If AI Chooses What You Buy, Whose Customer Are You?
Retailers want AI to send them customers. They just don't want AI to own the relationship.
Imagine you're buying a new TV.
Today, you might search Google, visit Amazon or Walmart, compare prices, read reviews, and choose one yourself.
But that process could get much shorter.
You could simply tell an AI:
“Find me a 65-inch TV under $1,500 with great picture quality for movies.”
The AI could search products, compare specifications, check reviews, and narrow the options down for you.
Eventually, with your permission, an AI agent could even complete more of the purchase process.
Now imagine Samsung made the TV and Walmart sold it — but the AI decided which one you should consider.
That raises a surprisingly complicated question:
Whose customer are you?
Samsung's?
Walmart's?
Or the AI company's?
A Customer Is Worth More Than One Sale
Suppose Walmart sells you $100 worth of products.
The value isn't limited to that $100 transaction.
If you keep shopping there, Walmart can learn more about your behavior.
What do you search for?
What do you actually buy?
Which brands do you return to?
How much are you willing to spend?
When are you likely to shop again?
That knowledge can help a retailer recommend products, build loyalty, and sell advertising.
And advertising has become a serious business for retailers.
Walmart's advertising business, Walmart Connect, grew 44% in its most recently reported quarter as of August 2026.
Reuters reported that analysts estimate its gross margin at around 70%, and that the business contributes roughly one-third of Walmart's operating profit.
In other words, knowing who is likely to buy something can itself be extremely valuable.
Retailers aren't just selling products.
Increasingly, they are also monetizing their understanding of shoppers.
What If the AI Knows the Customer First?
Now put an AI between the shopper and the store.
The old path looks roughly like this:
Customer → Retailer → Product
A future shopping journey could increasingly look like this:
Customer → AI → Retailer → Product
At first, that may not seem like a dramatic change.
The retailer still makes the sale.
But look at the information flowing through the transaction.
If you buy a box of cereal from Walmart, the retailer knows what you bought.
But imagine that before the purchase, you told your AI:
“It's for my kid's breakfast.”
“I don't want anything too sugary.”
“Keep it inexpensive.”
“He didn't like the one we bought last time.”
The AI has now been exposed to something potentially more valuable than the final purchase.
It knows something about why you're buying.
It may know what you rejected.
What you're comparing.
What matters to you.
And what you're willing to pay.
That doesn't mean today's AI shopping services automatically own or monetize all of this information. What data is stored, shared, or passed to a retailer depends on the platform and transaction.
But it explains why the relationship is becoming so important.
Retailers Want AI Shoppers
Retailers aren't trying to stop AI shopping.
Quite the opposite.
AI-referred shoppers may be valuable customers.
Adobe Analytics found that in May 2026, U.S. retail visitors referred by generative AI services produced 53% more revenue per visit than visitors from other traffic sources, according to Reuters.
That doesn't prove AI itself makes people spend more. People using AI for product research may simply arrive further along in the buying process.
Either way, retailers have a reason to want that traffic.
Companies including Walmart, Ulta Beauty, and Wayfair have been working to make their products easier for AI services to discover.
AI could become another major entrance to online shopping.
Just as Google once did.
But They Don't Want to Lose the Customer
This is where the tension begins.
Retailers want AI to send shoppers their way.
They don't necessarily want the entire relationship to happen inside the AI.
Imagine a future in which you search through an AI, receive a recommendation from the AI, pay through the AI, and return to the AI the next time you need something.
You may barely interact with the retailer itself.
The store still gets the sale.
But it could lose part of the relationship.
It may know less about what you considered before buying, why you chose one product over another, what you might want next, or how to bring you back directly.
Payment company Adyen highlighted this risk in 2026.
As AI systems move from recommending products toward choosing merchants and facilitating transactions, merchants risk becoming less visible to their own customers.
