May 18, 2017
By Sean Neary, Subject Matter Expert in Financial Services, Featurespace
There is no doubt that both fraud and customer management are constantly evolving. Fraudsters are becoming increasingly sophisticated in their attack methods, and data breaches continue to rise. At the same time, consumer pressure to provide a seamless purchasing experience is a top priority for many merchants and retailers. It is something I experienced personally during more than a decade managing fraud operations within the banking and payments industry.
The future is looking challenging in our CNP world, not least due to:
- A rise in global CNP attacks: Data breaches continue to rise, compromising customers’ card details and making the e-commerce channel a target for fraudulent spending. During the 2016 holiday season, CNP fraud rose by 31 percent over this short period. Legacy fraud systems that rely on outdated, inefficient methods of customer analysis from consortium data cannot keep up with the changing fraud types. These systems are not able to learn from new types of customer activity events without manual training of the system, and therefore the accuracy of fraud detection declines rapidly.
- Increasing consumer pressure for a smooth ordering and purchase process: With global consumers expecting 24/7 access to online accounts and purchasing, the ultimate goal for merchants and retailers is to reduce friction throughout the purchase process. Extra authentication steps might help reduce fraud, but must be balanced against the risk of genuine customers leaving before completing their transactions.
- Balancing costs of fraud and customer management: Fraud operations teams are under increasing pressure to manage the cost of preventing fraud against the expense of chargebacks and blocking genuine transactions. All while minimizing manual intervention during a customer’s ordering and purchase process, to maximize operational efficiency.
With these challenges ahead, it is time for some good news: machine learning and AI are enabling merchants, retailers and card issuers to stop fraud in real time, while also reducing false positives, to accept more business and reduce friction for genuine customers.
To gain this competitive advantage, it helps to understand the business benefits behind the buzzwords.
Behind the Buzzwords: AI and Machine Learning
Artificial intelligence, big data, machine learning: these buzzwords have been heard more frequently over the past few years, but how can they benefit merchants and card issuers in attempts to detect and prevent real-time fraud?
The buzzwords simply represent scientific techniques that use mathematical algorithms, data, and powerful computing hardware to build models. A model, based on multiple data inputs, provides an output to a user, which can then be used to make an informed decision. Self-learning systems using these techniques go one step further by automatically learning and updating the system based on user feedback.
How are AI and Machine Learning Solving Fraud Prevention Problems?
For merchants and card issuers, machine learning is the key technique giving organizations a competitive edge in fraud prevention and customer management—especially when combined with real-time adaptive behavioral analytics.
Adaptive behavioral analytics works by building a ‘normal’ pattern of behavior, in real-time, for each individual customer and their peer group—and then spotting the exact moment when this behavior changes.
By detecting anomalies across customer activity from every channel, a machine learning, adaptive behavioral analytics approach evolves throughout each customer’s lifecycle and detects new fraud attacks as they occur.
At the same time, this robust understanding of each customer means that genuine customers are easily recognized by the fraud system, dramatically reducing false positives. Genuine customers can transact without being blocked, and fewer authentication steps are needed before transaction—reducing friction during the purchase process and increasing revenue for merchants.
All this takes place within a self-learning fraud prevention system, which automatically updates itself as new behavior types are identified, reducing manual intervention during transactions.
The Practical Benefits of Understanding Real-Time Behavior
Practically, merchants and card issuers are seeing significant benefits from adopting a machine learning, adaptive behavioral analytics approach, including:
- Quick, accurate fraud decisioning and acceptance to protect customers and increase revenue
- Up to 70 percent reduction in volume of genuine transactions declined (false positives) including in CNP fraud target areas—increasing revenues from genuine customers
- Reduced chargebacks and smoother ordering processes, by reducing manual intervention during transactions
- Reduced friction for authentication as behavioral activity and device information can be used to identify genuine and registered customers
The Future of AI and Machine Learning is Now
Machine learning and adaptive behavioral analytics are enabling merchants, retailers and card issuers to understand their customers at a deeper level for advanced fraud management and reduced friction during the purchasing process.
The time is now to embrace the latest advanced machine learning methods for informed, real-time, automated decision making. It is an opportunity to have the highest level of real-time fraud protection in the payments industry while increasing control over customer management – moving towards the seamless purchasing process that genuine customers crave.
Sean Neary has been in the financial services risk management industry for more than ten years. Starting his career in the operational environment at Barclaycard, Sean progressed from frontline analyst to complex risk management and its technology. He brings his Financial Services industry expertise and insight to his role at Featurespace.