The Commonwealth Bank of Australia says generative AI tools have helped it lower frauds by 30% and reduce call center customer wait times by 40% in 2024
- Customer scam losses at Commonwealth Bank cut in half — CEO Matt Comyn says many more AI use cases in development Bluesky: @msbrumfield.bsky.social , @radiobeartime.com , @mfulk.bsky.social , and @benjialpha.bsky.social . X: @business Bluesky: Cynthia Brumfield / @msbrumfield.bsky.social : Excellent if this bears out. — “The technology helped lower frauds by 30% and its AI messaging app trimmed call center customer wait times by 40% this year.” — AI Cuts Scam Losses, Speeds Up Home Loans at Top Australia Bank — www.bloomberg.com/news/article... Mark O'Neill / @radiobeartime.com : Which frankly raises some very deep questions about their current business processes which the banking regulator might like to explore. [embedded post] @mfulk.bsky.social : Because people hung up on it? [embedded post] @benjialpha.bsky.social : 20 odd years ago, I'd get a phone call before weird international transactions could go through. Now they just let them go through no worries. [embedded post] X: @business : Australia's Commonwealth Bank said generative artificial intelligence tools cut customer losses from scams in half. It has also helped to shorten the process for mortgage pre-approvals to as little as 10 minutes. https://www.bloomberg.com/...
Context & Ripple Effects
Banks had already been pushed toward AI defenses as deepfakes and voice cloning made financial scams more scalable, as covered in the rise of AI-enabled financial scams. Commonwealth Bank now offers a concrete operating case spanning fraud controls and customer service rather than a single pilot.
The bank says it has further generative-AI uses in development, making these reported results a meaningful test of whether AI can improve both loss prevention and service throughput in a regulated, customer-facing workflow.
First-order effects
- Commonwealth Bank reports lower fraud and scam losses, while customers using its AI messaging channel face shorter call-center waits.
- The bank can move more routine customer interactions through messaging and accelerate mortgage pre-approvals, shifting immediate operational attention toward scaling the tools it has deployed.
Second-order effects
- Other banks and fintechs face greater pressure to match AI-assisted fraud detection and service responsiveness, particularly as deepfake- and voice-cloning scams raise the cost of slower defenses.
- Success in both fraud and service creates a stronger business case for prioritizing AI investment around measurable outcomes—loss avoidance and faster resolution—rather than broad, undifferentiated chatbot rollouts.
Third-order effects
- If repeatable across institutions, banking AI adoption could become organized around end-to-end operational workflows where a model’s value can be tied to a lower cost per useful task, not simply customer-facing novelty.
- The same automation trend may reshape bank staffing and process design over time; the survey projecting AI-related banking job reductions signals the labor question, though this report itself documents service gains rather than workforce changes.
The trend: Financial institutions are moving generative AI from experimentation into fraud prevention and high-volume service workflows where outcomes can be measured in losses, speed, and capacity.