AI Agents Aim to Change Shopping. Some Retailers are Locking the Doors.

Dow Jones
Yesterday

Tech executives are selling their vision of a retail industry transformed by "agentic commerce," wherein artificial-intelligence agents shop and can even make purchases with no direct human involvement.

Meta's release of the Muse agent, capable of autonomous shopping, only supercharged the anticipation. Retailers including Gap, Walmart and Best Buy have publicized partnerships with Muse, while Shopify said Muse will be able to buy from every merchant on its platform.

But retail brands aren't entirely sold, with some rushing to accommodate third-party agents and others deliberately locking them out.

Retailers with hesitations cite factors including the desire to control consumer data and concern over the customer experience. There is also no evidence yet that significant numbers of people are telling AI agents to buy things for them.

Shopping networks QVC and HSN treat agentic commerce as a channel for marketing and product discovery, allowing shopping agents to complete purchases while proactively sending their catalogs to companies like OpenAI and Google for visibility, according to David Thompson, senior vice president of product and technology strategy at parent company QVC Group.

"We need to be out in the places that already are grabbing her attention," he said, referring to the company's typical shopper. Thompson said QVC can't confirm whether agents have actually completed any purchases without conducting extensive traffic analysis. Some agents might also be inadvertently blocked by the company's legacy security systems designed to defend against bad actors, he said.

Gap decides which third-party agents can browse and which can check out on a case-by-case basis, said Damon Berger, senior vice president of marketing shared services.

"We're trying to be open yet continue to have the trust of our consumers," Berger said.

Retail conglomerate Tapestry, parent company of Coach and Kate Spade, allows agents to browse its brands' sites but doesn't let them complete purchases, largely because of concerns over fraudulent purchases and "inventory hoarding," where bots fill carts without buying.

However, the company is optimizing its systems to allow agents at checkout, so it will be ready if it changes its strategy.

"One day, if you're like, 'You know what, I will have my own agents who will do my shopping for me,' we want that journey to be smooth for you," said Yang Lu, Tapestry's chief information and digital officer.

The department store chain Kohl's prevents agentic purchases, a spokeswoman said, declining to elaborate. Chewy and Neiman Marcus each block certain agents, according to Billions Network, a firm that helps companies verify humans and AI agents. Neither company responded to requests for comment.

Others are more restrictive. Amazon blocked Meta's Muse, saying the tool failed to identify itself as AI and raised data-privacy and security concerns, including the capture and storage of data regarding individual Amazon users' accounts.

The company also blocks agents from OpenAI, Google, Anthropic and Perplexity, according to research from Billions Network.

Muse is designed to protect users by generating a one-time card number for each transaction and asking for a user's permission before buying anything, said a Meta spokesman.

EBay in January barred agents, including those designed to shop, from accessing or taking action on its platform without express permission. The company believes some AI can't make the nuanced evaluations necessary for shopping handmade and preowned goods, according to a person familiar with the matter.

Some retailers that don't welcome agents still see agentic shopping strategies as "no-regret investments," putting them in position to capitalize if more consumers adopt the technology, according to Isabel Perry, chief strategy officer at Dept, a digital marketing firm.

Muse's ability to theoretically shop anywhere with user permission means retailers are out of time to start making decisions, Perry said.

It is unclear how eager consumers are for agents to shop on their behalf.

Only 3% of U.S. adults would trust agents to complete purchases, according to a survey by payments firm NMI. One-third of active AI users told marketing firm VML they would never give agents the power to buy or even reorder products.

Despite being "optimistic" about the technology, Amazon CEO Andy Jassy told investors last year that "the customer experience is not good," citing AI's tendency to bungle prices and other basic product information. Amazon released Buy for Me, a semiautonomous agentic shopping tool, last year.

Investors seem convinced that agentic commerce is about to take off. A jump in Meta's share price following the Muse rollout coincided with hits to companies in the banking, insurance and travel sectors, which investors theorized might be disrupted.

But the bulls might be getting ahead of themselves, according to Sky Canaves, principal analyst at research firm Emarketer.

"There is this view in tech circles that consumers don't like shopping," Canaves said. "But we actually do like going to the grocery store. Consumers still like to make their decisions themselves and be in charge, especially where money is concerned."

 

At the request of the copyright holder, you need to log in to view this content

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10