Problem/Objective:

Loan fraud is a major cause of concern and losses for the banking sector. Therefore, the RBI has prescribed guidelines on dealing with such frauds and red-flagging of accounts based on early signs of default. Punjab National Bank (PNB) has recently suffered a high-profile scam to the tune of $1.5 bn and faces other willful defaulters as well, clocking 14.11% of its aggregate loans as bad loans. Presently, PNB is using an in-house post sanction credit monitoring tool, Preventive Monitoring System, for all corporate borrowal accounts. The current model covers a number of indicators for evaluating the conduct or health of a borrowal account, and it seeks to measure the performance of the account on the said signals on a continuous basis. However, in an effort to strengthen its borrower evaluation criterion and curtail bad loans, the bank seeks to deploy AI-powered tools to predict the occurrence of unethical practices in the banking system.

Solution/Approach:

PNB has sought to put in place an end-to-end comprehensive solution for Early Warning Signals (EWS) and intelligent transaction monitoring so that timely corrective action can be taken in the event of a potential fraud. The bank aims to leverage advanced capabilities such as artificial intelligence, web crawling and Optical Character Recognition (OCR) to get early warnings on a dynamic basis, based on alerts generated using borrower‘s information collected from various internal and external sources. Using AI and ML, the solution would have intelligent facts extraction capability from sources as diverse as news, social media, government databases, rating agencies, intelligence agencies, SEBI, RBI and other international regulators. It would also be capable of finding the relationships among borrower‘s related parties. A final score would be assigned on the basis of severity of triggers and risk categorisation of the customer. By accessing numerous data points to establish the credibility and financial health of the borrower, the software would throw up alerts on a real-time basis to help the bank officials take data-driven decisions about accepting or rejecting a transaction.

Impact/Implementation:

The bank has planned to outsource the supply, implementation and maintenance of such an EWS system, and has already rolled out a notice inviting applications from interested vendors. By using AI to create a 360-degree profiling of the borrower, PNB aims to aid the prevention, early detection and prompt reporting of frauds to the RBI and investigative agencies. This will be a step closer to realising a zero-tolerance policy of the Indian banking system towards unethical practices.

Image from Pixabay