Agentic Commerce: Revenue, Risk, and Complexity Across the Payments Stack

Agentic commerce will be great for consumers.

For issuers and acquirers, it’s about to introduce a lot of complexity. And that complexity is where new revenue pools tend to show up.

I already use GenAI for shopping discovery on a near-daily basis, and I honestly look forward to the days of full agentic commerce.

From a consumer perspective, the value proposition is clear.
If an AI can handle reorders, subscriptions, rebooking, or “buy this when it meets my rules,” that’s real convenience. Less friction. Less thinking. More time back.

For merchants, it feels like a mixed bag. Agentic commerce changes discoverability. Instead of competing for human attention, merchants get selected by agents based on signals like total price, delivery certainty, return clarity, and trust.

That’s a real upside for merchants with clean data and strong operations. But it also compresses differentiation and reduces the influence of UX, bundling, and checkout persuasion.

For issuers, there’s upside, but it’s not free. Cleaner intent and more constrained, delegated transactions could mean fewer false declines and more “always-on” spend.

At the same time, it creates new requirements: new controls, token orchestration, fraud models that don’t freak out over non-human behavior, and entirely new dispute conversations that start with “my agent did it.”

For acquirers, this might quietly become a new revenue line. Most merchants won’t want to figure agentic commerce out on their own. Agent-ready acceptance, verification and trust signals, policy controls, dispute audit trails, fraud tooling that understands agent behavior… this starts to feel like a new category of value-added services, not just an extension of what exists today.

And then there are the networks. Even if authorization requirements don’t change overnight, agentic commerce likely drives other changes: new data expectations, richer context, delegated controls, and trust signals layered into existing rails.

Over time, it’s hard not to imagine that turning into new network requirements and, eventually, new network (scheme) fees tied to tokenization, verification, or enhanced risk and data services.

Still early. Still a lot of unknowns.

But it feels like one of those shifts where the consumer benefit is obvious, and the rest of the ecosystem has to re-learn where the money actually gets made.

Previous
Previous

In BaaS, Partner Operations Become Issuer-Level Network Risk

Next
Next

Card Network (Scheme) Fee Analysis: How Issuers and Acquirers Can Finally See What They’re Paying