AI Agents Are Splitting The Customer Relationship Into Pieces

On September 8, Meta launched Muse, a free personal AI agent that can open a browser, fill out forms, negotiate, and make purchases on its user’s behalf. Twelve days later, people who sent it shopping on Amazon met a pop-up warning that “continued access by an unauthorized AI agent violates Amazon’s Conditions of Use.” Third-party shopping agents, an Amazon spokesperson said, “should operate openly and respect service provider decisions about whether or not to participate.” (Amazon’s own shopping assistant, meanwhile, will buy items from other retailers’ websites on its users’ behalf.)

In a previous article for Forbes, I argued that AI agents are likely to split many markets between the cheapest offer and the best fit. The next question is who owns the customer, and the Amazon standoff makes that look like a winner-take-all fight. My interviews for this series point to something else, a relationship coming apart into pieces that different players claim. That leaves plenty of room for companies that will never build an agent, including most B2B suppliers.

Discovery Has Separated From the Sale

“The biggest strategic risk isn’t losing the checkout page,” Leonid Myroniuk, a former senior product manager at Visa, told me. “The biggest risk is losing the moment when the customer decides who to trust.”

In March, OpenAI seemed to agree when it backed away from checkout inside ChatGPT, saying the first version of its Instant Checkout “did not offer the level of flexibility that we aspire to provide.” It let merchants use their own checkout and concentrated on product discovery, and by August its new advertising business had reached an annualized revenue run rate of $1 billion.

Four Pieces of the Relationship

The business relationship has distinct pieces, and agentic AI enables distinct contenders to vie for each one: