This document discusses a project to provide micro loans to millions of unbanked mobile customers based on their mobile billing history. A self-learning credit scoring model will be built to determine creditworthiness from customers' mobile transaction and payment data in order to offer loans of up to $250 via their phones. The project aims to introduce financial services to underserved customers and is forecast to be the fastest growing service line in 2017. It has a 3-month timeline per country and $100,000 total budget plus localization costs.