Payment processing startup Stripe Inc. has launched a new version of its Radar fraud prevention system to improve its ability to take on fraudulent payments.

Dubbed Radar 2.0, the updated system features new tools designed for enterprise organizations to tailor defenses for their individual company using machine learning. The enhanced protection uses hundreds of new signals that distinguish legitimate customers from fraudsters, including purchase patterns that are highly predictive of fraud.

Stripe claimed the new system helps businesses reduce fraud by up to an additional 25 percent while keeping payment acceptance rates high.

A completely new feature, “Radar for Fraud Teams,” also offers improved visibility and granular control for identifying and preventing fraud for fraud professionals within large organizations. Features include faster and more accurate reviews, delivering more in-depth insight into attributes such as typical purchase pathways and mismatches between country of incoming IP address and country where a card was issued; custom rules with real-time feedback, allowing security professionals to customize their fraud scans; custom risk thresholds; block and allow lists; and “rich analytics” on fraud performance, a feature that allows easy access to dispute trends for a user’s business.

“Stripe’s machine learning models are now trained on hundreds of billions of individual data points drawn from the Stripe network,” Stripe Engineering Manager Michael Manapat said in a statement. “We’ve used these data points to update our fraud models, helping businesses on Stripe more accurately identify fraudsters and reduce fraud rates by up to 25 percent while still keeping payment acceptance rates high.”

Stripe remains one of an increasing number of unicorns, startups with a valuation more than $1 billion, that has yet to go public. As of April last year, the company said it had no plans to go public soon, but that hasn’t stopped recent speculation that it may eventually do so.

Image: Stripe