Opinions expressed by Entrepreneur contributors are their own.
You are reading Entrepreneur India, an international franchise of Entrepreneur Media.
Alternative credit scoring models, and the lenders that embrace them, are seriously penetrating the parts of the market that were considered largely impenetrable or too difficult to guarantee. Advances in AI/ML and innovations in using data beyond the preset list of mainstream lending practices have made this possible.
Pexels
Some of the visionary players in the market segments where insufficient data is a major barrier to underwriting and thus lending are making extensive use of alternative credit scoring models that use AI/ML on unconventional data to profile and evaluate customers. These models often combine elements of various computer vision algorithms (for image segmentation, object detection), geospatial analysis, and NLP methods for information extraction from textual data.
This approach has proven to be a game changer in “new-to-credit” segments. For some of the early movers in the lending space targeting the lower end of the SME sector where mainstream insurance and credit history files are quite thin, AI/ML-driven alternative credit scoring models are increasingly becoming an integral part of lending processes and will an important differentiator in the future.
Conventional credit scoring methods followed by lending institutions rely on adequate credit history (credit bureau data), formal banking and accounting records, multi-year tax filing information, etc. Alternative credit scoring models, on the other hand, use data other than the types listed above. Fintech firms lending in markets where adequate credit histories, banking and tax return records, etc. are not available, particularly rely on such alternative credit scores for their underwriting.
These alternative credit scoring models use data such as geolocation-based data on various economic, demographic and risk indicators, certain comparable types of indicators derived from satellite image data, other sectoral economic trends at the location level are used quite extensively in alternative credit scoring AI/ML models. Another type of data where AI/ML algorithms (e.g. certain variants of deep learning models) prove to be quite useful is business image data (e.g. goods inventory, retail space, storefront and location street, etc.). Also, modern alternative approaches to AI/ML-driven credit scoring make use of allowed mobile data (transactions, SMS data, informal mobile app accounting data, for example) using certain regular expression-based and/or NLP methods followed by ML-based modeling. An important aspect of the alternative approach to credit scores is that it makes use of the alternative data, along with the limited banking data available or even a small credit history (“thin file”) that may be available in some scenarios.
As we noted, the alternative approach to credit scores not only uses unconventional data, but the data types are of a wide variety (images, texts, along with numerical data). This necessitates specific computer and data extraction techniques and AI/ML algorithms to ingest and use most of this type of alternative data (such as images, SMS scraches, etc.) that would not have been suitable for traditional data. -analysis methods. Carefully developed and rigorously tested ML models using such comprehensive data from multiple sources are capable of highly accurate credit risk prediction. This enables fintech firms to address the critical data gap by replacing conventional credit scoring with AI/Ml-based credit scoring models that use alternative data.
The alternative approach to credit scoring allows for multiple extensions of the scope of lending to a significant portion of the subordinated segments, increasing revenues with appropriate credit risk management and pricing for the lenders, and also addressing the social cause of financial inclusion.
AI/ML solutions that enable such alternative credit risk modeling will also be a critical factor in bringing (almost) all-digital lending products into hitherto untapped segments. The early movers who have adopted AI/ML ahead of the rest will have a major advantage in that space thanks to their significantly evolved AI/ML practices and rich, organized in-house alternative data they have collected, along with a deeper understanding of the markets.