A fraud prevention cloud system uses machine learning and big data analytics to detect potentially fraudulent activities and protect multiple customer service channels in banks, financial institutions and government agencies, says cybersecurity multinational Kaspersky lab head of fraud prevention Alexander Ermakovich.
The cloud platform provides fraud prevention solutions for end points and mobile devices in organisations’ environments, as well as fraud prevention cloud technologies, including a device-reputation database, device and environmental analysis, behavioural analysis and biometrics, and clientless malware detection.
“The new fraud prevention offering from Kaspersky Lab delivers multichannel protection for organisations and users, resulting in reduced losses because of fraud and controlled prevention costs. “The solution deploys advanced technologies to improve the visibility and detection of suspicious activity without undermining user experience,” he avers.
Behavioural analysis and biometrics help to identify whether a person’s claimed identity is true without any additional actions or procedures required by the user. Behaviour is analysed through mouse movements, clicks, scrolls, keystrokes on personal computers, and accelerometer or gyroscope position and gestures, such as touches and swipes, on mobile devices.
“The Kaspersky Fraud Prevention Cloud gathers and analyses user behaviour, device, environment and session information – anonymised and depersonalised big data in the cloud – making it available to expert forensics and automatic off-line analysis.
“This information feeds into an organisation’s in-house enterprise fraud management system, allowing for proactive fraud detection in real time, even potentially before a transaction occurs. This approach is based on a combination of big data and threat research analysis, with machine learning algorithms bolstered by the expertise of the company’s best security teams, called Humachine intelligence,” explains Ermakovich.
Additionally, risk-based authentication assesses the risks before a user is permitted to log into a digital channel, providing decisions for in-house back-end systems on whether to proceed, request additional authentication information or block access until further verification. This feature improves usability for legitimate users by decreasing the number of authentication stages, while unauthorised users are detected before they can commit any fraudulent activity.
The cloud solution’s continuous session anomaly-detection technology further helps to improve fraud detection by identifying account takeover, new account fraud, money laundering and the use of automated tools or any suspicious processes that occur during the session.
The Kaspersky Fraud Prevention Cloud is active during not only the login process but also the whole session, building statistical models of various behavioural patterns with the help of machine learning technologies,” he adds.
The clientless malware detection combines direct and proactive detection techniques.