Around a year ago, the topic of the moment in the cross-border payments industry was artificial intelligence. Buoyed by the rise of large language models and generative AI, the technology was the primary source of hype for the sector, with almost daily announcements focusing on AI-based solutions and a host of thinkpieces extolling the potential of the technology.
One year on, things are a little bit different. AI hype has by no means vanished – in the wider business world it’s barely out of the headlines, as ongoing interest in productivity meets the growing concerns of an AI bubble. However, within cross-border payments, the tone has visibly shifted.
AI is still a focus of hype, but it is now sharing the stage with stablecoins, which has dampened some of the most hyperbolic impulses around the technology. Meanwhile, public companies that began AI initiatives are slowly starting to report initial results, providing tangible metrics on which to build further strategy.
AI discussion has become more focused in payments, not less prolific
With stablecoins drawing so much attention, it would be easy to think that the rate of discussion of artificial intelligence has gone down. However, that does not seem to be the case. Analysis of the public earnings calls of 25 major cross-border payments-related players by my own company FXC Intelligence shows that the number of mentions of the technology more than doubled in Q3 2025 compared to the same quarter in the previous year.
Press releases about the technology’s use in payments have also increased, albeit more slowly, growing by around 10% year-on-year.
However, what is notable is how the technology is being discussed. Previously, the topic has often been brought up by companies who only discussed it in abstract, or with vague plans light on details or firm timelines. By contrast, the discussion now is highly intentional, and typically leans towards launched products or tangible results.
For many, the focus so far has been on using AI to improve productivity, with banking giants Citi and Goldman Sachs both highlighting the technology’s use in operational settings such as, in the case of Citi, code reviews. This has even extended to projected financial benefits, with business payments major Corpay projecting a margin expansion from its own productivity efforts using the technology.
Similarly, the technology has seen particular benefits to streamline customer service and cut costs, with digital remittance player Remitly seeing its customer support and operations costs improve as a result of the technology.
These are relatively early examples, and for many payments businesses unlocking the technology’s true productivity potential is set to take some time, but it does at least show that the technology can deliver on its productivity claims in the right settings.
Agentic AI is still drawing hype
While AI-led productivity has begun to yield some financial benefits, there is one area where it remains more nascent and so at higher risk of untested hype: agentic AI. Referring to the use of autonomous agents that can handle entire tasks from beginning to end, this has been a topic of particular interest to payments because those tasks can include the initiation and completion of payments. And while agentic AI can in theory navigate human-focused payment interfaces, many in the industry have seen greater potential in agentic AI-specific payment interfaces and partnerships.
Use of the term agentic has only become widespread in the past year, and many companies have only launched full solutions in the last few quarters, so inevitably discussion is currently more at the potential stage rather than showing results.
However, there are positive steps. Mastercard and Visa both began processing agentic payments on their respective networks in the last quarter. Meanwhile, PayPal has announced a number of moves in the space, most notably its partnership with OpenAI, which will see it begin to allow ChatGPT users to checkout using its platform.
It’s clear that agentic AI has potential, with payment processors across the space, alongside consumer facing brands and banks, discussing the technology more in the last two quarters than in all other quarters combined. However, how much potential remains to be seen, particularly as it ultimately requires consumers to trust such AI to spend their money on their behalf.
A long-term technology for the payments industry?
As one of the most vital parts of the world’s infrastructure, payments is likely to be on the front lines of any AI transformation, and is well placed to seize it. However, as with any technology, AI needs to show genuine impact if it is to gain a lasting place in the industry rather than becoming a passing fad.
The technology is clearly showing potential as a productivity driver, and it is likely that we will continue to see further gains as organizations optimize further. However, it is not yet clear whether the most overzealous projections of the technology will be delivered on.
While 2025 was the year that AI went from excitement to reality, 2026 may ultimately prove to be the year where we truly reckon with how much it can transform our businesses – and where the line between expectation and reality lies.