Machine learning and AI for more secure banking

Published on the 26/04/2016 | Written by Donovan Jackson


Turning data into real-time decisions to fight financial fraud…

Ever had a credit card declined owing to a false positive? It’s not an unusual experience and when it does happen, embarrassment and the hassle of dealing with the bank’s compliance people is sure to follow. With a new partnership for the APAC region, enterprise solutions outfit Unisys is looking to improve the situation by more accurately identifying and stopping actual fraudulent transactions and eliminating the bad calls.

Achieving all that comes as the result of a partnership it has forged with Brighterion, a provider of real-time analytics software. Unisys, with its hooks into the financial services sector, is to act as Brighterion’s solution integrator within APAC.

iStart got on the phone to Singapore-based Ian Selbie, Unisys’ APAC banking and financial crime expert, to find out more, starting with the nature of the problem that the solution addresses. “There are three layers of security for banks, generally – the front end systems, which really don’t get compromised very often, then a second line of authentication including perhaps biometrics, which won’t stop fraudsters who have genuine [stolen] credentials, then the final line is real time analytics.”

While banks have long used systems to detect anomalies in credit card usage, those anomalies are notoriously difficult to identify. After all, people travel, use credit cards for every kind of transaction online and on the street, and buying something in India can be done from a mobile phone on the bus in Wellington.

That goes to the heart of the Brighterion hookup, explained Selbie. “At the core of it is smart profiling of each individual customer’s accounts, their cards, the POS devices. Every element of the network profile is examined in detail and the system looks for exceptions using rules and machine learning to find patterns; if something exceptional is detected, a real time alert is raised.”

It’s the false positive which is a real problem for customer experience, said Selbie. “If a dodgy transaction is identified ‘too late’, a loss can be incurred, if not by the cardholder, then somewhere in the value chain. If a valid transaction is stopped, it is a hassle which can involve cancelled cards and a very bad experience for the customer.”

Putting a figure to it, he said the Brighterion software means one tenth of the number of false positives. “That means a flow on effect in the reduction of the costs of resolving potential fraudulent transactions and a better experience for bank customers.”

Selbie said Unisys is in the process of engaging with and demonstrating the solution to major banks in Australasia.

Post a comment or question...

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

MORE NEWS:

Processing...
Thank you! Your subscription has been confirmed. You'll hear from us soon.
Follow iStart to keep up to date with the latest news and views...
ErrorHere