Beyond binary checks
In the past, catching fraud was mostly about looking at where things happened and what methods were used, like stolen credit card details or forged cheques. But with the growing demand for fast payments, we need to update our methods. That’s where machine learning systems, like Paytm Labs’ Pi, come into play. Pi doesn’t rely on old-school checks; instead, it uses artificial intelligence to identify new and evolving types of fraud in real time.
Personalised fraud defence
Personalised checks in embedded systems are key to real-time fraud detection. These checks rely on special computer programmes (ML algorithms) to closely examine how people behave online. This groundbreaking technology quickly looks at things like when you last used your banking app, how long you use it, who you send money to and many other things about how you act online. This detailed way of looking at fraud helps banks make personalised calculations to see how risky a transaction is based on your unique behavior.
Legacy systems vs. innovative solutions
Thus far, banks and financial institutions have been hesitant to fully adopt state-of-the-art software solutions like Pi. Part of the problem is that many core and processing systems in use throughout the industry cannot support real-time transactions. This makes it challenging to integrate new technology that identifies possible real-time fraud. Even though legacy systems will continue to exist, the need to process real-time payments is pushing for a change to more innovative solutions that can keep up with the speed of modern transactions.
The pseudo real-time dilemma
Services like Zelle, frequently dubbed “pseudo real-time,” underscore the urgency for financial institutions to overhaul their infrastructure to incorporate elements of real-time fraud detection. Senator Elizabeth Warren estimates that “fraud tied to the Zelle network topped $255m last year.” However, it’s worth noting that Zelle’s claim to real-time processing conveniently leaves out its reliance on legacy processes, albeit on the back end. While transactions appear to take place immediately, settlement of those transactions occurs overnight. The gap between processing and settlement would suggest that there would be some time when existing fraud detection could take place. However, current legacy fraud detection capabilities weren’t built with peer-to-peer payments like Zelle in mind.
Ultimately, the difference between claimed real-time capabilities and actual processing accentuates the need for robust, fully realised solutions like Pi.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below formBy GlobalData
In the never-ending arms race with fraudsters, faster transactions beget faster fraud. Reliance on tools built for payments processed in batches will not be appropriate in the future. Until now, financial institutions in the US have not been required to up their game, as Zelle and other peer-to-peer payments are not entirely included in the definition of Regulation E liability. As noted by CNN, “under the federal rule known as Regulation E, banks are technically only liable to cover fraudulent activity when it involves ‘unauthorised’ transactions. Say when someone steals your credit card and makes purchases without your permission. But if someone persuades you to send them $500 through a phishing scam, banks consider that ‘authorised’ and won’t reimburse those funds.”
As banks move to use the appropriate processing in real-time, such as FedNow or RTP, the fraud problem will grow, resulting in regulatory pressure to shift liability. Such a move would guarantee that banks will move to real-time fraud detection. Banks that tackle that sooner rather than later will be in a better position for compliance.
A significant shift in the financial landscape occurs at the intersection of real-time payments, fast fraud and machine learning. What used to be just ideas on paper, including personalised fraud detection and behavioural analysis, are now confirmed realities. With the advent of advanced software and hardware engineering, these concepts are now actionable and within reach, paving the way for a more secure and efficient financial ecosystem.
As this transformative period in finance continues, we’ll have to decide whether to stick with the legacy systems we know or move towards new and advanced ways of spotting fraud in real time. This choice will shape the future of finance.
Alex Jimenez is Managing Principal, Financial Services Consulting at EPAM Systems, Inc.