Feedzai Partners Azul Systems to Deliver Ultra-Fast Fraud Prevention Solutions

Blog entry

Azul Systems has signed up Feedzai, a company that uses real-time, machine-based learning to analyse big data in commerce, as an independent software vendor that will use its Zing Java virtual machine (JVM) to power ultra-fast processing of big data within the Feedzai Fraud Prevention That Learns software.

In the financial services sector, Feedzai works predominantly in retail banking, delivering enterprise technology solutions that combine big data and machine learning to allow analysts to predict and prevent electronic payment losses in real time on the basis of behavioural analysis of customers. The company has customers among the world’s top five retail banks and notes that its technology is an equally good fit for investment banks, which like retail banks, for example, must apply Know Your Customer (KYC) and Anti-Money Laundering rules.

Loc Nguyen, chief marketing officer at Feedzai, explains: “KYC requires banks to gather more customer data than they have in the past, yet banks want to take the friction out of customer on-boarding and ask less questions. With Zing, we can reduce the number of questions on an application form from 20 to 30 to just four or five. Less is asked, but more information is gathered in the background as our fraud engines dig into a customer’s history and make it possible to see if an account opening request makes sense. The speed of Azul’s Zing JVM means we can consume and process more data than we could before in the same time. This will allow us to build ultra-smart artificial intelligence machines.”

Feedzai was using a JVM to process big data before partnering with Azul, but was pushing the solution to its limits and looking for more speed and consistency. Azul’s Zing provides not only speed, but also the elimination of Java garbage collection pauses that can be disruptive.

Nuno Sebastiao, CEO of Feedzai, says: “The real-time analysis of data in the financial industry is key to predicting and preventing fraud. It’s almost impossible to have ultra-low latencies, in the range of 5 to 10 milliseconds, with a standard JVM and our customers demand that. Azul powers the largest banks in the world and with peak load demands of up to 50,000 transactions per second, Zing will help ensure that we can deliver the best that artificially intelligent machines can offer.”

Scott Sellers, CEO of Azul Systems, says the evolving and ever more demanding financial services sector is a focal point for the company, and adds: “Feedzai’s work in fraud protection is an interesting space for us as more and more organisations are looking at better ways to detect and prevent fraud.”