Broadcom Inc. (NASDAQ: AVGO) announced the Arcot Smart Ruleset, a machine learning-powered fraud detection system for 3-D Secure payment authentication. The system replaces manual fraud rule-writing with automated logic that adapts to evolving threats and regulatory requirements.

The Arcot Smart Ruleset uses machine learning models trained on 20 years of transaction data from over 5,500 financial institutions. The system analyzes real-time transaction signals and automatically creates fraud prevention rules based on patterns identified across billions of annual transactions.

Velera reported preventing over $2 million in additional fraud in the past six months using the system. "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 prevention systems rely on static "if/then" rules written and maintained by humans. Broadcom states these manual systems cannot scale against fraudsters using artificial intelligence and often result in false-positive declines that block legitimate customers.

The system includes a 3-D Secure Regulatory Engine that manages protocol upgrades, certifications, and compliance requirements automatically. Matt Cooke, vice president and general manager of Broadcom's Payment Security Division, said the system "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 division focuses exclusively on 3-D Secure payment authentication technology. The company is based in Palo Alto, California, and develops semiconductor and infrastructure software solutions.