Broadcom Inc. (NASDAQ:AVGO) has launched Arcot Smart Ruleset, a machine learning–based fraud detection platform designed to enhance security for 3-D Secure payment authentication.
The new system replaces traditional manual rule-writing processes with automated decision logic that can adapt to evolving fraud patterns and regulatory requirements.
Arcot Smart Ruleset is built on machine learning models trained using more than 20 years of transaction data collected from over 5,500 financial institutions. By evaluating real-time transaction signals, the platform automatically generates fraud prevention rules based on patterns observed across billions of transactions each year.
Payments organization Velera reported that it prevented more than $2 million in additional fraud during the past six months using the technology.
“The Arcot Smart Ruleset strengthens our fraud prevention by using machine learning to uncover risk patterns that complement our existing controls,” said Sergio Valencia, vice president of authentication risk management at Velera.
Traditional fraud systems typically depend on fixed “if/then” rules written and updated by human analysts. According to Broadcom, these manual approaches struggle to keep pace with increasingly sophisticated fraud tactics powered by artificial intelligence and can often generate false positives that block legitimate transactions.
The platform also includes a 3-D Secure Regulatory Engine designed to automatically handle protocol updates, certification processes and compliance requirements.
Matt Cooke, vice president and general manager of Broadcom’s Payment Security Division, said the technology “frees fraud teams from tedious manual rule-writing so they can focus on higher value activities such as investigations and fraud strategy.”
Broadcom’s Arcot business unit specializes in 3-D Secure authentication technology for digital payments. The Palo Alto–based company develops a broad range of semiconductor and infrastructure software solutions.