Commonwealth Bank has rolled out an agentic artificial intelligence system designed to identify emerging fraud and scam patterns across payments and transaction data. The system also generates detection rules to help block suspicious activity as it develops, according to the bank.
The deployment is part of the bank’s broader investment of around $1 billion annually in fraud prevention, cyber security and financial crime controls. The system builds on existing monitoring capabilities that analyse more than 80 million data signals daily across transactions, cards, and digital banking channels.
AI Moves from Detection to Action
Unlike traditional fraud systems that rely on predefined rules, the new AI agent is designed to detect unfamiliar patterns, assess their severity, and recommend new controls. It operates continuously, analysing large volumes of data and adapting to new threats as they emerge.
When the system identifies suspicious behaviour, it proposes new fraud detection rules, which are then reviewed by the bank’s fraud analytics team before being deployed. This “human-in-the-loop” model ensures oversight while allowing faster response times to evolving risks.
The bank processes more than 20 million payments each day and sends over 40,000 alerts to customers through its mobile app. According to internal data, its fraud systems contributed to a more than 20% reduction in fraud losses in the first half of the 2026 financial year compared with the same period in 2025. The new AI agent has also been used to develop or update roughly three quarters of the bank’s card fraud detection rules.
The system was developed internally by the bank’s data science and engineering teams in around three months. It is built on Snowflake’s data cloud and integrated with the bank’s cloud-based core banking platform to enable real-time data access and processing.
About Commonwealth Bank: Commonwealth Bank is one of Australia’s largest financial institutions, providing retail, business and institutional banking services.
The shift toward agent-based AI signals a move from static fraud controls to adaptive systems that can evolve alongside increasingly complex financial crime patterns.
Digital Trade Outlook Analysis
As cross-border payments and digital transactions scale, fraud risks are becoming more dynamic and harder to detect using traditional systems. AI-driven detection models that can adapt in real time are likely to become a standard layer in payment and trade finance infrastructure. For banks operating across global trade flows, speed in identifying and responding to threats is now as critical as prevention itself.
Source: Commonwealth Bank.