Making Banks Smarter: How the AI Revolution is Compelling Banks to Change
(Image Credits: Moneylion.com)
The most pressing issue plaguing the Indian Banking system is the gradual rise of
Non-Performing Assets or NPAs. The COVID-19 Crisis has made things worse, and as Dr. Ranjan points out and most would concur, we are going to witness an astronomical rise in NPAs within the next six to nine months.
Even prior to the COVID-19 crisis the Indian Economy was in fix, dropping GDP numbers was just the tip of the iceberg. Small and Medium enterprises were suffering because of a loss of consumer confidence and consequently, the banks were not eager to lend. Globally long term rates were falling signaling a steep recession. The US economy was propped up by cutting interest rates and Europe had entered the uncharted territory of negative yield. During this, not so rosy time the world was hit by the worst economic disaster since the Second World War.
Can AI help lead us out of this socio-economic chaos?
In this piece, we explore the impact that AI can have on the banking system.
(Image Credits: Verdict.co.uk)
Banking has evolved to be one of the most central institutions, in modern society. It functions as a resource allocator in any economy, and at the same time, manages risk. Both these aspects require the banks to function in a way that may be regarded as anything, but human. Even a slight bias can tip the scales towards pandemonium, as was evident in the Financial Crisis of 2008 and the many crises before it. Consequently, the domain of banking has been enterally dependent on quantitative objective measures. AI is nothing but math in action and can help the modern banking system improve and thrive.
Banks have been shifting resources from those who have too much to those in need. Credit is the tool that enables this transformation and efficient credit management has helped build empires - we wouldn't have seen the brutal dominance of the great British Empire if it were not for the institutions in the City Of London; the glamorous American Dream would not have captured the imagination of the world if it were not for the banks on the Wall Street. As with anything old, fragility has crept upon the modern banking system. The way forward is disruptive, with the removal of existing biases and improvement in accessibility. These concepts are as good in practice as they seem in theory. In the digital age, we have witnessed the disruption of major industries, entertainment, travel, and education naming a few, banking is the next frontier.
An AI disruption in Banking is in the works. Credit assessment has been the central area of quantitative study for the past decade. However, this has mainly relied on the credit history of the borrower through their financial records. What about the untapped market for first-time borrowers? Such individuals usually face the challenge of showing their creditworthiness through their income and assets. This changes with AI and big data.
Today every transaction is being recorded, every financial move is creating a digital trail. Alternate data to predict consumer behavior is becoming a new tool for the banks to figure out the creditworthiness of an individual. Deep Learning and ML models now seek to use data pertaining to bill payments, micro-transaction, and even social media interactions to predict the chances of delinquencies. Such strategies are not only beneficial to first-time borrowers but also to the overall economy. Extending consumer credit can help break the vicious cycle of loss of consumer confidence and improve the status quo. The banks benefit immensely as they tap a huge consumer base which had been largely neglected up until now due to regulatory issues. If every potential buyer has the resources to buy, the economic situation is bound to turn around.
Automation on a large scale would help make banks processes more consumer-centric and streamline operations. A major source of risk to banks, more often than not, comes from the institutions themselves. Eventually, after all the required paperwork and objective measures to make things fair, it is a person who processes and judges loan applications. This bias reflects on the financial health of firms and is the main reason for the evergreening of distressed loans. The opportunity cost of extending credit to a substandard prospect is, a good prospect getting rejected. De-biasing the system and making it fairer is only possible through automation. Neural nets can be artificially designed to remove the bias in the training data. They can help make the system more fair and profitable.
Banks have begun embracing this change, still, the progress is slow and requires more effort on the part of the regulators. Regulation is important in the banking system however the discourse and debates are highly opaque. Disruption is inevitable, when is the question.
Head over to neuralastic.com to see how we are making our mark in the AI transformation.