AI shopping agents poised to transform digital payments

Agentic commerce, where autonomous digital agents initiate and complete transactions on behalf of consumers, is rapidly moving from concept to commercial reality. As large language models (LLMs) evolve from advisory tools to active decision-makers in online purchasing, the payments industry is preparing for one of its most significant structural shifts in a decade.

Industry analysts expect agent-driven commerce to reach meaningful scale by the early 2030s, with global revenue projections nearing $5 trillion. Unlike previous disruptions that introduced new rails or alternative payment methods, the next wave focuses on intelligence: adding machine-driven orchestration to existing financial infrastructure.

Attila Doğan, Chief Product Officer at PPRO, said, “A trend that will continue to grow in 2026 is new agentic protocols layered into existing financial rails. Agentic commerce is not creating a new payment rail; it is adding intelligence on top of the rails and methods we already have. Customers will shop through large language models (LLMs) that search, compare and purchase on their behalf. This creates a new commerce channel where the buyer is an autonomous agent rather than a human at a checkout page. Agentic commerce is bound to become a major force, with global estimates reaching 5 trillion in revenue by 2030.”

In this model, consumers delegate shopping tasks to LLMs capable of searching, comparing and purchasing autonomously. What changes is not the payment instrument itself, but the buyer, shifting from a human interacting with a checkout page to an AI agent executing purchase intent on behalf of its user.

To make this transition viable at scale, the industry will need new technical infrastructure. One emerging component is the merchant commerce protocol (MCP) server. These structured endpoints enable machines to communicate purchase intent in a form that merchants can understand, while coordinating the downstream payment processes. According to Doğan, MCP servers will be central to selecting the optimal method, route, credential format and risk posture for each transaction, effectively creating an “intelligence fabric” above existing rails.

For merchants, this means preparing acceptance stacks that can reliably support both cards and local payment methods within autonomous agent flows. For payment providers, the shift may require rethinking routing logic, tokenisation handling, authentication processes and fraud models to accommodate transactions where the initiating party is not a human user.

As autonomous agents begin handling higher-value, high-frequency purchases ranging from travel bookings to recurring household spending the competitive advantage will shift toward entities capable of enabling seamless machine-to-merchant commerce.

Agentic commerce is still in its early stages, but the strategic intent is clear: the next era of digital payments will be defined not by new infrastructure, but by new intelligence layered on top of the infrastructure already in place.