Stripe's new 'caller ID' machine learning platform is tackling online fraud - and making card payments safer
FYI, this story is more than a year old
Fraud has been a lifelong problem for businesses, and with the evolution of online technologies, the problem is only going to get bigger. By 2020, card fraud will total $183.29 billion worldwide.
However, antiquated technologies mean that online businesses are financially liable and are being left behind when it comes to effective fraud management. Legacy systems also tend to use manual rules that can block genuine payments.
Machine learning is a rapidly emerging technology that can help businesses block fraud, and that is exactly what online commerce platform Stripe has done. Its network has previously helped the company vet four out of five credit cards, and now it's the first company to use machine learning-based fraud tool with no setup and works from the first customer transaction.
Today the company launched Stripe Radar, a set of machine learning-powered tools that use data from hundreds of thousands of transactions happening on the Stripe network every second, as well as data intelligence from banks and credit card networks.
The company says machine learning uses artificial intelligence and computers to train themselves on data and its application to new situations without intervention or instructions from human programmers. Radar works by using algorithms to detect patterns and instances of likely fraud. As it learns, its defences become more powerful over time.
John Collison, Stripe president and co-founder, says that Stripe Radar is caller ID for incoming charges, using machine learning.
“Because of this network, Stripe Radar can effectively spot patterns and detect fraud, protecting every Stripe user from the moment their first charge comes in," he says.
The company says that if a business launches tomorrow and receives its first credit card payment, there's more than an 80% chance that Stripe will have seen that card before. In fact, it has seen the average card at least six times before.
At no additional charge, Stripe Radar is fully integrated into Stripe users' accounts, there is no integration or setup or need for manual review teams, so businesses are instantly protected from their first transaction.
Stripe says its technology is also simple and effective, built on top of its own machine learning algorithms. These include prioritised lists of flagged charges, the ability to preview and set custom rules without coding, and additional fraud checks such as verification codes and address verification.
The company says that during the beta period of two months, Stripe Radar blocked more than $40 million in fraud attempts for a non-profit medical funding agency Watsi. Teespring, a custom t-shirt design website, was also able to reduce manual review charges and redeploy its software developers to product and customer experience focuses.
Stripe has previously received $300 million in funding from investors such as Sequoia Capital, Visa, American Express, and PayPal founders Peter Thiel, Max Levchin and Elon Musk.