When AI does your shopping, will you still pick the brand?

AI shopping helpers may soon choose products for you, and early research suggests they do not always choose the way people expect. An analysis by Christophe Lachnitt on the Superception site, published November 10, 2026, pulls together findings from Adobe, Wharton and Accenture on what happens to brand loyalty when software does the buying.

What the research shows

Adobe found that visits to brand websites that come from AI tools convert 90% better than visits from other sources. Each visit also earns 83% more revenue. In plain terms, people who arrive after asking an AI tool seem to be nearly ready to buy.

Wharton researchers tested how AI models pick products. They ran 26,000 trials in which an AI chose one smartwatch from eight. Sometimes it saw extra information, such as consumer reviews or media recommendations. The order of that information also changed.

• Adding just one source sometimes changed the recommendation radically.
• Product quality alone did not guarantee the AI would favor it.
• The AI choices were not as objective as one might assume.

What people will hand over

Accenture's international study found that shoppers are most willing to delegate purchases they find tedious. They want to keep control over buys that feel emotionally meaningful or tied to who they are. The pleasure of choosing certain products will not disappear, and handing over the task will not be total.

The analysis also describes a possible new habit called delegated loyalty. A shopper tells an agent to repeat usual purchases or stick to certain brands, perhaps only under conditions such as price. Automated reordering already exists, but on a far smaller scale than agents could bring.

Who is most at risk

Brands that sell mainly on convenience may gain from more automation. Brands whose value comes from the experience of choosing will want to protect that direct moment with customers.

The most exposed are brands that inspire no emotional preference and have no clear functional edge. Without a real point of difference, they are more open to being traded away by an algorithm.

What businesses can do

Winning new customers may depend more on whether AI agents can identify, evaluate and recommend a brand. That puts weight on the proof behind its claims. The experience after the purchase matters too, because it can still build lasting human preference. Where a shopper turns that preference into a standing instruction to an agent, it becomes a stronger way to keep customers.

What is not clear yet

The analysis does not say how fast shoppers will adopt AI buying, or how much they will delegate in each category. Much of the evidence comes from tests, so real shopping behavior may differ.