One merchant discussed by Adyen generated roughly 70% of its transaction volume through direct channels and wanted to protect that relationship.
The concern is simple:
AI can bring you a customer.
But if that customer always returns to the AI, whose customer is it really?
We've Seen This Fight Before
The internet has repeatedly created new middlemen between businesses and customers.
Brands wanted people to visit their websites.
Then Google became a major gateway to those websites.
Retailers wanted shoppers to come directly to their stores.
Then Amazon became a place where millions of people discovered, compared, and bought products.
Instagram, TikTok, and YouTube created another layer by influencing what people discover before they ever reach a store.
AI could become the next layer.
But there is an important difference.
Google largely helps you find information.
Amazon gives you a place to compare and buy products.
An AI agent could increasingly help decide:
“This is probably the best product for you.”
And, with permission, it could eventually take more actions after making that decision.
That puts AI in a potentially powerful position between intent and purchase.
AI Isn't Doing All Our Shopping Yet
There is an important reality check.
In 2026, most people are not handing their entire shopping lives over to autonomous AI agents.
Forrester says much of today's so-called agentic commerce is still closer to conversational product discovery and recommendation than fully autonomous purchasing.
Humans usually remain involved in the final decision and payment.
Some attempts to move checkout directly into AI interfaces have also shown how difficult changing shopping behavior can be.
So saying “AI already controls shopping” would be an exaggeration.
But the direction is worth watching.
A 2026 global study from Visa Acceptance Solutions found that 47% of consumers surveyed had used AI during their most recent purchase journey, while roughly one in three said they would be comfortable allowing an AI agent to complete a purchase for them.
Discovery is changing before purchasing fully does.
And that may be enough to start changing who controls the customer relationship.
The Next Valuable Shelf May Not Be a Shelf
For much of the past two decades, one of the most valuable places for a business to appear was near the top of Google.
On Amazon, appearing prominently in search results could also make an enormous difference.
AI shopping could create a new kind of shelf.
Except this shelf may contain only a few products.
Instead of showing you 100 results, an AI might say:
“These are the three I'd consider.”
For brands, that creates a new competition.
Not simply to rank on a search page.
But to become one of the products an AI considers worth recommending.
That's where this question connects to a larger shift in commerce.
If companies begin competing to enter an AI's shortlist, the next question is unavoidable:
Who owns the relationship with the shopper who trusts that shortlist?
The Answer Hasn't Been Decided Yet
It would be premature to say AI companies will simply take over the customer relationship.
Retailers aren't standing still.
New commerce infrastructure is being designed to let AI systems work with merchant catalogs, carts, accounts, payments, and loyalty programs while still involving the retailer.
The outcome could take many forms.
AI companies may control discovery while retailers control checkout.
Retailers may retain customer accounts while AI becomes the preferred interface.
Or some AI platforms may eventually manage much more of the journey.
That battle is only beginning.
And it means the future of online shopping may not be decided simply by who sells the product.
It may also depend on:
Who receives the customer's first question?
Who understands what the customer wants?
Who makes the recommendation?
And, most importantly,
Who does the customer come back to next time?
The biggest change in AI shopping may not be that we click a few fewer buttons.
It may be who gets to understand us before we buy.
Sources
Reuters — Retail stagnates, ad dollars help Walmart stay the course (Aug. 18, 2026)
Reuters — Retailers tap AI shopping traffic but fight to keep customer data (Aug. 7, 2026)
Reuters — AI-referred U.S. shoppers browse longer, spend more per visit, data shows (June 15, 2026)
Reuters — Rise of AI shopping pushes merchants to protect loyalty, Adyen says (Aug. 13, 2026)
Forrester — The State of Agentic Commerce in Mid-2026
Visa Acceptance Solutions / PYMNTS Intelligence — 2026 Global Digital Shopping Index: The Agentic Inflection Point
Google — Universal Commerce Protocol and agentic commerce
BEYOND THE OBVIOUS.