Delinquency prediction in a cash-driven economy

Context

A lender serving salaried borrowers in a largely cash-based economy.

Challenge

Thin borrower data made early delinquency hard to anticipate.

What we deployed

Machine-learning delinquency model for 0–60 day risk.

Typical architecture

1. STEP 01

   Repayment and payroll signals combined

   →
2. STEP 02

   Model scores accounts weekly

   →
3. STEP 03

   Collections prioritised by risk

Illustrative of how this solution is typically structured and governed; deployment specifics are available under NDA.

Results

• 90% accuracy predicting 0–60 day delinquencyPartner-reported

NoteResults from deployments by GoGenesis and its platform partners. Outcomes vary by portfolio and operating environment; your results are set in the pilot.

Governance approach

• Collections staff decide each contact strategy

Illustrative of how this solution is typically structured and governed; deployment specifics are available under NDA.