Marqeta Delivers AI-Driven Risk Decisioning to Enhance Real-Time Fraud Prevention

WHY THIS MATTERS: The global payments industry is battling an alarming increase in sophisticated fraud, with projections indicating a 153% rise in global losses between 2025 and 2030. This surge is forcing card issuing platforms to move beyond static, rules-based systems toward dynamic intelligence. Marqeta’s integration of an AI-powered risk score directly into the authorization flow signifies a critical shift in how enterprises manage risk at scale. This development is not merely about blocking bad transactions; it’s about maximizing revenue by reducing false declines—a major hidden cost for merchants. By analyzing hundreds of transaction attributes in milliseconds, the new Real-Time Decisioning capability allows businesses to confidently scale their operations. It positions the platform as a core defense against evolving threats like enumeration attacks and the forthcoming challenges of agentic commerce. This is a value-first necessity for any organization relying on modern card programs to drive growth.

Marqeta, Inc. (NASDAQ: MQ), the modern card issuing platform, announced it enhanced its Real-Time Decisioning (RTD) offering with an AI-powered risk score that analyzes transaction risk levels at the point of the authorization decision. With the new integrated capabilities, Marqeta enables its customers to make smarter, data-driven risk assessments to prevent payment fraud and reduce false declines. Part of its suite of risk services, the company’s RTD offering allows customers to create custom rules and controls to manage fraud.

The risk score is integrated into Marqeta’s platform, analyzing proprietary transaction data from its card programs to detect risk patterns specific to each customer’s use cases. The score evaluates 300+ real-time transaction attributes against historical behavioral patterns, automatically adapting to unique cardholder behaviors and market shifts. This enables customers to assess each transaction’s risk level, take immediate action, and run risk scenarios that are based on their actual cardholder data – all with millisecond-level response times. 

Payment fraud is getting more sophisticated, with global payment fraud expected to rise by 153% between 2025 and 2030. With this level of sophistication, fraud detection models need to be able to detect new fraud patterns and stay ahead of developing threats. Marqeta addresses this challenge by combining its RTD authorization rules with the predictive power of machine learning in order to continuously identify new fraud patterns and fight against emerging threats. 

“Today’s fraud threats are evolving faster than ever, requiring businesses to keep pace as they scale their card programs,” said Anthony Peculic, Marqeta’s Interim Chief Product Officer. “By embedding AI-powered controls and advanced machine learning into the authorization process, we enable customers to expand confidently while also strengthening their fraud defense as they scale.”

Marqeta’s suite of risk services, RiskControl, includes Know Your Customer (KYC), 3D Secure, Real-Time Decisioning (RTD), and end-to-end disputes management. Learn more about Marqeta’s Real-Time Decisioning product here. 

FF NEWS TAKE: This upgrade is a necessary evolutionary step in fraud prevention that moves the needle on platform capability. The trend of embedding machine learning into core infrastructure is essential for keeping pace with organized crime and sophisticated bots. The next major challenge, however, will be adapting these controls to govern autonomous AI agents that transact on behalf of consumers. Future development must focus on strengthening authorization frameworks to ensure accountability in the age of agentic commerce.