It is the reality for all businesses today that customers’ demand for speed and convenience are only some of the factors that are driving the rise in digital transformation programmes. At the same time, fraudsters are rapidly evolving their methods of attack – particularly in the digital space – meaning businesses are in a perpetual race to protect customers and maintain a seamless customer experience. Real-time machine learning is setting the stage for creating a positive balance between customer protection and experience for today’s risk and fraud professionals operating within an evolving global landscape of regulatory changes, faster payment cycles, and Open Banking.
Why this is the right time for machine learning
Machine learning models are not new – data scientists have been using machine learning models for over 60 years. However, it is only now that computer processing power is enabling businesses to use these models at speed and scale to solve practical, real-world problems, such as tackling fraud and risk.
Modern computing architecture and advances in machine learning techniques have made it possible to look at large volumes of multiple data sources and build much richer profiles with the reliability that is required to make real-time decisions during the transaction lifecycle. Until recently, processing multiple data sources at this speed and scale was not possible.
The solution is not to reduce fraud checks, but to complete more checks quickly, while maintaining and increasing accuracy. This is where machine learning is giving businesses the edge.
Machine learning models ensure performance does not degrade over time
Featurespace’s proprietary machine learning invention, Adaptive Behavioral Analytics, works by continuously adapting to consumer and fraud trends by monitoring genuine customer behaviour. Unlike traditional rules-based decision making, the power of machine learning models is that they do not degrade over time, as they can adapt to the changing data it analyses. This enables businesses to make more accurate fraud and risk decisions at speed and scale, with minimal manual intervention to update the fraud management system.
Featurespace continues to produce new inventions with the launch of Automated Deep Behavioral Networks for the card and payments industry. This deep learning innovation provides a much stronger layer of defence to protect consumers from scams, account takeover, card and payments fraud, which cost an estimated USD 42 billion in 2020.
The challenge and the discovery
Deep learning technology has various applications, such as in natural language processing or NLP, for the prediction of the next word in a sentence however, its use in preventing fraud in card and payments fraud detection had not been optimised to protect companies and consumers, until now. With the invention of Automated Deep Behavioral Networks by Featurespace, that capability gap has now been closed.
Research published in the latest Nilson Report clearly illustrates the critical need for advanced card payments protection like that provided by deep learning. Nilson found that card-based payment systems worldwide generated gross fraud losses of USD 28.65 billion in 2019, an increase of 2.9% from USD 27.85 billion in 2018, with gross fraud losses projected to reach USD 38.50 billion by 2027. In a shocking report from banking industry body UK Finance, a record USD1 billion was stolen in the first six months of 2021, up 30% from the same period in 2020, and up more than 60% from 2017 when figures were first compiled.
It is for these reasons that Featurespace developed Automated Deep Behavioral Networks. The technological ecosystem automates feature discovery and introduces memory cells with native understanding of the significance of time in transaction flows, improving upon the market-leading performance of the company’s Adaptive Behavioral Analytics. Ultimately, Featurespace’s belief is that detecting fraud before the victim’s money leaves the account is the best line of defence against scams, account takeover, card and payment fraud attacks.
So, how can these machine learning principles apply to solving the practical problem of spotting and blocking fraud in real time?
Online payments are increasing to the levels where they are surpassing in-store payments. In the United States alone, ecommerce sales are expected to surpass USD 740 billion by 2023 according to a report published by Statista Research. For online businesses where the card is not physically present during the payment, it gets even tougher to distinguish a genuine transaction from a fraudulent one in real time. At the same time, merchants and financial institutions are wary of introducing additional security checks which can increase customer drop-out during the transaction process or increase the number of genuine transactions declined.
For the following groups, the benefits of adopting a machine learning fraud prevention solution include:
Consumers:
enabling genuine transactions with reduced verification;
automatically identifying scams, account takeover, card and payment fraud attacks before the victim’s money leaves the account.
Data scientists:
automatically discovering features in transaction events;
pushing machine learning logic through the entire modelling stack;
leveraging the irregularity of human actions to identify anomalistic behaviour.
Card and payments industry:
improving risk score certainty across all transactions – fraud detection during the transaction is increased and genuine behaviour is more accurately identified to facilitate the acceptance of more transactions;
providing performance uplift for all payment types, including card and ACH/BACS, wire, P2P, and faster payments;
improving the detection of high-value, low-volume fraud (and detection of low-value, high-volume fraud);
reducing step-up authentication;
providing strict model governance documentation, with explainable logic, fair decision making, and reason codes;
delivering stable, real-time scoring with high throughput and low latency response times for business-critical enterprises, even under surge conditions.
Deploying a fraud prevention solution that is powered by the most advanced machine learning techniques gives organisations access to systems that make real-time decisions based on changes in customer interactions. And as a result, the customer, payments systems, and financial institutions are protected from rapidly changing fraud attacks that continue to grow in sophistication and speed.
Get your copy of the 2021 Forrester Enterprise Fraud Management Report recognising Featurespace as a Strong Performer.
This editorial is part of The Fraud Prevention in Ecommerce Report 2021/2022, the ultimate source of knowledge that delves into the evolutionary trail of the payments fraud ecosystem, revealing the most effective security methods for businesses to win the battle against bad actors.
About Steve Goddard
About Featurespace
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now