With SAFE, F5 Employs AI to Take Fraud Detection to New Level

Cybersecurity vendor F5 has announced a services- and AI-based fraud detection solution it says will catch more fraudulent online transactions than other tools.

The Shape AI Fraud Engine (SAFE), offered as cloud-based service, uses an artificial intelligence engine to examine every online transaction throughout a user’s journey. That includes everything from site arrival, registration and login to updating stored account details, redeeming voucher codes, checkout and payment. According to the company, SAFE uses AI to detect and block malicious intent, and it can find twice as much fraud compared with other fraud detection tools.

To do this, SAFE uses JavaScript to collect advanced telemetry data from each application session, including a broad set of environmental and behavioral signals, explained Mike Plante, F5’s vice president of Shape product marketing. Those telemetry signals are analyzed by a real-time processing engine, which determines in milliseconds if a user's intent is malicious.

“Instead of simple rules that can be reverse-engineered by fraudsters, SAFE focuses on understanding three basic questions: Are you human, are you good or bad, and are you who you say you are,” he said. “No one telemetry signal is enough to accurately answer these questions — that is, to determine if user intent is bad or not. Shape AI Fraud Engine uses machine learning and artificial intelligence to evaluate different combinations of signals to determine what patterns are indicative of fraud.”

For example, scammers are more likely than legitimate users to use copy-and-paste keyboard shortcuts to fill out login or new application webforms and switch in and out of the browser. “If you stack up enough of these combinations of signals into behaviors and evaluate them correctly, you can accurately determine when someone is intending to commit fraud. SAFE evaluates hundreds of environmental and behavioral factors to evaluate intent, and identify and stop online fraud,” Plante said.

Akif Khan, a senior director at Gartner, said F5 isn’t alone in using AI for fraud detection. At the same time, he said the vendor’s approach of looking at multiple events across the entire customer journey is right in line with a broader shift in the market. “Many legacy solutions in the market today are still overly focused on only the point of purchase,” he said.

While reducing fraud is a worthwhile goal, Khan said the real pain points for the customers he talks with are how to maintain “acceptable” levels of fraud without negatively impacting the customer experience, without declining good customers and without increasing operational costs with manual interventions. Any new solution that wants to be successful in the market today needs to be able to clearly address those issues, he said.

F5 plans to continue to add new features to SAFE that support both its financial services and e-commerce customers, including enhanced reporting and data visibility, Plante said.