AI-based advanced analytics is making credit, debit cards smarter with real-time responses

Banking and fintech firms have been using artificial intelligence (AI) for the last few years to improve fraud detection on credit and debit cards, analyze patterns of defaulters, caution users from overspending and even help them determine their spendings. 

Some companies have now also begun using predictive analytics to enhance how credit and debit cards are being used in real-time.

Consider the example of a Philadelphia-based fintech, cred.ai, which launched its 'Unicorn Card' two years back. The card, which was licensed by payments network Visa and issued by the Wilmington Savings Fund Society, FSB, uses a credit optimizer tool. This tool uses an AI algorithm to improve the user’s debt-to-credit ratio, which accounts for up to 30% of a FICO score that evaluates a person's creditworthiness in the US. Apple, too, uses AI algorithms to determine the user’s credit limit on the Apple Card, which was launched in 2019.

Closer home, Gurgaon-based fintech firm OneBanc has developed a card to connect various banking systems. The company, which counts 60 corporates among its partners for providing cards on employees’ salary accounts, claims this card brings all of the employee’s payments into one. OneBanc’s card use the company’s smart platform to include employee benefits like meal balances, etc., on a single card. When the card is swiped, it takes a decision in real-time about what balance to use for making the payment.

OneBanc has already partnered with RuPay and Visa, and said its "AI Card" will be hit the market in the next four to six weeks.

Vibhore Goyal, the founder and chief executive of OneBanc, explained that the problem with banking right now is that different systems like credit cards, loans, etc., are not connected. According to him, when banks issue credit cards on fixed deposits (FDs), for instance, the system is manually programmed to disallow customers from overshooting the amount put in the FD. “Those two systems are actually not talking. It's a manual process. So, someone will be deputed to reach out to the customer if they break the limit. Our (AI- powered automated) systems enable us to do all of this in real-time,” he added.

The card can also connect with a company’s human resource (HR) and finance platforms and make recommendations based on the same. For instance, it can recognize a recurring payment the user makes every month and start suggesting that payment at the prescribed time monthly.

Further, Goyal also said that the AI helps enhance security on the card. The AI Card doesn't have magnetic stripes, instead embedding the card data on a Europay, Mastercard, Visa (EMV) chip. When transactions are made, the AI records location data, and takes into account information like a user's travel information, employment profile etc. to detect possible frauds.

EMV chips were prescribed by the Reserve Bank of India (RBI) back in 2018, and card issuers have been moving to replace magnetic stripes with those. The chips, however, are susceptible to something called shimming, where attackers steal data off the chip in order to duplicate the card and use it for unauthorized purposes. OneBanc's algorithms aim to protect against the same by using the fraud detection AI explained above.

Likewise, Bengaluru-based Scienaptic AI said this January that payments provider Uni Cards would use its AI algorithms to provide a "refined application process" to its users. "Our credit decisioning platform allows Uni Cards to experiment, test-and-learn faster, and over time, build the best decisioning strategies for their customer segments, balancing growth and profitability," said Pankaj Kulshreshtha, chief executive of Scienaptic.

According to Mihir Gandhi, partner, Payments Transformation at PwC, a couple of major private sector banks in India too have already been working on using AI to predict the payment methods that would be most suited to making a payment. “Even if you look at non-financial companies, like say Amazon, which has my payment details saved. When I’m doing a transaction on Amazon, it suggests a preferred payment option,” said Gandhi.

Prashanth Kaddi, Partner, Deloitte India, said the market for data analytics in the banking and financial services sector is in the range of Rs.7000-10,000 crore in India alone. Of this, the use of advanced analytics, which includes AI and machine learning (ML) is expected to be just under Rs.1000 crore.

Going forward, experts say platforms that power smart cards will eventually start understanding users’ transaction patterns, and be able to predict transactions even before they happen. They will also help reduce confusion among customers who have many different cards to choose from today, and carry multiple cards. In future, cards like these could also choose how to pay for a transaction instead of the customer knowing which card would be best for a particular payment.

Kaddi said the industry is “reasonably mature” but the direction today is towards hyper personalization. In future, a credit card app may suggest a specific retail offer when a person walks into a mall, having recognized their location. Doing so will eventually allow banks to become the preferred plastic in the user’s wallet.

Gandhi believes the future of payments is when customers walk into a kirana store and the payment app or platform suggests the mode of payment automatically. According to him, the return on investments on such platforms for banks is that they can provide more convenient payment methods, and get insights into transactions that customers are doing, which can then be subject to more analytics. However, he said that the ability to monetize this data has not yet happened. “Everyone talks about data, analytics and AI/ML (artificial intelligence and machine learning), but monetizing it is the difficult part. Using it for ensuring a smoother process is what the focus is on today,” he said.