Major payment networks and technology firms are moving quickly to lay the groundwork for what they see as the next stage of global shopping, where artificial intelligence agents search, compare and complete purchases for consumers.

The emerging model, known as “agentic commerce,” reflects how shoppers increasingly rely on chatbots to handle daily tasks such as finding products or hunting for deals online. Until recently, those tools stopped short of processing payments, forcing users to leave chat interfaces to finalize transactions.

That limitation is now beginning to fade. Visa and Mastercard say they have spent the past year building systems and partnerships that allow AI agents to securely complete purchases inside chat platforms. Early pilots are already underway, and payment executives expect the technology to reach consumers in 2026.

Payment leaders told CNBC the shift could prove even more disruptive than the rise of e-commerce marketplaces like Amazon.

“A big shift in commerce happened when payments moved from a mostly brick-and-mortar world to an e-commerce world,” said Sandeep Malhotra, Mastercard’s EVP for Core Payments in Asia Pacific. “Now, we are seeing the next shift, which is moving from the e-commerce world to an agentic commerce world. We have gone from cash to digital, now we’re going from digital to intelligent.”

The concept generally refers to AI systems that act on behalf of users by finding products, comparing prices and completing payments directly within a chatbot. Instead of browsing multiple websites or apps, shoppers could receive curated options based on precise requests.

One early use case could be travel bookings. A consumer might ask an AI agent to find the cheapest red-eye flight from Singapore to Tokyo under $500 with no stops. The system would scan options, confirm details, book the ticket and pay using stored credentials without leaving the chat.

Malhotra said the technology may also allow users to authorize agents to make purchases while they are offline, such as automatically buying an item once its price falls below a set level.

Visa and Mastercard have already rolled out initial frameworks designed to secure transactions initiated by bots. Both companies have completed pilot programs with selected merchants and users.

T.R. Ramachandran, Visa’s APAC Head of Products and Solutions, said commercial deployment of personalized and secure agent-driven payments could begin as early as the first quarter of 2026.

With more than half of Visa’s total volume already coming from e-commerce, and rising consumer demand for AI shopping assistance, the conditions are favorable, Ramachandran said. A Visa survey conducted in December found that nearly half of U.S. shoppers now use AI to improve their shopping experience. Separate data from Adobe showed AI-driven traffic to U.S. retail sites surged by 4,700% in July compared with a year earlier.

Agentic commerce transactions are expected to take place across widely used AI platforms such as ChatGPT and Google’s Gemini, as well as through merchant, bank and app-based agents. Payment firms have also been working closely with AI companies including OpenAI to prepare for the transition.

OpenAI introduced a “Buy it in ChatGPT” feature in September that enabled instant checkouts. In November, Perplexity partnered with PayPal and launched a free agentic shopping product for U.S. users.

Large retailers are also experimenting on their own. Amazon began testing its “Buy For Me” tool earlier this year, while taking steps to limit external AI agents from accessing its website, amid concerns about pricing pressure and loss of direct customer relationships.

Despite growing momentum, agentic commerce raises unresolved issues around security, accountability and liability. One focus area has been the development of so-called agentic tokens, which use cryptographic authentication to verify authorized AI agents and separate them from malicious bots.

Visa launched its “Trusted Agent Protocol” in October with Cloudflare, creating cryptographically authenticated records for bot-initiated payments. Ramachandran said Visa also plans to introduce additional “payment signals” for banks to improve transaction visibility and strengthen authentication using behavioral intelligence.

Liability remains another concern when AI agents make mistakes, such as ordering the wrong product or booking incorrect travel dates. Traditional payment disputes typically involve four parties, but AI introduces a new participant.

“Now there is a fifth player in the value chain — AI platforms who have become inserted into the value chain because the customers want them there,” Ramachandran said. “You almost have to assume mistakes will happen and create guardrails and protection around that.”

Supporters argue the benefits outweigh the risks. They say agentic commerce will save time, reduce search costs and improve access to information and deals.

“I can only see benefits with agentic commerce. Consumers will have better access to information, better access to goods, better access to services and better experiences,” Malhotra said.

Merchants, however, may need to adapt quickly as AI-driven price discovery becomes more common.

“When price discovery and shopping ubiquity become the norm rather than the exception, it will be fascinating to see how companies adapt,” Ramachandran said.

Payment executives expect merchants to verify AI agents, develop their own bots, expand loyalty programs and rethink upselling strategies. While the exact pace of adoption remains uncertain, industry leaders agree the shift is inevitable.

“As to exactly when it will scale, that’s less clear,” Ramachandran said. “But based on our experience and overall [large language model] platform adoption, we’re likely talking months rather than years.”