Why AI shopping answers recommend big chains, and what one word changed.
Ask an AI assistant where to buy running shoes, a sofa or a bottle of wine, and the answer will probably be a big chain.
That is the short version of a study Lightspeed Commerce published on 29 September 2026. It is worth reading closely if you run an independent or founder-led business, because the most useful finding is not the headline.
What the study tested
The research was run by Vaer AI for Lightspeed. It put 20,000 shopping prompts across 10 product categories to ChatGPT, Google AI Mode, Google AI Overviews and Gemini, set in Los Angeles, San Francisco, New York City and Montreal, between June and August 2026. That produced about 460,000 AI responses, some with live web search and some without.
What it found
When the AI relied only on its own training data, it chose the larger retailer 90–94% of the time in head-to-head comparisons.
With live web search switched on, the picture improved but national chains still led, at 46–58% of responses. The products behaved differently. Google AI Overviews recommended no local store in 68% of shopping responses, while ChatGPT and Google AI Mode surfaced a local retailer in roughly 70%.
A separate consumer survey commissioned by Lightspeed (Censuswide, 2,000 North American consumers) found that 56% have used AI to help make shopping decisions. So this is not a niche behaviour.
The one word
Here is the finding I would pay most attention to. When the researchers added the word "independent" to the prompt, the share of recommendations going to large national chains fell from about 44% to as little as 9%.
Nothing about the businesses changed between those two prompts. The question did.
What the study does not show
Two limits matter before drawing conclusions.
First, Lightspeed sells software to independent retailers. It has an interest in this result, and the research should be read with that in mind.
Second, the study covers retail shopping prompts in four cities. It does not test service businesses, consultants or B2B companies, and it does not explain why the AI behaves the way it does. It measures what came back.
My read: this is a positioning question first!
What follows is my interpretation, not a finding.
The AI answered the question it was asked. A generic question ("where should I buy this?") got the biggest, most documented answer. A more specific question got a more specific one.
That is how recommendations have always worked. Ask a friend for "a restaurant" and you get the famous one. Ask for "somewhere quiet with good vegetarian food" and you get a different answer, if one exists and your friend knows about it.
So before asking how to get AI to recommend your business, I would ask an older question: what is your business specifically known for? If neither you nor your customers can say it in a sentence, there is nothing specific for anyone to ask for, whether they are asking a search engine, a neighbour or an AI assistant.
Three questions worth answering this quarter:
If a customer had to describe you with one specific word or phrase, the way "independent" worked in this study, what would it be?
Is that description stated plainly where people and machines can read it: your homepage, your product and service pages, your business listings, your reviews?
Do your customers use the same words when they talk about you?
None of this is an AI tactic. It is positioning, and it would be worth doing if AI assistants did not exist. The study simply suggests that being specific may matter even more when the first place a customer looks is a chat window.
Source: Lightspeed Commerce, "AI favors big retailers over local stores 9 out of 10 times, new study finds", 29 September 2026. Research by Vaer AI; consumer survey by Censuswide.
