“Agentic AI is going to unlock a lot more commerce,” Rob Cameron, global head of Visa Acceptance Solutions, told PYMNTS CEO Karen Webster.
That vision is already underway in what has become known as the prompt economy. Consumers are beginning to act as their own personalization engines. Rather than scrolling through endless product listings, they will opt to engage in curated conversations where prompts replace clicks and decisions unfold in real time.
The shift is not only a change in how people shop. It represents a fundamental re-wiring of the ecosystems that connect consumers, processors and merchants, one that Cameron said Visa is working to secure and scale.
“It’s all moving very quickly, and faster than many consumer technologies have ever evolved,” Webster told Cameron.
The transition has not been uniform or without friction, but it’s still early days.
Advertisement: Scroll to Continue
“Sometimes, based on my prompt, I get exactly what I’m looking for, and other times I don’t,” Cameron said.
He attributed the inconsistency to how prompts are structured and to the lack of embedded history in most agents today.
“Usually, I start from scratch with every search and every prompt, and that makes it a little harder for the agent to really return what I’m looking for on the first run,” he said.
Additionally, not every category is equally prepared for agentic commerce. Cameron pointed to travel as “a terrific use case” for agentic AI, although one that still requires some consumer involvement at the end.
“It’s a high-dollar expense and something that I’d want to ‘see’ before I click ‘buy’… but I think that the agent can do a lot of the planning to make life so much easier,” he said.
Reconstructing Commerce for Agents
In this new environment, virtual storefronts will not disappear overnight but will evolve alongside AI, Cameron said. Agents will continue to use existing storefronts for a time, but over the long term, commerce will be reconstructed as agents handle more of the decision-making. That means merchants will need to think differently about how consumers discover products.
Visa believes secure data tokens can play an important role in that process, Cameron said. Tokens can carry privacy-preserving signals — derived from shopping history and preferences with consumer consent — while maintaining anonymity, allowing agents to deliver results that reflect consumer intent.
Many consumers are no longer interested in an “endless aisle” of search, Webster said. Instead, they want curation. That preference sets the stage for agents to provide trusted, personalized itineraries and recommendations.
Visa’s Role: Building the Foundation
Visa is positioning itself at the center of this shift, creating the underlying infrastructure that ensures agentic commerce rests on secure, flexible and trusted rails. One step is allowing merchants and acquirers to test and learn, Cameron said.
“Almost every larger merchant or acquirer that I speak with is [asking], ‘How do I participate in a pilot? How do I get some experience?’” he said. “That’s why we’re making our MCP server available. We’ve got [our] Agent Toolkit on the Acceptance Platform currently in pilot.”
Another foundational element is tokenization, which Visa first launched more than a decade ago. Tokens give consumers control over how much of their data they choose to share, Cameron said.
“One of the key tenets of Visa and our tokenization … is for [the consumer] to have the control on what [data] … they … share,” he said. “I might want to stay as a guest and not share very much other than to make a payment with you, this merchant, but I might be willing to expose my purchase history and other things with this merchant or this agent with whom I’ve got a more valued relationship … so that it can make more informed decisions. This will all be consent-based.”
In an era of agentic AI, flexibility is critical to preserving consumer trust, Cameron said.
Visa is also rethinking credentials themselves, he said. Visa’s Flexible Credential allows a single payment identity to shift between debit and credit depending on context. Consumers can, for example, set preferences to use BNPL for purchases over $100 and debit for smaller ones.
This evolution could eventually lead to a single smart credential that dynamically optimizes how and when consumers pay, Webster said.
These initiatives are not only about payments at the point of sale. They are also about embedding intelligence into the infrastructure of commerce. Visa is working to capture intent, amount and location up front, feeding that data into the agent-merchant interaction so that purchases are secure, authenticated and protected from the start, Cameron said. This intent-driven model preserves the trust consumers already associate with their Visa credentials while protecting merchants with more reliable signals.
Preserving Trust in a New Ecosystem
“We’re trying to set up a framework based on trust … to set up the opportunity for that exchange of payment in exchange for a good or service,” Cameron said.
Visa’s post-purchase and dispute protections help cement the ecosystem between merchants and consumers, he said. Agentic AI could even streamline dispute resolution, turning processes that today are largely manual into ones that flow more smoothly over time.
Looking ahead, Cameron said Visa’s priority is to accelerate pilots and expand partnerships while making sure that global brands and small businesses can participate. With data tokens, flexible credentials and secure rails, Visa is positioning itself to power agent-driven commerce at scale and ensure that the foundation of this new economy is built on trust.
Source: https://www.pymnts.com/