Philip Plambeck, Senior Vice President, International, Computop
For retailers, fraud is a serious issue they are actively trying to battle. And it’s only growing. Research earlier this year found that 60% of e-commerce merchants experienced a rise in overall fraud levels.
Why? The growth of e-commerce has been exploited by criminals who have combined it with the fundamental digital abilities required to conduct fraud, even exchanging tips with other fraudsters on social media platforms. So, while retailers used to be reluctant to talk openly about fraud, everyone now recognizes the severity of the problem that needs to be addressed, especially when it comes to payments.
The research also noted that credit and debit cards account for most fraud losses due to the growing frequency and scope of data breaches compromising consumer data (including card information) and their widespread use. However, newer payment methods like digital wallets, payment apps, and Buy Now Pay Later (BNPL) account for 37% of fraud losses, so they, too, need to be considered when addressing fraud.
How is this fraud being committed? There are several ways. One is through the use of AI, specifically generative AI. However, AI can also play a critical role in combating payment fraud.
Helping retailers prevent payment fraud with AI
Technologies that leverage machine learning to provide AI fraud scores can generate an automated risk indicator for retailers. The score is an algorithm that learns on its own by analyzing individual payment transactions and determining how different data pieces relate to one another. Using a comprehensive approach, an AI fraud score can identify even the most intricate fraud patterns that may escape human detection or more basic automated systems. It can adjust and bolster its accuracy over time by continuously learning from fresh data, which results in steadily declining fraud rates.
As an extra benefit to retailers, an AI fraud score also helps to minimize false positives – instances where legitimate transactions are incorrectly flagged as fraudulent – and they can frequently be tailored by a retailer to match the demographics of their customers and their own risk tolerance. Because of its adaptability, the solution offers a more customized approach suited to the particular requirements of each retailer’s operations rather than a one-size-fits-all strategy. One retailer may configure the tool to be more strict when handling high-risk transactions, while another may choose to make it more lenient.
Going beyond AI for fraud prevention
Retailers can use the AI fraud score with various other risk management solutions. These consist of static pay gate risk evaluations that have worked well or are utilized to support advanced functionality that restricts the number of transactions made within given time periods. Additionally, it’s critical that risk management solutions work with 3D Secure and check the solvency of customers in different countries for domestic and international sales.
Consumers’ and retailers’ role
With the advancements the payments industry is making in helping to stop fraud, bad actors are turning to consumers directly more and more in attempts to steal their identity for personal gain through phishing attacks and the like. Consumers need to be vigilant in protecting themselves.
Retailers must also play their role in fighting payment fraud. They need to be aware of detecting breaches in their payments process and determine if a customer is actually who they claim to be rather than a fraudster.
Next steps
Regarding AI, for machine learning to learn, the industry needs to start sharing information about fraudulent transactions. Very often, fraud is not apparent at the moment of transaction processing but instead shows afterward. In most cases, these failed payments don’t appear in the transaction records as there’s typically no connection back to payment processors’ fraud tools or scoring providers.
Fraud includes stolen credit card numbers, known suspicious addresses, baseless chargebacks, and delivery fraud. The industry should establish ways to regularly inform upstream providers in the processing chain about misbehavior and damages so their tools can also learn about these cases.
In closing
The issue of payment fraud can’t be resolved overnight. Ironically, technology is the answer to a problem that is fundamentally technological. By combining machine learning and AI with customizable, adaptive functionality, it is possible to detect fraud and improve security, reducing risk for retailers while bolstering customer satisfaction at the same time. With an AI fraud score and other fraud prevention tools and data sharing, the payments industry can significantly minimize the impact of retail fraud.
About the Author
Philip Plambeck is Senior Vice President, International, for Computop. In this role, he is responsible for all of Computop’s non-EU business. Previously, Plambeck was Senior Vice President of Sales for Computop in the UK. Before that, he was Sales Manager at Elavon Inc. Europe.