After its launch in 2011, HiGear, a San Francisco, Calif.–based luxury car sharing service, was seemingly on track for a successful run with 5,000 members and over $1 million in initial funding.
Then suddenly, it hit a wall.
The company’s successful start caught the attention of fraudsters who made off with $400,000 worth of luxury cars with stolen identities and credit cards. The loss obliterated the company, which was forced to shut down.
While the fraud-fighting technology available six years ago didn’t stand a chance of protecting HiGear, the current car-sharing market, which is projected to be worth $16.5 billion by 2024, continues to experience similar fraud attacks.
And car sharing is just one example of emerging industries that seemingly have bull’s-eyes on their backs, said Nuno Sebastiao, cofounder and CEO of Feedzai, a fraud-fighting, machine learning platform.
“We see people in new industries trying to build as much adoption as they can, and then, as soon they create sufficient critical mass, the bad guys come in,” Sebastiao said.
Nonetheless, for businesses old and new, reaching that critical mass has also meant digitizing their operations, making machine learning and artificial intelligence (AI) their new-age fraud fighting tool.
Machine learning gets the problem-solving call
Deep learning algorithms, AI and neural networks may be getting their day in the sun, but they’ve been around for a while. It’s just in the past 15 years that they have gained traction, technology that was once niche is now becoming more mainstream and cost-effective. Additionally, today’s marketplace is increasingly technology-dependent.
However, even today it’s not uncommon to find companies eager to combat fraud but lacking in customer data, Sebastiao explained.
To help such merchants, Feedzai taps into its network of partners, marketplaces and large acquirers that aggregate data for 60,000 to 100,000 merchants.
It also utilizes a Segment-of-One approach, which entails tracking transactions that occur on the company’s network and using that data to verify the authenticity of consumer identities and their transactions.
Cracking the code on digital identity in luxury retail
As machine learning has evolved over the years, so has the level and type of fraud.
And in the past few years, one of the biggest victims of fraud has been the luxury retail industry. Between Q2 2015 and Q2 2016 alone, fraud attacks on luxury retailers went up by 87.2 percent, the highest among any other retail segments, according to the latest PYMNTS Global Fraud Attack Index™.
While chargebacks, botnets and account takeovers are some forms of fraud that merchants of all types face, luxury retailers relying on selective distribution are often a favored target for fraudsters because of the high-ticket value of the goods and, in turn, the potential reward to outweigh risk.
These high-end, often limited-edition products are usually smuggled by fraudsters through a network of reshipping companies, Sebastiao explained.
For luxury, selective distribution of products in different geographies is often a way of creating exclusivity and boosting consumer demand. Take luxury retailer Hermes, for instance. In 2016, the company recorded 9 percent growth in Japan, which the company attributed to its selective distribution network.
For many luxury brands, Sebastiao said, losing some volume of sales to friendly fraud is a smaller concern than maintaining their brand value.
Luxury retailers want to make sure that they reach the right consumer base in each of their markets, as opposed to having someone buying their products and then reselling them in a secondary market. he explained. “That’s a huge problem.”
This is where machine learning comes in, Sebastiao said, adding that an AI can learn the digital identity of customers, their payment methods and geographic locations in order to restrict the buying and reselling of exclusive products.
“We know that it is very unusual that someone buys a [high-end] pair of shoes every half hour,” he said. “It’s very unusual.”
Such nefarious transactions are readily evaluated and flagged by AI, he added.
AI joining what will likely be a long, drawn-out war
With the evolution of machine learning, its use cases are now applicable to a wide array of industries ranging from travel and apparel to gas stations and car sharing.
But when it comes to fighting fraud, the right solution doesn’t just serve the merchant side of the transaction or the issuing side alone, Sebastiao said.
“The problem to address is, how do you cohesively use an AI with the data that you have to effectively mitigate the issue while ensuring that the customer experience is there?” he said.
After all, consumers expect to have seamless interactions with brands, with or without the security guaranteed by an AI.
Ultimately, Sebastiao said, the underlying factor to remember is that while AI has evolved to become a crucial piece of the puzzle, it will continue to evolve and address very specific use cases in the future.
Meanwhile, whether it’s car sharing services like HiGear or luxury retail, fraud is unrelenting and seemingly magnetically drawn to any opportunity to benefit from susceptibility. According to Sebastiao, machine learning is the best tool in existence today to fight it.
To download the April edition of the PYMNTS.com Digital Identity Tracker™, click the button below.