How data-driven fraud prevention can act as catalyst for secure and seamless digital payments ecosystem?

Digital payments are no more a thing of the future, they are already transforming the here and now. The 30.19 percent growth recorded by digital payments for the year ended March 2021, as per the RBI’s Digital Payments Index (DPI), is proof of its growing clout in India’s banking and payments landscape.

Consumer behaviour is increasingly leaning in favour of digital payments, as evidenced by the most-recent survey on festive season spending. Three-fourths (75 percent) of those surveyed are now more likely to use digital payments than they were a year ago. However, this increased familiarity and acceptance brought about by the pandemic has also given rise to a new wave of fraudsters who probe for vulnerabilities exposed through such rapid change. If this growing threat is not addressed by the industry as a whole, it could undo the progress made by digital payments, and the trust consumers have placed in it.

Although Indian consumers have displayed greater faith in digital payments, they are still concerned about payments fraud and are exercising greater caution than they did a year ago. These concerns are echoed in the Consumer Trust and Fraud Survey conducted in June 2021. 71 percent of consumers said they are more concerned about scams and fraud because of COVID-19, rising from 47 percent last year at the onset of the pandemic. Almost half (49 percent) are most concerned about fake UPI payment links that ask for money transfers via text or email.

As the industry knuckles down to tackling fraud, it is equally important to identify good, genuine consumers and to provide them with a seamless experience. In fact, Aite Group estimates that the cost of false positives – the inadvertent cancelling of good transactions – will grow to USD $443 billion this year. This is damaging as it ultimately impacts profitability as well as customer loyalty.

The surest way to combat fraud is to try and make sense of the data that is available. Historically, the industry used a limited set of data points in fraud detection. There are five that sit at the heart of a good identity engine –

1. Name

2. Email

3. Phone

4. IP

5. Address

But over the last five years, by combining first and third-party data, the industry has been able to significantly expand the number of data points available to them as part of the fraud detection process. To detect and discourage bad actors, different parties within the ecosystem can tap into the power of network intelligence, machine learning and positive profiling. The vast number of data points being generated in our 24/7/365 world is creating new opportunities for the automated validation and verification of genuine customers — and helping in the fight against fraud.

Given the complex nature of fraud, it is vital to have a multi-layered approach to fraud management. At the heart of a modern, multi-layered fraud management system is the data that empowers the ecosystem to validate across solutions, data types and action points. In the past, data has shown that such a multi-layered approach yields a 20 percent improvement in acceptance rates, compared to the industry average.

For instance, in a multi-layered approach, real-time identity verification solutions can be seamlessly added to existing fraud management platforms to improve accuracy, without disrupting the customer experience. Additional data sources cross-validate information in an automated, orchestrated, real-time manner, enabling higher confidence in customer verification and scoring.

Today’s fraud scenarios necessitate a dynamic and secure fraud management system. The conventional siloed approach of implementing channel-specific monitoring solutions, without building integrated defense-in-depth at the enterprise level, is ineffective against attacks from multiple sources.

Financial institutions need to ramp up usage of relevant technologies across multiple channels to quickly determine whether consumers’ transactions and accounts are legitimate. The increasing digitization of payments and sophistication of identity or credential theft (and proliferation of synthetic IDs) require a combination of artificial intelligence (AI), biometrics, machine learning (ML) and big data analytics to help the digital payments ecosystem flag such fraudulent transactions.

For example, AI-powered user behaviour analytics (UBA) goes beyond static machine learning (ML) by assessing micro-behaviour patterns such as a swipe on the phone screen or a stroke of the touchpad. It provides additional, unique behavioural data that helps distinguish a real user from a fraudster – whether human or robotic.

This is just one example of how banks and financial institutions can develop a better and more accurate understanding of users’ behaviour, their reactions and consumption of various features across multiple channels. It is an approach that not only blocks fraud attacks but also enhances customer experience, conversion rates and revenues.

While a multi-layered approach to fraud is crucial in warding off malicious elements, real-time fraud prevention can be strengthened only when the larger ecosystem coalesces to share data.

With real-time payments leading to a shorter window for fraud prevention and payment recovery, the RBI’s recent initiative to enable banks to report and access fraud information from a central payments fraud registry is another step in the right direction. Not only does it encourage collaborative learning, but it also facilitates faster response times. These data-driven initiatives will go a long way to securing the digital payments ecosystem and elevating the customer experience to a new level.