repayment", "production_history", "market_transactions" ] }
{
"score": 19,
"scale": 24,
"tier": "established",
"factors": {
"production_income": 2,
"mobile_airtime": 3,
"group_savings": 2,
"utility_repayment": 3,
"child_education": 2,
"child_health": 3,
"climate_resilience": 2,
"protective_engagement": 2
},
"child_factors_floor": "never_negative",
"human_review_required": false
}