Visa, Mastercard Pave the Way for AI in Payments; What Does It Mean for Workforce?
By Yamini Kalra
The payments industry, which has operated on established frameworks for decades, is adapting – or forced to adapt – to a new kind of intelligence. For industry heavyweights Visa and Mastercard, artificial intelligence is becoming a tool to drive fraud detection, customer experience and operational agility.
“The payments industry evolves in waves … and [AI] is going to be a bigger wave than we have seen before,” said director general of The Payments Association Tony Craddock.
The payments industry is inherently dynamic, driven by consumer expectations for speed, security and convenience. AI offers that and much more, with an obvious sprinkle of challenges and workforce disruption concerns. So far, companies are making AI their friend, investing billions of dollars in anticipation of a faster and safer payments landscape.
Visa has already deployed 500 generative AI applications as part of its “go-fast” strategy aimed at staying ahead of cybercriminals. One such AI tool is aimed at identifying enumeration attacks, which cost the company around US$1.1 billion annually in fraud losses. The payments giant has spent more than US$3 billion on AI and data infrastructure in the last decade. Last year, it launched a US$100 million venture fund for gen AI startups.
“While much of generative AI so far has been focused on tasks and content creation, this technology will soon not only reshape how we live and work, but it will also meaningfully change commerce in ways we need to understand,” said chief product and strategy officer at Visa Jack Forestell.
Meanwhile, Mastercard invested over US$7 billion in cybersecurity and AI over the past five years. The company claims it is using gen AI to “double the speed” at which it detects potentially compromised cards. Its integration of Risk SDK technology allows merchants to capture device data, including geolocation, IP address and device fingerprinting. This data is then analyzed in real time, creating a multi-layered fraud prevention system.
Visa has reaped similar benefits. The company processes more than 500 million transactions daily, using advanced AI models to detect fraudulent activity.
Fraud Prevention: AI’s First Line of Defense
Companies are using machine learning to detect and block fraudulent transactions in milliseconds. In 2023, Visa’s AI-driven tools blocked US$40 billion worth of fraudulent transactions, and Mastercard said the technology boosts fraud detection by 300 per cent. These systems learn from an immense volume of transactional data, finding patterns and anomalies that even seasoned fraud analysts may miss.
American Express experimented with ML back in 2010 for fraud prevention and credit modeling. The company, in its annual report for 2023, claimed investments in AI helped maintain the lowest fraud rates in US among major credit card networks.
But for every dollar saved in fraud prevention, there’s a question about how much autonomy AI should have in sensitive financial decisions. And while AI secures such transactions, it also places companies in a position of balancing trust with control – a tension especially acute in finance.
Reduction in False Positives
The use cases for integrating AI is strong in other areas as well. Businesses lose out between 1.5 per cent and 2.2 per cent of their revenues due to “suboptimal payments acceptance” or false positives, where legitimate transactions are mistakenly denied by the payment processor. More than 45 per cent of consumers do not retry payments after a false positive, choosing instead to take their business to a competitor. It is a problem that has persisted – even grown – in the payments industry.
But AI can mitigate some of these losses. Mastercard claims its in-house predictive technology, trained on 125 billion payment transactions, reduces false positives by 200 per cent. And Visa said its VAAI Score helped achieve an 85 per cent reduction in false positives compared to traditional risk models by honing in on specific signals unique to enumeration attacks.
But despite the clear benefits of AI, its adoption is uneven. A staggering 86 per cent of merchants say their payment service providers do not provide ML solutions for better fraud detection and increased authorization rates. With 68 per cent of CFOs citing the need to improve payment acceptance rates, investment in AI solutions is no longer optional.
The Double-Edged Sword of Gen AI
Sophisticated gen AI tools can bypass voice authentication protocols using audio cloning, demonstrating how technology can be both a shield and a weapon in the hands of bad actors. Future fraud detection systems may integrate behavioral biometrics, analyzing typing patterns, swiping motions and other user-specific behaviors to authenticate transactions more securely.
Amex anticipated risks to its “brand and reputation” due to the increased sophistication of AI that can assist with the creation of deepfakes.
Amex uses gen AI tools to analyze transaction data and anticipate customer needs. For instance, frequent travelers might receive real-time notifications about travel-related benefits or fee waivers, creating a seamless and personalized user experience. Mastercard’s AI-driven virtual assistant, launched in 2024, provides personalized financial insights and handles routine customer inquiries. By analyzing spending patterns, the assistant helps users manage budgets and optimize their finances, reducing the burden on human agents and cutting down response times.
Companies are also using ML to optimize transaction routing, reduce fees and accelerate settlement times. Stripe processed over US$1 trillion in payments in 2023, with AI playing a pivotal role in managing the complexities of global transactions.
Workforce Disruption: Adapting to Automation
The Wall Street Journal reported Visa will lay off 1,400 employees – including 1,000 tech roles – by the end of this year as it accelerates its AI adoption journey. The company envisions a future where human employees oversee AI “digital employees,” with one worker supervising 8 to 10 AI-driven systems handling various tasks.
AI-driven automation is changing the nature of traditional roles within these companies. With chatbots handling routine inquiries, customer service roles are shifting from direct interaction to managing AI systems, analyzing data and handling escalated issues. Employees are expected to transition from transactional roles to more strategic ones. Mastercard has invested in training programmes to equip staff with these evolving skills. Yet this transition can be challenging for employees unfamiliar with data science or ML.
The rapid pace of automation means not all employees can adapt quickly, leaving gaps in workforce readiness. Customer service roles, back-office operations and routine transaction monitoring are particularly susceptible to automation and fear redundancy.
The industry stands at a pivotal moment, where the right balance between innovation and responsibility can unlock unprecedented opportunities. Data privacy, however, remains a critical concern, and a standard AI governance framework to ensure responsible use of AI is becoming increasingly important.
