AI: Fintel Refining Credit Scores

The Marseille-based start-up Fintel was founded in 2019 and launched a tool for assessing applicants’ creditworthiness in just a few seconds using APIs.

This AI-powered full-digital solution performs analyses in a matter of seconds, searching through up to 5 years of history to define the customers' banking profiles. The idea is to collect data from the customer’s bank account and from tax services, analyses it (revenue, credit lines, savings, etc.).
Besides, it can also detect typical evidence of risk-inducing behaviours, including intervention fees or recurring overdrafts.

Further refining risk assessments using AI technologies to be consistent in analysing and calculating credit scores. The point is to assess the applicants’ trustworthiness (attrition score) and their appetite for savings and investment products.
Automating scoring processes. Despite the advent of digitalisation in the banking sector, the assessment of applicants’ creditworthiness often remains a manual process. Fintel, for their part, allows for significant time savings with a simplified customer process and faster integration.
Taking advantage of PSD2-driven opportunities to automatically collect customer data and speed up the scoring step.

Fintel enjoyed support from the Erasmus University Rotterdam and had access to data from nearly 300,000 bank accounts to improve their algorithm’s efficiency. This algorithm achieves over 99% of correct detections when it comes to identifying profiles likely to fail. This approach allows for finer risk evaluation, bypassing common beliefs regarding a priori risky profiles.
This tool has already been adopted by two international banks: Ikano Bank in Sweden (IKEA’s banking arm) with issues nearly 1.5 million credit lines each year, and Shinhan in South Korea.
Fintel charges a fee for each request. They hope they can achieve a turnover of €1 million by 2020.

Category : Alertes quotidiennes