dcsimg
Online sales currently generate around €2 trillion in revenue worldwide, and are growing at a rate of roughly 25% per year. But this is a market that involves costly risks for online sellers. One such risk is credit card fraud. Munich Re and the start-up Fraugster have joined forces to offer a solution that reduces this risk, lowering online sellers’ costs and increasing their profits.
© Getty Images/KTSDESIGN/Science Photo Library
Digitalisation

Artificial intelligence improves security of e-commerce

Online sales currently generate around €2 trillion in revenue worldwide, and are growing at a rate of roughly 25% per year. But this is a market that involves costly risks for online sellers. One such risk is credit card fraud. Munich Re and the start-up Fraugster have joined forces to offer a solution that reduces this risk, lowering online sellers’ costs and increasing their profits.

23.04.2018

Credit card fraud costs online traders around €25bn a year. In order to avoid payment defaults, traders, payment services and credit card companies currently check payments by hand or by means of unwieldy rule-based solutions, which is a laborious, expensive and ineffective process. 

The German-Israeli start-up Fraugster has developed a self-learning algorithm that checks in milliseconds whether online payment transactions are legitimate. This service allows payment companies to check transactions faster and more reliably, offering them, their clients and online sellers a new level of security for their business. Not only does the algorithm detect and, with a high level of probability, prevent fraudulent transactions in the first place, it also lowers the number of valid payments that are incorrectly rejected. Online sellers can thus reduce losses while at the same time increasing revenues from legitimate transactions. With several million transactions a day passing through the system, the AI technology ensures that the algorithm is continually honing itself and adapting to changes. Fraugster is liable for the costs clients incur if the system incorrectly deems a fraudulent transaction to be legitimate. 

Munich Re has analysed and tested Fraugster’s algorithm and found the quality of transaction checks and IT infrastructure to be fit for purpose and reliable. In doing so, Munich Re made a detailed analysis of the statistical models used, the self-learning principle of the algorithm, the predictive quality of the algorithm with data it has not yet encountered, the quality of Fraugster’s data scientists, the company’s desire for further development and the quality of its IT infrastructure. On the basis of this analysis, Great Lakes Insurance SE, a Munich Re subsidiary, insured Fraugster should the software return any incorrect results and fail to recognise fraudulent transactions. This means Fraugster offers its clients additional financial security and gives the service the seal of approval that Munich Re trusts the AI predictions of its product.

The cooperation between Munich Re and Fraugster is creating macroeconomic value: the innovative type of risk appraisal and tailored insurance solution combine to offer greater security, reducing sellers’ costs and increasing their revenue and efficiency. At the same time, Fraugster’s software also lowers the number of valid payments that are incorrectly rejected, thus breaking down barriers for those wishing to trade online. 

Munich Re Experts
Michael Berger
Business Development Manager
Knoerchen
André Knoerchen
Head of New Risk Solutions within the Special Enterprise Risks unit of Corporate Insurance Partner (CIP)
Munich
Print

We use cookies to improve your browsing experience and help us to improve our website.

By continuing to use our websites, you consent to the use of cookies. Please see our cookie policy for more information on cookies and information on how you can change your browser's settings.
You can disable cookies, however please note that disabling, deleting or disallowing cookies will affect your web experience.