Sift , an AI-driven fraud prevention company, has announced updates to its account takeover (ATO) solution, enhancing fraud risk protection throughout the entire consumer journey. Sift’s detailed analysis of its customers shows that its ATO solution prevents a median of up to USD 4.2 million per year in fraud losses per customer. The system is designed to protect user accounts from login through to post-transaction activities, targeting the rising incidence of ATO, which accounted for nearly USD 13 billion in losses in 2023.

With the growth of the digital economy, ATO attacks rose by 24% in Q2 2024 compared to the same period in 2023. Traditional methods that primarily focus on login-stage defences have proven insufficient, as they often fail to prevent attacks occurring at other stages of user engagement. ATO can sometimes lack dedicated oversight within organizations, leading to gaps in accountability between fraud and security teams.

Sift’s solution aims to close these gaps by integrating data and processes accessible to both departments. The latest updates to Sift’s ATO solution include: Identity-Centric Accuracy: Incorporates AI-driven behavioural analysis and device fingerprinting to assess risk and user intent throughout the customer journey. Enhanced Integrations: Supports existing Customer Identity Access Management (CIAM) systems with low-code integrations, including compatibility with Ping Identity PingOne DaVinci and Okta Auth0, to strengthen identity management across security systems. Advanced Controls: Provides detailed multi-factor authentication and session monitoring, coupled with pre-configured, industry-specific automation tools. Additionally, Sift introduced new behaviour tracking features and a VIP Fast Pass for industries with high transaction volumes, such as iGaming and Fintech.

Expanded functionality in RiskWatch aims to streamline manual reviews for faster, more comprehensive assessments. Sift provides AI-driven fraud prevention for over 700 customers globally, offering solutions built on machine learning and extensive user data networks that process approximately one trillion events annually. .