Smarter debt collection
Collecting from overdue customers meant charging them through payment APIs, and every attempt costs money. So I built a decision engine, mixing rules and a machine-learning model, that decides what to collect, when, and how much to try. Less blind retrying, lower cost for every dollar collected.
How it works
Tap a step for the details.
Customers who are behind on payments.
Business rules plus a machine-learning model weigh up each customer.
It picks when to try and how much to ask for.
The attempt goes through the payment APIs, by ACH or card.
Every attempt costs money, so fewer blind retries means less waste.