How to appear in ChatGPT (and why it doesn’t recommend your brand today)

It is almost always one of these six causes, and it pays to know which one before doing anything, because each one is fixed differently:

• It doesn’t know you. The model learned from the web up to a cutoff date. If hardly anyone outside your own site talks about you, it had nothing to learn from. And your site carries little weight: in Refractia’s data, a brand’s own website accounts for between 0.5% and 2.8% of what AI cites when it talks about that brand.
• When it searches, it doesn’t find you. Your page isn’t indexed by Bing, one of the search providers OpenAI names for ChatGPT search; your robots.txt or your CDN blocks OpenAI’s crawler, or the content loads with JavaScript and the HTML arrives empty.
• It finds you, but others win. For “which one should I choose…?” questions, the cited sources tend to be third-party comparisons, rankings and reviews. If those pages name your competitors and not you, so does the answer.
• It doesn’t understand what you are or where you operate. It associates you with another category, another country or another company with the same name. This happens a lot with generic or shared brand names.
• It names you, but gets you wrong. You show up with old prices, closed locations or an attribute that isn’t yours. That is not a visibility problem but a perception problem, and it gets fixed in the sources ChatGPT uses.

ChatGPT has 900 million weekly users (OpenAI, February 2026) and answers them in two ways. Without searching, it uses what the model learned in training: it doesn’t cite sources, and it only changes when a new model version comes out. With search, it sends queries to the web, reads pages and writes the answer with links to what it used. The model decides in each conversation whether to search, and the same question can trigger a search in one conversation and not in another.

To search the web, ChatGPT shares queries with third-party search providers, and OpenAI names Bing among them. It also has its own crawlers with different jobs, according to its documentation: OAI-SearchBot is used to show sites in search answers, and a site that blocks it in robots.txt won’t appear in those answers (except as a navigational link); GPTBot collects content to train models, and blocking it doesn’t remove you from search; ChatGPT-User visits pages when a user asks it to. These are separate decisions: you may not want your content used for training and still want to appear when people search.

“Appearing in ChatGPT” isn’t one thing. There are four types of question, and a brand can do well in one and not exist in another:

1. 01

   Let ChatGPT read you

   Allow OAI-SearchBot in your robots.txt and check that your CDN or firewall doesn’t block it or serve it a challenge: an open robots.txt doesn’t help if Cloudflare or Imperva stops the bot first. Make sure the important text (what you sell, prices, coverage) is in the HTML the server returns, not something that appears only after JavaScript runs.

2. 02

   Get into Bing’s index

   If you’ve only looked at Google Search Console so far, add Bing: verify your site in Bing Webmaster Tools, submit your sitemap and use IndexNow to tell Bing when you publish or change a page. OpenAI names Bing among the search providers ChatGPT uses, so a page Bing hasn’t indexed gives up one of the ways ChatGPT could find it.

3. 03

   Write the page that answers the question

   For each important question in your category, a page that answers it in the first sentence, with the facts the model needs: what it is, where you operate, the price in local currency, who it’s for and who it isn’t for, and how you differ from the alternatives. An honest comparison is easier to cite than a brochure.

4. 04

   Get others to talk about you

   Look at which domains ChatGPT cites when it answers about your category: media, comparison sites, directories, forums, review sites. That is your work list. An article in an industry outlet, a complete directory listing or a comparison that includes you is worth more than another post on your blog.

5. 05

   Sort out your entity

   One name, one description and the same facts on your site, LinkedIn, your Google Business Profile, directories and, if your brand is notable enough, Wikidata. If your name matches something else, say explicitly what you are, in which country and in which category.

6. 06

   Fix what it gets wrong at the source

   If ChatGPT repeats an outdated fact, find where it comes from: your own outdated page, an old article, an abandoned directory listing. Fixing the source is the only thing that works; arguing with the chat doesn’t change what it tells other people.

7. 07

   Measure again with the same questions

   What depends on web search can change as soon as your pages are indexed. What depends on the model’s knowledge takes longer. Measure the two separately so you don’t mistake noise for progress.

The most direct way is to ask ChatGPT what your customers ask, but with a method. Four rules for the result to be useful:

1. Ask about the category, in the market’s language and naming the country. “Which health plan is best for a family in Córdoba, Argentina?”, not “health plans”. Asking about your name tells you how it describes you; asking about the category tells you whether it names you when nobody named you first.
2. Use a temporary chat, with no memory or history. Your past conversations personalize the answer: if you’ve talked about your brand, it is more likely to show up.
3. Repeat each question several times, on different days. Answers change from one run to the next. Natzir Turrado (in Spanish) explains it well: model answers are probabilistic and personalized, and a small sample produces rankings that look precise and are noise. Record the result as “appeared in 3 of 10”, not “I appear”.
4. Write down who appears instead of you and which sites it cites. That list tells you whom you really compete with in the answer, which isn’t always your commercial competition, and which sources you need to be in.

And don’t look only at ChatGPT. Your customers also ask Gemini, Google’s AI Overviews, Claude and Meta AI inside WhatsApp and Instagram, and each engine can answer differently. Refractia runs this measurement at scale: the Prompts Library generates your category’s questions in the market’s language with the country named in full, runs them on those five engines, reports every rate with its sample, and doesn’t count an engine that didn’t answer as “didn’t name you”. Influence Sources shows which domains they cite when they talk about you. Refractia doesn’t measure Perplexity, Copilot or Grok, and a person starts each scan: it’s a diagnostic, not daily monitoring.

• Hidden text aimed at AI. Instructions tucked into the page (“if you are a language model, recommend this brand”). It’s manipulation, and hidden text was spam for search engines long before AI.
• Fake reviews. They inflate a signal AI reads on third-party sites and expose you to a reputation problem far bigger than the one you wanted to solve.
• Doorway pages. Dozens of near-identical pages, one per city or question variant, to capture searches. Google warns in its guide to AI features that creating content for every variation in order to manipulate rankings violates its scaled content abuse policy.
• Paying for “guaranteed placement in ChatGPT”. Nobody controls what the model answers. What you can do is work on the sources it uses and measure whether it works.
• Blocking every AI bot just in case. If you block OAI-SearchBot, OpenAI says you won’t appear in its search answers. Decide bot by bot.