Two much-discussed payment trends are at the forefront of the emerging payments ecosystem this year. Biometrics are becoming an everyday part of how consumers authenticate themselves, while artificial intelligence is moving from novelty towards a more competitive and established market. Yet one of the most anticipated developments—AI agents that shop and pay on consumers’ behalf—remains largely a promise for the future.
The 2026 Emerging Payments Survey: Agent Adoption Surges, Biometric Authentication Grows report from Javelin Strategy & Research examines how consumers are fueling and responding to these shifts. Drawing on Javelin’s Emerging Payments Survey data, the research takes a long view of technology adoption, recognizing that meaningful behavioral change rarely happens overnight. As Christopher Miller, Lead Analyst of Emerging Payments noted, automation is a thousand-year trend, not a five-year trend.
Growth in Biometrics
Biometric authentication is becoming familiar largely because consumers encounter it everyday. Mobile operating systems use facial recognition and fingerprints to secure devices, turning these technologies into routine parts of the digital experience.
Usage patterns closely follow the devices consumers own. Those with Apple devices generally report using facial recognition, while Android users are more likely to report using fingerprint recognition.
Payments, however, have been slower to adopt biometrics. Nearly half of consumers say they are very likely or likely to use biometric authentication at a merchant terminal. Rather, consumers simply aren’t being given many opportunities to use biometrics when they pay.
“In my own more detailed research about providers of biometrics for point of sale, they are telling me that payment is fading as one of their use cases,” said Miller. “The growth opportunity for biometrics is as part of a broader identification of an individual for a variety of reasons in a shopping scenario.”
“A merchant has an interest in identifying someone for many reasons other than making a payment,” he said. “Who are you? Do we have records about your shopping preferences? Do you have loyalty points? Are there opportunities to upsell you things? The notion of identifying someone has many more vectors than simply authenticating a payment. And what we’re increasingly seeing is that’s how the biometric will be deployed.”
No Mandate for Payments
The shopping experience is being reshaped in ways that could make biometrics a bigger part of the consumer journey. But the notion that shoppers will simply press their thumb on a fingerprint reader to make a payment is unlikely to play out in exactly that way.
“The reality today is that ultimately biometrics is to be more than the payment,” Miller said. “The payment is secondary to the broader mission of recognizing the individual and all of the things that you can do once you recognize the individual.”
Previous changes in the way consumers pay have often been prompted by regulation. The shift from magnetic-stripe cards to chip cards, for instance, was driven largely by regulatory and industry mandates rather than consumer demand for chip technology.
In addition, there is relatively little money to be made from using biometrics to reduce fraud at the point of sale. Transactions in which the physical card or device is present at the terminal are already relatively low risk. If retailers can expect only minimal incremental benefits, there’s little incentive for them to invest in biometric payment technology.
A More Competitive AI Market
The AI landscape is changing just as quickly. A year or two ago, ChatGPT was arguably the only AI tool with broad consumer recognition. Since then, competitors including Perplexity, Claude, Copilot, and Gemini have gained ground in both awareness and usage, while ChatGPT has seen mild decline.
The result is a more mature and competitive landscape, with consumers beginning to distinguish between tools based on what each does best. For some, the deciding factor may be advertising or brand familiarity; for others, it may be the quality of writing, search, reasoning, or another specific capability.
That fragmentation is significant for commerce. As more consumers incorporate AI into everyday tasks, the market becomes large enough for new tools to establish themselves—and for merchants to begin thinking about how they will interact with an expanding ecosystem of AI agents.
“If you are a merchant trying to decide what your forward-looking strategy is for handling agentic purchases, if you say we should do partner with ChatGPT, you have not built a durable strategy,” said Miller. “Your strategy has to recognize that current market dominance is likely to be transient, because none of these products are likely to be the end state for how consumers actually consume agentic capabilities.”
The Outlook for Agentic Purchases
AI is already changing how consumers discover products. Increasingly, consumers are turning to AI tools as a more sophisticated form of search. The next step—having an agent actually select and purchase a product—has yet to become mainstream.
“There is almost no evidence for the use of agents to complete transactions without your intervention,” said Miller. “That is in part because the infrastructure for such transactions largely does not exist in production. To the extent that such transactions have been completed by anyone anywhere, it is generally speaking been in pilot or has been very specific implementations. It is not random consumers having their agent buy them toilet paper or whatever. That stuff just hasn’t come to pass.”
The primary reason consumers haven’t used agentic shopping agents is straightforward: the option isn’t yet widely available. There is therefore little evidence yet about whether consumers will want to use agents to complete purchases because most have not been confronted with that choice.
There are, however, some obvious segments where agentic shopping could play an important role. Business purchasing is a particularly promising opportunity. Companies often need to purchase larger quantities of parts or supplies and may primarily want the best available deal. In such cases, an agent could potentially compare options, negotiate trade-offs, and complete routine purchases with considerably less human involvement.
“Will individuals do this?” Miller said. “In some cases, yes. But you might have thought that the option of having groceries delivered to your home would be a no-brainer for everyone but it’s just not the case.”
“Some people prefer to walk through the grocery store,” he said. “In some cases, it’s because they like to get out of the house. In some cases, they want to touch the avocados themselves. For certain kinds of folks, shopping retains a social meaning that is not tied to the mere provision of goods. That in and of itself is a barrier on all transactions ever being agentic. But it’ll happen.”
