How does ChatGPT recommend businesses?
How does ChatGPT recommend businesses? It builds each answer from two places: what the model absorbed in training, and the pages, listings and feeds it can pull in at the moment someone asks. A business gets named when those sources describe it the same way, say plainly what it sells and where, and match the wording of the question.
There is no form to submit and no ranking to check. SOCi's Local Visibility Index 2026 found that ChatGPT recommends 1.2% of multi-location businesses, against 11% for Gemini and 7.4% for Perplexity.
What ChatGPT reads when it answers
Three inputs feed a recommendation. They behave differently, so they need different work.
Training memory
The model has a general sense of which brands exist and what they are known for. That sense comes from how often and how consistently a brand is described across the web. You cannot edit it quickly. You change it over months by getting accurately described in more places.
Live retrieval
When ChatGPT searches, it reads pages and quotes from them. Pages that state the answer in the first two sentences are easy to quote. Pages that open with a story about the founder are hard to quote. The page types that tend to get pulled in are guide pages, comparison pages, pricing pages, FAQ pages and location pages, because each one answers a specific buyer question.
Listings and feeds
For local questions, ChatGPT and Perplexity use Bing Places and Yelp. Gemini grounds its local answers on Google Maps. For products, OpenAI publishes a product feed spec with 9 required fields. Shopify's catalog needs a standard category, an image per variant, a price above zero and full policies. A store with a thin catalog is hard to recommend, however good its blog is. If you run a store, our ecommerce page covers the feed side in more detail.
How does ChatGPT recommend businesses for local questions?
Local questions lean heavily on listing data. If someone asks for the best dentist near them or a plumber who does same-day work, the engine needs a business record it trusts: name, address, hours, categories and reviews that agree with each other.
| Engine | Local data source | Share of multi-location businesses recommended (SOCi, 2026) |
|---|---|---|
| ChatGPT | Bing Places and Yelp | 1.2% |
| Gemini | Google Maps | 11% |
| Perplexity | Bing Places and Yelp | 7.4% |
Many owners keep a polished Google Business Profile and have never opened Bing Places. That owner looks fine in Gemini and may be missing from the data ChatGPT leans on. Yelp is the other half of the same problem: an unclaimed or out-of-date Yelp page sends a conflicting signal.
Signals that appear to matter
Mentions on other sites
Ahrefs studied 75,000 brands and found that third-party brand mentions correlate with AI Overview visibility at 0.66, against 0.22 for backlinks. That study covers Google's AI Overviews. ChatGPT was not part of it. I would still plan around the pattern. Being named in trade roundups, local directories, review sites and comparison articles carries more weight than link counts alone.
Pages outside the top 10 still get cited
Ahrefs, in March 2026, found that 38% of AI Overview citations come from the top 10 results and 31% from positions 11 to 100. A page that answers a narrow question precisely can get cited without ranking first for the broad term.
Consistency
If your site says you are open until 6, Yelp says 5 and Bing Places says nothing, the engine has three versions of you. The same goes for product data: a price on the page that differs from the price in the feed is a reason to skip you.
What does little
Two popular fixes deserve less of your week than they get.
- llms.txt: AI crawlers rarely request the file, and Google discourages it.
- Schema markup on its own: structured data correlates only weakly with AI citations. It still matters for rich results and for keeping feeds consistent, so keep it accurate, but do not expect it to get you named.
Why the effort pays
Locus reported in Q2 2026 that 45% of US online shoppers use an AI tool somewhere in the purchase journey. Adobe measured AI traffic to US retailers up 393% year over year in Q1 2026, converting 42% better than non-AI traffic. Volume is still small next to Google for most sites, but the visitors who arrive have already been told you are a good answer.
Six things to do this week
- Write 10 questions a buyer would type before choosing a business like yours. Use the words your customers use.
- Run each question in ChatGPT, Gemini and Perplexity. Record who gets named and which sources are cited. Save the results so you can compare in a month.
- Claim and complete your Bing Places listing and your Yelp page. Match the name, address, phone, hours and categories to your site exactly.
- Give each offer its own page: what it is, who it is for, price or price range, where you deliver or serve, and the policies. State the answer in the first two sentences.
- Look at the sources cited next to your competitors. Those are the trade lists, directories and articles that already mention them. Pitch the ones that fit you.
- If you sell products, check the catalog for the gaps above: missing categories, variants without images, zero prices, thin policies.
A worked example
JCS Fashions is a 4,878-product Shopify apparel store in Milpitas, California, and the first customer we worked with at LynkAI. In 2026 its sessions from AI assistants went from 20 in January to 569 in October. Over the same period we fixed its offers and catalog, built the entity layer and published about 290 quality-gated articles between April and September. By the end, 233 guide pages were earning impressions on 24,164 queries, and Google clicks had grown from 205 a month in January to 626 in September. The store's own numbers are on the JCS Fashions customer page. One store is one data point, but the order of work (clean data, then clear pages, then mentions) is the same one the steps above follow.
Where to start
Step two is the one most owners skip, and it tells you the most. If you want the answer without running 30 prompts by hand, the free AI visibility scan shows where you are named today and where a competitor is named instead.