When you swipe your card or click “buy” on that website, many processes and interactions take place that you’ve probably never heard about: A complex web of organizations engaging with each other to process your payment fairly, efficiently, and safely. That web is largely driven by cutting-edge technology, which you also don’t hear about.
One of those technologies is artificial intelligence (AI). In this analysis, I break down key uses for automation and AI in the payments industry, and take a deep dive into RTP Prevent, Visa’s AI-powered fraud prevention tool for real-time payments.
For context, I run a startup called Sweet that is building peer-to-business payments for bars and clubs. Think Venmo but for sending money to businesses on-premise. We’re at an early stage and currently building AI into our product roadmap in different ways.
Real-Time Payments
One major trend in the payments industry is a move towards real-time payments. Bank transfers take days. Checks take time to cash. Card processing doesn’t settle in real-time. But services like RTP, Zelle, Venmo, and CashApp are riding on a wave of new consumer demand for instant payments. For this reason, they’ve collectively gained hundreds of millions of users in a short amount of time, and businesses are increasingly accepting real-time payments as a method of payment.
But real-time payments are tricky. RTP transfers must settle near-instantaneously, leaving no time for the transacting banks to prevent fraud, confirm the transaction, and build in a buffer time during which parties can request refunds. Financial institutions aren’t able to recall or retrieve funds once they’ve been sent via a real-time payments network, a juicy proposition for criminals or fraudsters.
This is where AI comes in.
Which companies are the most important vendors in AI and hyperautomation? Check out the Acceleration Economy AI/Hyperautomation Top 10 Shortlist.
Visa’s RTP Prevent
Visa’s recently announced RTP Prevent uses AI to analyze transaction data in real time and provide a rapid risk assessment to the financial institutions involved in a real-time payment.
This helps financial institutions make a better decision on whether to authorize the transaction, catching fraud before it happens and increasing the overall safety and security of real-time payment networks.
RTP Prevent is currently an API that integrates with other risk assessment tools. Visa continuously improves the model with new data, helping financial institutions counter new fraud methods as they are discovered.
Visa’s innovation is especially relevant given the recent release of FedNow, which is the public Fed-operated payment network for instant transfers. While FedNow’s RTP has existed for many years, there are certain advantages that a Fed-operated real-time payment system will bring. Leveraging AI to reduce the risk of these systems has tremendous value both to the exchanging banks and the final consumer.
Conflicts with AI and Fraud in Payments
Fraud is obviously a big topic in payments. New technology frequently creates new ways of exploiting payment systems and taking other people’s money. But financial institutions and payment companies use fire to fight fire: In many cases, the technology that enables the problem is also used to solve it. AI is no different.
One major concern is that generative AI is allowing malicious actors to create synthetic profiles and skirt existing know-your-customer (KYC) measures like checking your ID and taking a photo of yourself. With ever-improving generative AI systems, someone can generate a suite of paperwork, biometric information, pictures, and even things like social media accounts, all to imitate or create a person that doesn’t exist. While the implications of this in relation to the movement of money are very bad, they go far beyond payments. Governments will need to improve their ways of tracking citizen identities.
But AI is also helping to prevent payment fraud. From Visa’s RTP Prevent to other risk assessment tools to computer-vision-enabled KYC, AI is already solving many problems in the industry. Another key use case is the categorization of transaction types, which can help both consumers (how did I spend on shopping versus groceries?) and businesses (how are my customers transacting and what areas experience the most fraud?).
The Future of AI in Payments
AI is undoubtedly impacting all industries. In payments, the tech will continue to accelerate larger trends within the industry that are already taking place — whether that’s cross-border payments, real-time payments, increased levels of personalization for users, more efficient compliance, and even the industry’s crossover with other tech like blockchain or DeFi.
One of the things that excites me about the payments industry is its speed. People really care about their money and how it moves. If a technology makes that process easier, smoother, or safer, it gains traction rapidly. AI is no different. I can’t wait to see what’s next and how my startup Sweet will evolve around the tech.
