ChatGPT brand mentions: which brands it names, and how to get yours named

ChatGPT brand mentions are the times ChatGPT names a company or product in an answer, whether or not it links to it. The brands it names most are the ones its model already knows well. A lesser-known brand gets named when ChatGPT finds a specific, checkable reason to pick it.

So brand visibility in ChatGPT is won with evidence more than with rankings: what your own pages state plainly, and what other sites say about you. Those are the places ChatGPT’s search finds a reason to pick one name over another.

Which brands ChatGPT mentions most

No neutral ranking of the most mentioned brands in ChatGPT exists. The public lists come mostly from tool companies’ own panels of prompts, so the order reflects which questions each panel asks as much as it reflects ChatGPT.

Independent research shows the pattern behind every list. Researchers at Maastricht, Utrecht and Zurich universities put 117 real shopping questions to ChatGPT, Gemini and Google’s AI Overviews in September 2026, three times each.

ChatGPT stated a first-person preference in 79% of the answers that recommended a product, against 7% for Gemini and 2% for AI Overviews, and it called one product the best in 73%. It likes to pick a winner.

The winner is usually a name the model already knows, and a controlled experiment shows why. For a B2B company, the ranking that matters is the one for its own buyers’ questions, and that one can be measured without buying a tool.

How ChatGPT decides which brands to name

ChatGPT draws on two things: what its model learned in training, and what its web search finds. OpenAI’s shopping help names both: the response the model generates before it considers any new search results, and product data from providers alongside other third-party content.

Search runs when a question needs current information. ChatGPT rewrites the question into searches of its own, may fold saved memories into them, and may use an approximate location from your IP address. Two buyers asking the same thing can see different brands.

Money is not an input. OpenAI’s ads principles, published in January 2026, say ads “do not influence the answers ChatGPT gives you”, and its shopping help says product results are not ads. An ad can sit beside an answer; it cannot buy a place in one.

Why known brands win, and what breaks the tie

A controlled study by researchers at Trine University and Texas A&M, last revised in October 2026, shows the mechanism. It gave GPT, Claude and Gemini models product cards for a well-known skincare brand and an invented rival, then asked which to recommend.

With identical specifications, the well-known brand won 100% of the time. A rating advantage of 0.075 stars was enough to break that hold, and once the products differed on quality, the brand name explained only 1.2% of how they were ranked.

The authors’ conclusion is the useful part for a smaller company: the barrier “is not brand equity itself, but the lack of any distinguishing information.” A familiar name wins when nothing else separates the options. What the invented brand added decided whether it won:

The right-hand column is our reading for B2B. The study tested consumer products through the models’ APIs, not the ChatGPT app, so treat it as the mechanism rather than a forecast for your category.

One result is a warning, not a tactic: the models treated a fabricated clinical citation as real evidence. Buyers do not. Forrester’s 2027 predictions expect two-thirds of providers to lose deals because buyers cannot verify claims, so only claims you can prove belong on the page.

What this means for a B2B brand

B2B buyers already ask. In Forrester’s State of Business Buying 2026, based on a survey of nearly 18,000 business buyers, 94% said they use AI during their buying process. They check what it tells them with peers, product experts and industry analysts.

Most B2B vendors also start further back than a consumer brand. A model may know a category’s leaders well and a mid-sized vendor barely at all, so a smaller vendor’s name tends to reach the answer through what ChatGPT’s search finds: its own pages, and other sites that name it.

Those other sites carry weight. In the Maastricht-led audit, editorial and product-review sites made up 56.7% of the domains ChatGPT showed with its product recommendations. For a B2B vendor, that means the trade publications, review sites and comparison articles its buyers read.

The buyer’s own software is next. Forrester’s October 2026 predictions expect buyers to deploy AI agents across the buying process, from supplier discovery to contract-renewal analysis, so the first shortlist may be drawn by an AI agent working for the buyer.

How to check which brands ChatGPT names for your buyers

A tool is optional. What you need is a fixed set of questions, asked the way buyers see ChatGPT, often enough to see past chance:

• Write 20 to 30 buyer questions. The best provider for your category and customer size, alternatives to your main rival, head-to-head comparisons, and the problem you solve, in the words buyers use.
• Ask in the app, not the API. In the Maastricht-led audit, ChatGPT’s interface and its API shared only 12.0% of the domains they showed for the same question asked moments apart.
• Switch off personalization. Memory and custom instructions shape what ChatGPT recommends, so use a temporary chat set to Unpersonalized, or a logged-out browser.
• Ask each question three times. In the same audit, two ChatGPT runs of one question had only about 18% of their combined products in common.
• Log four things per answer. Whether you are named, where you sit in the list, whether a rival is called the pick or the best, and which sources are shown.
• Repeat it monthly with the same questions, and compare the share of answers naming you with each rival’s.

Twenty-five questions run three times give 75 answers a month. The figures below are made up, to show the layout:

At 75 answers, a share near 50% can move about 11 points either way by chance alone. Read the trend over several months, because AI answers change from week to week even when nothing on your side has.

How to get ChatGPT to recommend your brand

• State what makes you the right pick, in checkable terms. Who you serve, how pricing works, what you integrate with, and results with numbers you can stand behind. That is the distinguishing information the Trine and Texas A&M study found breaks a known brand’s hold.
• Lead with what only you can say. Claims every rival also makes cancel out, as the first table shows, and pressure tactics barely register.
• Let ChatGPT’s crawler in. OpenAI says sites that block OAI-SearchBot will not be shown in ChatGPT search answers, so check robots.txt, your firewall and the other settings that keep AI engines out.
• Answer buyers’ questions on your own pages. ChatGPT rewrites each question into searches of its own, so a page that plainly answers one of those searches gives it something to find.
• Get named accurately where buyers check. A complete listing on your category’s review sites, coverage in the trade publications buyers read, and a fair place in comparison articles. Mentions are earned the way links are, by giving a writer a reason to name you.
• Measure the result in answers, not visits. A buyer who reads the answer often never clicks, so count branded searches and “how did you hear about us?” replies alongside the monthly check.

In the programs we run, this becomes a monthly loop. We rerun the questions, find the ones where a rival is named and the client is not, then fix the page or earn the mention that would change the answer.

Run that way, getting a B2B brand named in AI answers stops being guesswork and becomes a routine with a number attached, whether your own team owns it or hands it to a partner that does this every month.