Using AI to assess credit risk; developing a model

The article presents the results of the analysis which was conducted to develop a credit assessment model for SMEs and individuals based on AI and using the incremental information coming from PSD2. The AI models developed address areas of business that have been wary of algorithms and AI. The focus was on developing models with high accuracy as well as understandability. The article shows that that ML can improve credit selection decisions compared to linear models without reducing transparency. PSD2 data were also used to outline individuals’ consumption lifestyle. This research both improves credit assessment and offers a way to give consumers advice because it allows banks to suggest changes to customers’ expense patterns that could improve their creditworthiness. Finally, the development of an affordability model based on a ML algorithm to find patterns in financial or lifestyle Key Performance Indicators (KPI) to define instalments that make a loan sustainable is recommended.