Scam Indicator, currently in proof-of-concept stage and to be piloted over the next few months, will help protect joint CBA and Telstra customers from phone scams where criminals try and trick people into transferring them large amounts of money.

Early results of simulations suggest the prospect of improvements in the detection and prevention of scams, potentially mitigating customer losses of up to $15-20 million.

The tool is designed to detect certain high-risk scam situations in real time using a Telstra API that CBA will call on as part of its scam detection processes. This enables CBA to check if a customer is on a phone call - the prime indicator that a scam is occurring. The bank can then contact the affected customer or put in additional checks, including delaying, blocking or interrupting transactions, as well as sending a notification into the customer’s banking app. .

Matt Comyn, Commonwealth Bank CEO says: "We have been working with Telstra to produce a machine-learning scams detection model. We acknowledge that there is more to do given the rising volume and fast changing nature of scams."

The project builds on Telstra’s ’Cleaner Pipes’ initiative, which uses machine learning to actively block potentially unwanted email, SMS and phone calls with improved spam and scam filters. Since launching, it has blocked close to 230 million SMS messages and an average of 10 million scam or potentially unwanted calls every month.

Vicki Brady, Telstra CEO says: “Australians are under constant threat by scammers, who are getting increasingly sophisticated and persistent. Cyber security and addressing the issue of scams is a team effort, and we should be doing all we can to help stem the flow of this criminal activity. Part of this needs to be organisations working together like this to share knowledge and insights and partnering to develop solutions that can help better protect our customers.”

The pilot programme will be rolled out in the middle of this year, with more banks expected to join the initiative in the latter part of 2023.