Instead of retailers and marketplaces controlling discovery and checkout through branded storefronts, AI agents are beginning to sit between buyers and sellers, deciding what gets bought, where and at what price. The move matters because payments networks and retail platforms are now racing to embed themselves inside that agent layer, signaling a broader reordering of who controls demand and transaction flow.
Price.com Shows How
Price.com’s Buy with AI pushes AI beyond recommendations and into execution. According to the release, the agent accepts user intent, scans retailers across the open web, compares pricing and availability, applies savings and completes the transaction. The consumer does not manually browse or check out. The AI handles the process end to end. That design turns Price.com from a comparison site into a transaction-executing intermediary, one that can influence purchasing decisions before shoppers ever encounter a merchant’s site.
That shift aligns with a broader pattern emerging across commerce. CNBC reported that agentic shopping tools are bundling price comparison, discount application and automated checkout into a single flow, allowing AI agents to surface lower prices and complete transactions with minimal consumer input. The result is a buying experience where storefronts matter less than the agent making the decision.
PYMNTS Intelligence data shows that consumers are becoming comfortable with that tradeoff. About 30% of consumers say they are willing to let AI assistants handle everyday planning or shopping tasks, and in an earlier PYMNTS survey 32% said that they have used or would consider using generative AI for shopping. Convenience and savings drive that willingness, especially in categories where consumers already compare prices aggressively.
Price.com’s launch demonstrates what agentic commerce looks like in practice. The AI does not simply guide users. It acts for them. That distinction forces a change in how retailers, platforms and payment providers think about influence. When AI controls the buying moment, whoever controls the agent controls demand.
Visa, Amazon Expand Agent Layer
Payments networks and retail platforms are responding quickly. Visa unveiled “Find and Buy with AI,” a set of capabilities designed to let conversational AI systems discover products and complete purchases using Visa’s global payments infrastructure. The initiative positions Visa inside AI-driven checkout flows rather than at the end of traditional merchant funnels. As AI agents initiate transactions, Visa is working to ensure authorization, fraud controls and settlement still run through its rails. As reported by PYMNTS, the company also introduced Trusted Agent Protocol that allows exchange of information between agent and the merchant.
Amazon is testing a more controlled version of agentic shopping inside its ecosystem. The company recently expanded AI-driven shopping features within its app, including agent-assisted experiences that help users discover products and, in limited cases, complete purchases from participating brand websites without leaving Amazon’s interface.
That pressure explains why traditional retailers are investing behind the scenes. Tesco recently signed a three-year generative AI agreement with Mistral AI to improve data analysis, internal workflows and customer experience. While Tesco did not launch a consumer-facing shopping agent, the investment strengthens the systems that agentic commerce depends on clean data, personalization logic and real-time decision support.
As PYMNTS CEO Karen Webster mentioned the protocol power struggle shaping AI-driven commerce, the next phase of competition will center on who sets the rules that govern how AI agents transact. As checkout shifts from storefronts to software, protocols increasingly determine which merchants get surfaced, how payments are authorized and where value accrues.
