Case Study

advanced & Dynamic Credit Scoring

Inaccurate credit scoring leads to larger exposures for banks. This point will have a tremendous impact for financial institution with the coming crisis.

​

Contact us for more detail

IN BRIEF

KEY INFORMATION

SPEED

A Merton-style model for each factor (country, risk), for single “counter parties”, as many factors as necessary

PERFORMANCE

Opmitization of credit portfolios by maximizing the benefit

+

FLEXIBILITY

By easily comparing the optimized portfolio with the original one

Challenges

  • Not to be unswayed by the racial, gender, and socioeconomic biases that have skewed access to credit in the past
  • Detect early warning of potential failure
  • Give interpretability on the fly

Solution

  • The system relies on alternative data points (social media, online browsing, etc.) and enables the discovery of hidden relationships between complex variables
  • No modification of the code or intervention of human experts is necessary
  • Connection between scoring, segmentation and activation

Gains

  • Higher margin for the bank (extremely important in the coming weeks)
  • More interpretable results to avoid ethical issues
  • Better understanding of customers behaviors for better service design

Testimonial

“Finally, an industrial understanding of AI potential ethical issues!”