Credit cards Credit card scammers have no chance

At the UBS Card Center, human and artificial intelligence is working hard to combat credit card fraud.

by UBS Insights 09 Sep 2019

Anna Bettoni, team leader of the 24-hour fraud prevention team at the UBS Card Center, sits in front of the two big screens at her desk and watches the credit card transactions of all UBS clients like a hawk. The data flits rapidly across the screen. It doesn’t take long for the software to emit its first warning. Anna assesses the possible fraud incident straight away.

As in most cases, the system was not wrong: Six hours after a purchase in a St. Gallen boutique, the cardholder apparently visited an ATM in Manhattan and tried to withdraw cash. “The system recognized how unrealistic this situation is and prevented the 200-dollar withdrawal,” says Anna. The client is then swiftly informed of the incident by phone and their UBS Credit Card instantly replaced.

Artificial intelligence (AI)

The software that’s been successfully preventing credit card fraud at the UBS Card Center for over 20 years is the FICO Falcon platform. In the 24-hour operation, 500,000 transactions by people and machines are assessed every day. That’s 25% of all credit card transactions in Switzerland. The system compares all transactions based on model profiles, various rules and known practices used by scammers.

A myriad of rules and behavior patterns have been fed into the system. The volume, type and amounts of all transactions are assessed. AI also examines card use at branches and other locations and triggers a system warning if it detects an inconsistency.

Not every suspicious case proves to be true

“The system is sometimes wrong because people can behave unpredictably in certain situations,” explains Anna. She gives an example of a case from last summer: Before going on vacation, a lady had run out of time to stock up on some essential items. So, on the day of departure, she went to the Zurich main train station and bought what she needed in various stores. The system identified this as a common credit card theft pattern: a lot of small transactions under 40 Swiss francs made within a short space of time, using NFC and no PIN. “We called her immediately. She just laughed, thanked us and explained the situation,” says Anna, with a smile.

Human intelligence

Every day, the fraud team at the UBS Card Center works on making the early warning system even more accurate. The FICO Falcon platform is a learning system – the keyword here being “machine learning.” Based on employee input, the system constantly creates new rules and learns how to interpret behavior patterns correctly. According to Marcel Drescher, Head Fraud Services at the UBS Card Center, the biggest strength of his department lies in processing potential cases of fraud. “Our main expertise is interpreting and analyzing individual cases. The system only learns effectively if we feed it the right data.” For this reason, between 2015 and 2018 our fraud prevention rate increased by 84%,” he proudly notes.

Will credit card scammers soon be out of a job?

From 2000 to 2016, the number of fraud attempts rose continuously. The sale of stolen credit card data on the dark web was a major factor here. However, since 2017 this trend has stabilized for many reasons. Different security mechanisms and technological developments are increasing security for credit card users.

The four-party system

These mechanisms and controls run in the background of every secure payment transaction.

“Alongside the constant optimization of our early warning system, the introduction of the new 3-D Secure protocol has also had a direct impact. Immediately following its introduction, we found that the number of fraud cases among our clients went down,” says Marcel.

Pickpocketing is on the increase again

At present, approximately 90% of all fraud attempts take place online. As these attempts are becoming easier to foil, an old-fashioned type of stealing is making a comeback: pickpocketing. The stolen credit cards are used immediately for shopping or withdrawing cash from ATMs. “Our early warning system recognizes this pattern and prevents it.” Credit card users can foil possible theft attempts in two steps: First, they can activate transaction notifications in UBS E-Banking or the UBS Mobile Banking App. They can then block their card in the UBS Mobile Banking App if it's being misused. A replacement card will be issued immediately.

The early warning system currently detects 85% of all fraud attempts automatically. The other 15% is reported by clients. In the future, client input will also be entered actively into the system so that credit card fraud can be prevented even more proactively.

Outstanding security

In April 2019, the UBS Card Center won the “Security Innovation of the Year” award at the Retail Banker International Awards 2019. The international jury praised the excellent work of the team responsible for fraud cases. This team uses AI functions and machine learning on the FICO Falcon platform. In 2018, this approach stopped 84% more fraudulent credit card transactions than in 2015.