Greed, Threat, Help — What HDFC Bank Is Up Against In Combating Fraud

The promise of a job amid a pandemic. The threat of penal action unless a condition is fulfilled. The offer to help a consumer through a cumbersome process.

This is some of what banks are up against as they battle rising instances of small value frauds hitting digital transactions.

“Greed, threat and help“ are the most common emotions that leave consumers open to frauds, said Manish Agrawal, head - credit intelligence and control group at India's largest private lender HDFC Bank Ltd in an interview with BQ Prime.

Agrawal believes the pandemic just made it worse. Apart from the fact that a lot more transactions were happening online, scamsters too doubled down on efforts to defraud during this time.

"We have suddenly seen a drastic increase in these transactions," Agrawal said. Lenders, meanwhile, have not spent enough time educating customers on appropriate digital behaviour. "We have to start doing that," he said.

According to Agrawal, social engineering is the most common type of fraud being used by scammers. This is where the scammers make the user believe that they are undertaking a genuine transaction over the phone or email, to gain access to their money.

"These fraudsters may contact customers on their own or the customer may go in search for a contact number and come across a fraudulent entry on the search engine. Both the scenarios are creating a lot of unrest," Agrawal said.

A key reason for this explosion in social engineering frauds, is the rapid rise in adoption of digital payment technologies, which has greatly outpaced customer education. According to data released by the Ministry of Electronics and Information Technology, 7,422 crore digital transactions were conducted in FY22 as compared to 3,314 crore in FY19.

While it is spending time and effort on customer education, HDFC Bank is also investing in building specialised architecture which can protect customers at two levels: authentication and authorisation.

Through previously recorded transactions, the bank is mapping customer payment patterns and ensuring that its systems are able to catch suspicious transactions before they go through.

The increased attention to frauds has meant that the bank's Credit Intelligence And Control (CIC) function has seen its headcount double in last four years and covers close to 100 locations.

However, the bank's systems are still learning.

"As a unit we track over a billion transactions every month at HDFC Bank. The biggest challenge is not to catch the fraud, but to reduce false positives, since that inconveniences customers," Agrawal said.

Reserve Bank of India's latest annual report supports Agrawal's assessment that the number of such frauds is rapidly rising.

Card and internet related frauds accounted for nearly 40% of the total number of bank frauds recorded in FY22. This is only second to advances related frauds which contributed to around 42% of total frauds recorded.

The number of frauds have also been on a rise over the years. In the year ended March 2022, the number of card and internet related frauds stood at 3,596, up 41% year-on-year. Total number of frauds recorded in FY22 stood at 9,103.

The amount in these frauds, though low, has also been on a rise. In the last financial year, the total amount involved in card and internet frauds stood at Rs 155 crore, up 30% from a year before.

According to HDFC Bank's own internal assessments, about 75-85% of the customers falling prey to such frauds are under the age of 45. Metropolitan cities account for about 60% of all digital transaction frauds by number, while rural areas contributed less than 7%.

Around half of the frauds on the unified payments interface, which fraudsters use regularly, involved amounts less than Rs 10,000, while only 1-2% of the frauds involve amounts worth Rs 1 lakh or above, according to the bank.

"Changing customer behaviour and digitising the authorisation process is the only way to adequately tackle retail digital frauds," said Aseem Parashar, partner, PwC India.

Earlier, when the systems flagged off a suspicious transaction, the banks would manually check the authenticity in the back-end. "But the velocity of digital transactions has gone up so much, they have to rely on artificial intelligence and machine learning to tackle frauds, without impacting the speed with which a transaction is approved. Otherwise the customer experience gets impacted," Parashar added.