Payment fraud has been a problem for a long time. But in the past, successful fraudsters had to be patient and willing to navigate manual and time-consuming techniques. 

The time and effort that payment fraud required likely added enough friction to limit fraud. Unfortunately, this is no longer the case. The emergence of global e-commerce, fully digital banks and new business models, along with tools such as artificial intelligence, has made digital payment fraud enticing to criminals around the globe. 

Proof that digital payment fraud is nearly cost- and hassle-free can be seen in data about its volume and frequency. According to a report by the Centre for Economics and Business Research, fraud cost the global retail industry a stunning $429 billion last year. Adyen research also found that consumers were scammed out of 234% more than the previous year and that they lost an average of over $800 each time they were victimized. 

While many consumers will be made whole by their card issuer, those who make purchases on some apps won’t be. And all shoppers arguably pay higher prices because of the money merchants pay to respond to or prevent fraud. This is all bad news for both consumers who enjoy the ease and flexibility of digital payments and the businesses that work hard to create positive and frictionless customer experiences.

Fraud is always a problem companies want to tackle. But in the current business environment, it has become an even greater priority. 

“In the last 18 months, many businesses have gone from growth at all costs to trying to optimize for profitability,” said Karan Katyal, vice president of commercial strategy for digital at Adyen, a company that provides an end-to-end payment platform to businesses in industries ranging from retail to airlines to hospitality to the growing number of digital-only companies. “Businesses are paying even more attention to fraud to cut down on some of that cost.”

Many types of fraud and the challenge of AI

Defending against fraud is inherently challenging because companies are always trying to keep pace with new and ever more sophisticated fraud techniques — techniques that both evolve rapidly and vary based on geography and a company’s business model. For example, refund fraud is a common technique among fraudsters. As the name indicates, this flavor of fraud is when someone seeks a refund for a product or service they never received. In 2020, the National Retail Federation reported that American retailers lost over $25 billion to fraudulent returns.

Identity theft is another type that is both difficult to defend against and especially worrisome for companies that handle large volumes of digital payments. Identity theft is when criminals intentionally create accounts — often using fake identities and stolen credit card numbers — with the express purpose of committing fraud. Friendly fraud is yet another increasingly common crime in which someone orders and uses a product or service but claims they didn’t. 

The potential list of fraud types goes on and on and will inevitably continue to expand, particularly now that fraudsters can leverage artificial intelligence, including large language models. 

“Large language models open up massive opportunities for fraudulent users to design new techniques,” said Carlo Bruno, vice president of product at Adyen.

How a platform approach and machine learning defend against fraud

Individual merchants are outmatched when it comes to defending against payment fraud. Understandably, the only fraud techniques they are exposed to are the ones they’ve already experienced, which makes it impossible to recognize or anticipate new attacks. 

Which is why it is so powerful for merchants to partner with a global payments platform provider such as Adyen. “If you’re a retailer, you only see activity at your particular store,” Bruno said. “We have the luxury of seeing patterns evolve everywhere in the world, so we can help customers anticipate trends we have seen in other geographies and industries.”

A global reach is also invaluable in developing machine learning models that pinpoint potential fraud and automatically and instantly defend against it. Why? Because the models are trained on huge volumes of data. 

“We are seeing billions of data points, whereas any individual merchant is seeing just a fraction of that,” Katyal said. “Our models are trained on all of that data and then can recognize fraud and automatically prevent a purchase from going through.”

Leveraging automation at scale for flexible and precise fraud prevention

Machine learning-powered fraud prevention is flexible in important ways that address the limitations of traditional fraud defense. Fraud prevention has always been rules-based. For example, a rule may be that transactions initiated in certain countries at specific times of day will automatically be blocked. 

The problem, however, is that it’s impossible to update rules at a pace that will keep up with evolving fraud techniques and prevent good shoppers from being caught up in fraud prevention because they aren’t applied precisely.

“This is the problem machine learning solves for because it basically creates on-the-go business rules. No human can keep up with the millions of data points about fraud trends and the cyclical patterns of your business and e-commerce,” Bruno said. “It’s too complex, and business rules get outdated quickly.”

As sophisticated as a platform approach empowered by machine learning is at finding and deflecting potential fraudsters, it’s also very good at locating the good customers all businesses want to serve. This is important. The promise of digital payments is to reduce friction and improve the experience of customers. Leveraging machine learning simultaneously helps merchants make things very hard for fraudsters and exceptionally easy for loyal customers. 

“We are just as focused on creating that frictionless flow for good customers as we are on targeting the bad guys,” Bruno said. “That’s all about precision, and machine learning provides that.” 

For more information on how Adyen delivers payments, data and financial products all on a single platform, please visit Adyen.com.