New Visa service deploys machine learning-based model to help predict probability of token fraud before it happens.
Visa launches AI-based Visa Provisioning Intelligence (VPI) product. (Credit: CardMapr.nl on Unsplash)
Visa (NYSE: V), a global leader in payments, today announced the commercial launch of Visa Provisioning Intelligence (VPI), an AI-based product designed to combat token fraud at its source. Available as a value-added service for clients, VPI uses machine learning to rate the likelihood of fraud for token provisioning requests, helping financial institutions prevent fraud in a targeted way and enable more seamless and secure transactions for Visa cardholders.
Tokenization is a powerful fraud-fighting technology pioneered by that helps protect consumers’ account information from bad actors by replacing account numbers with a unique code. However, tokens may be illegitimately provisioned to bad actors, and Visa found that token provisioning fraud losses reached an estimated $450M globally in 2022 alone.
“While tokenization is one of the most secure ways to transact, we’re seeing fraudsters use social engineering and other scams to illegitimately provision tokens,” said James Mirfin, SVP and Global Head of Risk and Identity Solutions at Visa. “With VPI, we’re leveraging Visa’s vast network and data insights to help clients detect and prevent provisioning fraud before it happens.”
VPI is a real-time fraud propensity score between 1 (indicating the lowest probability of fraud) and 99 (indicating the highest probability of fraud)2 provided to issuers for each token provision request. VPI uses a segment-level supervised machine learning model to identify patterns in past token requests across device, e-commerce and card-on-file tokens to help predict the probability of token provisioning fraud. The VPI score is intended to provide financial institutions with the following benefits:
Improved fraud prediction by allowing issuers to detect provisioning fraud and decline a token provisioning request before the fraud occurs.
Accurate separation of fraudulent activity from non-fraudulent activity reducing the number of declines.
An increased number of legitimate provisioning requests, increased payment volumes and continued trust in the card payment network.
Source: Company Press Release