Mastercard is set to integrate generative AI into its existing Digital Intelligence platform, enhancing its current Decision Intelligence (DI) capabilities.
DI currently processes and scores 143 billion transactions annually for banks in real-time. With the addition of generative AI, the system will analyze more than one trillion data points to determine the legitimacy of each transaction. The upgraded Decision Intelligence Pro is slated for release later this year.
“The precision of the solution—achieved by scanning potential points of sale in real time—has been shown in our own analysis to not only increase accuracy, but also reduce the number of false positives by more than 85%,” said Ajay Bhalla, President of Cyber and Intelligence at Mastercard, in a prepared statement.
Generative AI’s Potential to Transform Payments
Generative AI is hailed as the latest breakthrough poised to revolutionize payments. It promises to enhance fraud detection by analyzing vast transaction data, improve customer experiences through conversational chatbots, and streamline data reconciliation and reporting processes.
Despite its promising potential, accurately measuring the impact generative AI to overall fraud detection remains challenging. In his latest report, Generative AI Comes to Life: Notes from the Field, Christopher Miller, Lead Analyst for Emerging Payments at Javelin Strategy & Research, delved into the opportunities and challenges associated with implementing generative AI within the payments space.
“I would say broadly that based on our conversations with teams working on generative AI applications, they are generally hesitant to provide numbers that frame the impact of gen AI models, so we can take even such a broad claim of 20%-300% reduction with a grain of salt,” Miller said.
“Many companies are working on leveraging gen AI’s ability to include much larger data sets and do analysis at the transaction level rather than simply risk scoring individuals. The upsides mentioned here, with two types of improvement in fraud—e.g. reduction in false positives, and increased detection of fraudulent behavior—do offer real promise. Increased fraud detection combined with lower false positive rates reduce the ‘cost’ of fraud detection and make it less invasive to the customer experience of paying. We are characterizing this type of gain as ‘the hidden impact of a visible technology’ for this very reason, it’s unlikely to be obvious in any way to consumers that this is happening.”
