Protect your customers & profit this Prime Day

Keep this year’s four-day shopping frenzy from becoming a fraud festival against your business
Amazon Prime Day 2025 is upon us, and this year’s event has a twist: four full shopping days instead of the two of previous years.
In 2024, online spending in the U.S. surged 11% year over year to $14.2 billion during Amazon’s 48-hour Prime Day event, topping estimates and setting a record. The mega-event has evolved into a retail ecosystem phenomenon, with merchants of all sizes capitalizing on heightened consumer appetite for deals.
While this year’s extended timeline promises record-breaking sales volumes for retailers across the board, it also opens the door to an equally extended fraud bonanza. Fraudsters have long recognized major shopping events as golden opportunities, and the doubled timeframe essentially provides them with a lengthened attack window. For merchants riding the wave of increased consumer spending, the mathematics are simple and sobering: more volume equals more fraud attempts.
AI: The primary tool in fraud’s evolution
This upcoming shopping week presents unprecedented challenges for fraud prevention. Fraudsters are now weaponizing generative AI to create sophisticated attack vectors that were unimaginable just years ago.
Fraudsters are deploying generative AI to craft convincing phishing campaigns, automate the creation of thousands of fake accounts for promotional abuse, and even power chatbots capable of social-engineering customer service representatives into granting fraudulent refunds.
“As the shopping event begins, retailers are gearing up for a significant and sustained surge in sales. However, this surge brings with it a critical challenge: a heightened risk of card-not-present fraud and, later, policy abuse,” explains Eyal Elazar, Senior Director of Market Intelligence at Riskified. “During major sales events, consumers often make impulsive purchases without thorough research, leading to buyer’s remorse. This not only drives a wave of returns but also opens the door to increased instances of return abuse.”
In fact, Riskified data shows policy abuse increases 2x during summer months, especially around sales events like Prime Day.
This technological evolution represents a fundamental shift in the fraud prevention landscape. Traditional rule-based systems and static fraud detection models are increasingly inadequate against AI-generated attacks that can adapt and evolve in real time.
The hidden costs of extended shopping events
This year’s four-day format creates a compounding effect for fraud exposure. Each additional day doesn’t just add linear risk — it multiplies the potential for sophisticated fraud rings to test and refine their approaches against merchants and then scale across entire industries. Early successful attacks can be quickly scaled and repeated throughout the extended shopping window.
Moreover, the financial impact often surfaces weeks after the event concludes. Chargeback disputes, return fraud, and account takeover consequences typically emerge long after the sales celebration has ended, creating a delayed reckoning for retailers who may have initially celebrated seemingly successful sales figures.
Keeping fraudsters out of the deal
The solution lies in fighting AI with AI. As fraudsters become more sophisticated, defensive technologies must outmatch their capabilities. This requires a broad network of data on customer orders and claims across accounts and merchants, advanced machine learning models, and a clustering strategy for identity resolution and risk evaluation.
You need systems that can instantly distinguish between genuine customers, opportunistic fraudsters, sophisticated bots, or serial policy abusers, and can maintain accuracy and effectiveness even across sustained high-volume periods like Prime Day.
Merchants who can accurately identify and protect their best customers while effectively blocking fraudsters will find themselves with both increased revenue and stronger customer relationships long beyond this week’s sales.
“At Riskified, we use AI-driven performance segmentation to deliver consistent and reliable results. Our platform identifies the best segmentation strategy and automatically adjusts risk thresholds to block fraud while minimizing the impact on legitimate customers.”
Eyal Elazar
Senior Director of Market Intelligence, Riskified