Yabx: Mapping the Transformation of the Digital Lending Landscape Powered by AI

Artificial Intelligence (AI) has permeated the landscape of most sectors. Most businesses and sectors have been on their toes due to constant changes in today’s markets, and the financial industry is no exception.

The lending sector is constantly changing due to digitisation and evolving client expectations. The processes that form the framework of digital lending, such as customer lead management, customer onboarding and underwriting, customer relationship management, and collections, are increasingly embracing the pure-play approach to digital operability. Much credit needs to go to AI for the reorientation of users towards the proffered digital dimension from the traditional procedure.

George Thekkekara, chief risk officer at YabxGeorge Thekkekara, chief risk officer at Yabx
George Thekkekara, chief risk officer at Yabx

George Thekkekara is the chief risk officer at Yabx, a fintech aiming to simplify financial access to over two billion unbanked people in the emerging markets of Africa, Asia, and Latin America using the mobile phone device.

With an ambition to democratise such access, Yabx aspires to provide financial services to the under-served. Yabx uses technology and analytics to reduce the cost of delivering financial services, thereby bringing banking to the unbanked. 

Speaking to The Fintech Times, Thekkekara analysed the impact of AI on the lending industry:

Mapping the scope of AI’s serviceability within the lending ecosystem

The applicability of AI and ML: The purchase experiences of customers for credit facilities have evolved substantially over time. Customers who are online the entire day leave a large behavioural imprint on the digital assets of lenders or their affiliates. AI and its subset, machine learning (ML), can assist lenders in understanding the customer behaviour better while predicting possible business outcomes related to risk underwriting, fraud patterns, product uptake and usage, and the user’s loyalty to the franchise. These predictions can provide insight into the creditworthiness and the customer’s lifetime value. A host of algorithms can be simulated to identify these behaviours. It includes both tried and tested methods like logistic regression and novel approaches using tree-based ML algorithms (random forest, gradient boosted trees), and graph-based network analysis, which are handy in identifying collusion.

Sales and prospect targeting: By leveraging traditional and alternate data sections like clickstream data, search data, and other records that include wallets, telecom, e-commerce, and utilities, AI can predict the objective of a customer’s purchase. Based on the results, users may be classified as ‘must reach’, ‘require more effort’, or ‘not interested’. There may be other categories depending on the needs of lending institutions. Based on his categorisation, lenders can reach out to more prospects in a focused manner, particularly at a very early point in the sales funnel relationship.

Leveraging NLP for a complete customer view: Pertinent questioning can gauge customer requirements. Through exploration, a perspective is framed, and discernment becomes very easy to make a decision. The use of the resourceful branch of AI known as natural language processing, or NLP, aids in recognising various data points, including the assessment of language and tone to evaluate credit risk. In business loans, NLP may be used to assess attitude and entrepreneurial acumen. In the same way, it may flag data that doesn’t make sense and send it to be reviewed further. It can also be used to include nuanced factors such as the sentiment of the borrower throughout the loan treatment plan.

All-embracing credit scoring: AI uses alternative credit scoring mechanisms to help create more inclusion within the lending landscape. To calculate a credit risk score for a consumer, AI can leverage 300–400 data points from their behaviour, financial history, and income tax history, among other activities. In the past, lenders used financial and other data to make loan decisions. Through the use of AI, lenders have a better competitive advantage as it can help them grow their loan books by analysing structured as well as unstructured data. Lenders can use alternate credit scores to contact existing clients to market a pre-approved loan product or contact new leads. This ensures a better and more inclusive credit ladder where smaller ticket and short-tenure loans are leveraged alongside larger ticket instalments.

The future of AI and digital lending

Trust isn’t built overnight, but the gradual experimentation has ensured that AI and its associated technologies have accrued the faith of many, especially in the digital lending space. They not only process data quickly, but they also do it with higher precision and without prejudice. This makes them a valuable asset in the financial services business, particularly in lending, where millions of applications are received but cannot be processed owing to human bias, resource constraints, and mistakes. AI. is surely capacitating the scope and driving the expansion of efficacy within digital lending through its subset technologies like ML, NLP, and natural language understanding.