Text Size:


First Data’s New Fraud-Detection Tool Features Real-Time Algorithm Updating
June 1, 2017

By Kevin Woodward

First Data Corp. is hoping that its scale—it says it processes more than 2,800 transactions per second coming from more than 6 million merchants—will give its new Fraud Detect service an edge in thwarting fraudulent transactions.

Image Credit: First Data Corp.

Guru: “The algorithm can ingest known bad activity in real time.”

Announced Thursday, Fraud Detect uses artificial intelligence and machine learning, fraud scoring, cybersecurity intelligence, and information gleaned from the Dark Web, where much of the criminal sale of stolen data takes place, to identify potentially fraudulent transactions. Atlanta-based First Data says the service works with transactions in-store, at the pump, online, on mobile, and in-app.

It can provide a score on each transaction prior to the merchant’s actually running the transaction or fulfilling the order, Ajay Guru, First Data head of merchant fraud solutions, tells Digital Transactions News. The score, which ranges from 1 to 1,000, can be used as an indicator of the probability of fraudulent transactions, he says. The higher the score, the greater likelihood of fraud.

Fraud Detect uses an algorithm that incorporates the multiple data sources available to First Data, he says. “Because First Data has visibility across multiple merchant types and categories, we actively see the same card or type of trends in an industry, and how those fraud trends go from one industry to another,” Guru says. “The algorithm can ingest known bad activity in real time.”

In a test with multiple merchants, Fraud Detect produced lower false-positive rates, which is the identification of legitimate transactions as fraudulent ones, and an 80% improvement in identifying fraudulent transactions before they happen, he says.

The need for fraud-prevention tools is paramount for merchants, especially as more commerce moves online and new types of transactions, such as those made via Internet of Things-connected products, grow, Guru says.

“We realized that commerce is moving into newer challenges with the advent of the Internet of Things and card-not-present channels,” Guru says, with no indication of abating. A corresponding increase in fraud comes with that growth.

“Use of the Internet for both e-commerce and for the functional management of businesses and households is expanding as the depth and breadth of fraud and related opportunities to compromise systems also broadens,” says Beth Robertson, an independent payments analyst. “Since many of the payment mechanisms used for e-commerce rely on existing networks, and newer fraud-deterrence mechanisms like chip cards are not as effective for e-commerce purposes, there is a continuing need for an evolution in fraud-detection options.”

First Data is launching Fraud Detect in the United States and Canada first, with planned availability in other regions, Guru says. Fraud Detect is available to merchants of all sizes, and will be available to independent sales organizations to resell, he says. Pricing is dependent on the size of the merchant, a First Data spokesperson says.

Share |


Read Digital Transactions Online
read more