AI Shifts Product Discovery, not Loyalty

The idea of bots buying on behalf of shoppers worries some ecommerce store owners, not because they fear AI, but because it shifts the customer relationship into one they cannot see, understand, or influence.

Agentic commerce systems can interpret buying intent, evaluate alternatives, recommend products, assemble carts, and potentially complete purchases.

Merchants still own inventory and fulfillment, but a portion of the relationship shifts to an intermediary AI. Rather than converting shoppers directly onsite, sellers may persuade the system, which, in turn, makes recommendations to or purchases for the customer.

Universal Cart

Consider, for example, Google’s proposed Universal Cart, announced at I/O 2026. The cart would allow shoppers to compare and purchase products from many retailers across Google Search and eventually even YouTube and Gmail. Google would “own” the cart throughout the product discovery, comparison, and checkout.

Merchants are the sellers, but Google becomes much more involved in purchase decisions. That possibility raises a difficult question for some independent retailers.

“The problem with agentic ordering is that it turns all products into commodities,” wrote Joel Moskowitz of Tools for Working Wood in an email. “The retailer has absolutely no chance to sell, upsell, or encourage browsing. If a bot is doing the buying, that’s certainly the case. If there’s a universal cart, it has basically the same effect.”

Tools for Working Wood

Familiar Concern

Yet ecommerce sellers have seen this pattern before.

One could argue that every major change in ecommerce product discovery has altered where customer relationships begin and who controls them.

Search engines and marketplaces drove discoverability. Social platforms offered access to a shared audience but changed what content worked and who saw it.

Each is an overlapping product distribution and attention channel.

• Search engines. Shopper relationships begin with search engines. Stores that rank prominently on a results page can market, merchandise, and influence shoppers. AI Overviews have made this more difficult.

• Marketplaces. Shopper relationships begin and end on marketplaces. Amazon customers are Amazon’s. Third-party sellers pay fees to access those customers. Price competition can pinch margins.

• Social media. Relationships begin on social platforms. Merchants who please the algorithm can get clicks leading to sales. But social media sites don’t like outbound clicks.

• Agentic commerce. Relationships begin and end with AI agents. Influence the AI, and merchants can make a sale.

Search still matters. Marketplaces remain dominant. Social spurs discovery. But each changes access to customers.

Agentic commerce may be the next version of that concern. If search rewards relevance and marketplaces participation, AI shopping may reward whatever signals its systems value.

Adaptation

Thus the challenge of direct customer relationships is familiar. The key for merchants is adapting.

Sellers must learn anew how to market and where to focus. The change will likely create ecommerce winners and losers.

“Historically, we got our customers from organic search. That was fueled by having a decent website, good word of mouth, good service, recommendations and reviews, and occasional bits of magazine and other media coverage,” wrote Moskowitz as he described his company’s early success.

Now, Moskowitz added, “organic search is dying…and while AI bots do occasionally recommend us for one thing or another, if agentic commerce becomes a thing, we will be left out.”

This is a legitimate concern. The evolution of distribution channels drives a commensurate evolution in skill sets and tactics to remain competitive.

Ecommerce businesses have long thrived by mastering channel optimization: search, marketplaces, and social. Success comes from the competence of differentiating the business and its customer relationships.

Recommended

A Preview of Agentic Marketplaces

May 17, 2026

Differentiation

Moskowitz noted that he’s not planning to chase every new AI platform. “Our approach is to manufacture a lot of our own products. More and more of our effort is spent on niche, unique items.”

That approach suggests a broader lesson. If AI systems influence product discovery, merchants could spend less time gaming the algorithms and more time creating reasons for shoppers to seek them out.

Strong products, memorable brands, useful content, direct relationships, and distinctive shopping experiences are always difficult to replicate.

Agentic commerce may change where shoppers begin, but not why they choose one merchant over another.