Increased usage of semi-automated fraud prevention solutions can lead to undesirable outcomes in e-commerce, but fully automated AI/ML-driven solutions can help

In Japan, a sneaker boutique was using a rules-based, manual fraud prevention solution for its online store, but fraudsters were able to exploit the system’s weaknesses.

The firm, Foot Locker atmos Japan G.K. (atmos), also deployed 3D Secure to provide an extra layer of authentication at checkout, but this introduced unnecessary friction for good customers and had apparently led to a higher cart abandonment rate.

According to the firm’s general manager of the e-commerce business department: “With the previous rules-based system, it was challenging to discern between legitimate and fraudulent transactions, and the need for manual reviews placed a heavy burden on operations.”

Subsequently, adopting a fully automated fraud prevention solution helped the firm to eliminate the need for manual reviews and all their associated problems. Since deploying the solution, the firm has significantly reduced chargebacks and maintained an approval rate of over 98%. The new system is said to maximize the “genuine customer experience and lifetime value” by automatically making precise decisions about customer trustworthiness at critical interactions. This has reduced friction for good customers and has led to a reduction in cart abandonment rate.

The platform applies machine learning to deliver decisions that are 100% automated, with average response times under 400 milliseconds, according to Yosuke Noda, Country Manager, Forter Japan, which provided the automated fraud prevention solution: “We are very pleased to be working with a leading brand like atmos to deliver precise decisions about customer trustworthiness at critical interactions. This decisioning helps atmos improve customer experiences while driving revenue. The platform is powered by the largest network of retailers, meaning a fraudster known to one of its customers is known to all.”

Since the platform has no dependency on manual reviewers, it can scale seamlessly as merchants grow, Noda said.