AI Takes the First Step in Shopping While Consumers Keep the Buy Button

Consumers who followed artificial intelligence to a new retailer spent considerably more on AI-assisted purchases than the average AI retail researcher who made a purchase.

    Get the Full Story

    Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

    Subscribe to our daily newsletter, PYMNTS Today.

    By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions.

    PYMNTS Intelligence found that consumers who tried a new retailer after using AI spent an average of $1,430 on AI-assisted purchases over three months, compared with $931 among AI retail researchers who made purchases. Consumers who made an unplanned purchase averaged $1,564.

    The spending figures give merchants a reason to look beyond the sheer number of consumers using AI for product research. Artificial intelligence is changing the choices those consumers make. Forty-three percent found a better price, 27% chose a different brand and 26% bought a different product, according to the PYMNTS Intelligence report “The 50 Million Consumer Migration: The Data Behind Retail’s Shift Toward AI Discovery.”

    For retailers, the opportunity is customer acquisition. A shopper who might have gone directly to a familiar marketplace or merchant can now ask AI to compare the available choices first. The retailer that wins that comparison can get access to a customer who wasn’t looking for it.

    Our data show that AI recommendations introduced 10.8 million consumers to a retailer or website they hadn’t used before and moved 15.9 million toward a different brand. The broader research found that 49.6 million U.S. adults now begin retail product research with AI, including 39 million who have stopped relying primarily on the traditional search channel where they once started.

    That puts pressure on merchants to make sure AI can accurately evaluate what they sell. Price and product specifications matter, but so do availability, shipping, warranties and return policies. Those details can determine whether a merchant makes an AI-generated shortlist before the consumer ever visits its site.

    AI consumers graphic

    Checkout Changes the Equation

    Getting onto the shortlist doesn’t require the consumer to give AI access to money. Completing the transaction does.

    As noted here, 48% of online shoppers used AI to research their most recent purchase. Fifty-six percent would allow an AI agent to search and compare products, but only 37% would allow one to authorize payments and 35% would give one access to saved payment methods.

    Consumers aren’t rejecting AI checkout altogether. We found that 49% of consumers interested in agentic AI would delegate both routine purchases and larger, research-driven purchases to an assistant. Among consumers already using dedicated artificial intelligence platforms, 58% prefer checkout to take place within the AI environment. The same percentage applies to consumers using AI embedded in an app or service.

    Who stands behind the transaction matters, too. A PYMNTS Intelligence report, “From Assistive to Agentic AI: Consumers Wade Into Autonomous Commerce,” found banks and digital wallets among consumers’ most trusted providers of agentic assistants, with card networks also competitive.

    Visa CEO Ryan McInerney described the gap between shopping and payment Tuesday (Sept. 8) at the Goldman Sachs Communacopia + Technology Conference.

    “We are seeing adoption for shopping, but not yet for autonomous payments,” McInerney said, according to PYMNTS’ coverage of his remarks.

    “The barrier to that, if I had to describe it in one word, would be trust,” he said.

    McInerney said three-quarters of consumers surveyed by Visa didn’t trust agentic platforms to make autonomous payments with their money and financial information. But 61% said they would trust an AI agent to make payments if Visa were involved. The share topped 70% among consumers who use large language models at least weekly.

    Merchants are putting limits on AI as well. Forty-six percent said pricing and the final price paid are the functions they are least willing to let agents handle. Thirty-one percent plan to invest within a year in automated product search and comparison, versus 26% for AI completing checkout on a consumer’s behalf.

    That helps explain where payments companies are concentrating their work. PYMNTS reported Wednesday (Sept. 9) that Mastercard introduced Agent Connect as part of an expansion of its Agent Suite for Merchants. The service is designed to connect merchants, AI agents, digital platforms and payment providers while supporting user-authorized transactions. Mastercard also said merchants retain control over how their products and services are represented.

    The $1,430 spending figure puts a value on solving the remaining payment problems. AI is already bringing consumers to retailers they hadn’t used before, and those shoppers spent considerably more on AI-assisted purchases than the average AI retail researcher who made purchases.

    Checkout adds requirements that discovery doesn’t. The merchant has to know which agent is presenting the transaction. The payments provider has to know that the consumer authorized it. Consumers need to be able to decide how much authority the agent gets, including when it can act without another approval.

    For merchants, there is money on the other side of those controls. AI has already shown it can change the seller, brand, product and price a consumer chooses, but what remains to be seen is how much of the transaction the consumer will allow it to control.