Despite near-constant industry buzz, the days when artificial intelligence agents dominate e-commerce—and consumers widely complete in-store purchases with a palm swipe—have not yet arrived.

This is not to say they will never arrive, but if the rollout of prior tech trends like biometric authentication and embedded finance is any indication, there is still substantial runway before this financial future becomes reality.

In a recent PaymentsJournal podcast, Javelin Strategy & Research’s Don Apgar, Director of Merchant Payments, and Christopher Miller, Lead Emerging Payments Analyst, cut through the noise surrounding recent payment innovations to assess the true progress of financial trends this year.

What they found is that all these still face challenges. Most notably, an increasingly sophisticated retail landscape only amplifies the questions merchants and financial services firms must answer as they adopt new innovations.

A Road Test for Agentic Commerce

No discussion of trends would be complete without artificial intelligence, and debate about AI’s role in financial services has intensified as models have become increasingly capable. This has led many experts to project the imminent rise of agentic commerce, where AI agents shop and make purchases with limited user direction.

Last year saw a wave of announcements around agentic AI, including new commerce platforms from Visa and Mastercard, as well as a Google-developed agentic protocol intended to serve as a framework for this new shift.

Despite these unveilings, very little true agentic commerce materialized in practice.

“The prediction was that this year we were going to see things live for the first time,”
Miller said. “These products—the ideas, the concepts, and the workflows—were all going to get road tested for the first time. My suggestion was that things might not go as smoothly as all the announcements suggested they would, and, frankly, that turned out to be the case.”

These kinds of false starts are not unusual with new technologies, where it takes time to test edge cases and build the underlying infrastructure. In agentic commerce, that infrastructure would need to cover everything from how consumers input an initial prompt to which AI agent is ultimately authorized to complete a purchase.

While many of these components are now being addressed, significant unanswered questions remain about what the finished system will ultimately look like.

“We’re getting to questions of who will use this and what will they use it for?” Apgar said. “How will we resolve trust issues? How do we resolve authority issues? How do we know that the action mirrors the intent, and the result mirrors the instruction? From a prediction perspective, as much of the buzz that we’ve seen about agentic commerce, 2026 is still going to pan out to be a building year.”

Agentic Search Versus Commerce

While agentic commerce may still be a work in progress, AI has already become firmly rooted in the consumer experience this year, especially as a tool for product discovery and comparison.

“One of the things that AI does well is digest large amounts of data efficiently,” Apgar said. “If you are searching for a bookcase that’s less than 26 inches tall and less than 38 inches wide, I’m sure you’ve gone through web searches where you’re muddling through product pages and you have to find the details of the specifications and you have to drill down to find the measurements—only to back out and do it again on another web page. And there are how many bookcases?”

AI can rapidly narrow search results, often producing answers and recommendations that consumers would not easily find through conventional search methods.

While these tools are a game changer for consumers, they are also changing merchant business models. Instead of relying on search engine optimization to surface in Google results, merchants are now competing to be visible within AI-generated recommendations.

At the same time, as AI increasingly becomes the buffer between merchants and customers, many retailers worry about declining website traffic. This shift could weaken brand identity and, in some cases, reduce businesses to little more than fulfillment engines operating behind AI interfaces.

On the other hand, merchants who do surface prominently in AI-driven discovery stand to reach new audiences and bolster their brand visibility.

These complexities are already beginning to impact merchants, and the sophistication is likely to deepen as agentic commerce evolves.

“If we had this vision that agentic commerce was a single-provider solution that a consumer might use from end-to-end and somehow it would just layer over the existing framework of e-commerce, that’s proven to be false,” Miller said. “Just layering OpenAI on top of the internet as it exists is not going to work for anybody.”

“In a sense, the internet—and more precisely the e-commerce version of the internet—will have to be reengineered for everybody’s benefit, in enabling things like software agents to do any of this work,” he said. “That’s where the building is going to be, it’s in that infrastructure layer.”

The Path to Biometric Authentication

A trend that appeared closer to mainstream adoption this year was biometric authentication at the point of sale. The benefits are well established, including stronger security and reduced friction at checkout. Unlike agentic commerce, biometric technologies have existed for years and have been piloted globally across a range of use cases.

Given this, it might have been expected that this year would mark a clear inflection point in adoption. So far, however, progress has been limited to continued trials, including the launch of additional Biometric-Authentication-as-a-Service platforms that integrate biometrics into existing payments stacks.

There has also been movement toward cross-experience, unified identity solutions. In many cases, when customers create a biometric profile with a company, their in-store purchase and loyalty data remain disconnected from their online profiles. Cross-experience identity solutions can connect these dots.

Still, these platforms are far from widespread adoption, which appears to reflect the current state of the biometric authentication market this year.

“I suggested that new products would come to market and we’d start to see some more launches, but I will say that it’s been a little bit light in terms of news on that front,” Miller said. “There is a path to market, but that doesn’t mean that any merchants have said, ‘We’re going to turn that on,’ and it doesn’t mean that the capability is ready to light up today.”

“We might be a little slower than what I thought, but we continue to see development in the marketplace, the creating of the business plans and of the go-to markets so that these products and capabilities are going to be available to be chosen,” he said. “That wasn’t true two years ago in a widespread way, so that’s a significant advance, even as we continue to wait on its arrival.”

The Boiling Embedded Finance Pot

There are notable parallels between the gradual rollout of biometric authentication and the evolution of embedded payments and finance. One of key challenges in embedded finance, however, is that banks and fintechs are often operating at cross-purposes.

Many fintechs have developed strong vertical Software-as-a-Service (SaaS) platforms that address a wide range of merchant needs, but these systems don’t always balance ease of use with financial services expertise.

For example, some fintechs may present a seasonal merchant with an interest-bearing deposit offer during the offseason, when cash flow is tight. Conversely, they may extend credit during peak season, when liquidity is already strong.

Financial institutions with deep experience in these products often struggle to integrate with newer merchant platforms. They may offer a SaaS-based point-of-sale system but lack the capability to fully leverage the data these platforms generate.

“The pot is still boiling, with the fintechs struggling to figure out banking and the banks struggling to figure out data,” Apgar said. “Everybody thought based on how fast the market was moving and the many partnership announcements that this would be a lot further along, and that one or two companies would have come out on top and stick the flag in the top of the mountain that says, ‘We̵...