ChatGPT names the brands that the sources it searches, and the text it learned from in training, already connect with your category. If those sources do not name your brand, ChatGPT will not name it either. When ChatGPT searches the web, it lists the brands that appear on the pages it picks, so a brand missing from those pages is usually missing from the answer.
Five causes explain most cases: ChatGPT cannot read your site, does not know who you are, does not trust what it finds about you, has the wrong category or old facts, or never sees you on the websites it uses as sources for your category. Each cause leaves a different sign in the answers, such as whether ChatGPT can describe you when asked by name and which websites it cites, meaning the links it shows as sources. Ask ChatGPT the six test questions below, match what you see to one row of the table, and apply that fix first.
One missing answer proves little, because ChatGPT gives different answers to the same question. Judge by how often you appear across many runs. The same five causes apply to Gemini, Perplexity and Google AI Overviews, the AI answer Google shows above its search results, though each engine reads different sources. If you show up in one engine and not another, read why AI engines disagree about your brand.
| Reason | What you see in the answers | Likely cause | First fix |
|---|---|---|---|
| Site not readable | Your own site is never cited as a source | Your site's bot rules or firewall block ChatGPT's search bot | Let ChatGPT's search bot fetch your pages |
| Unknown brand | Never named; asked about you by name, it hedges or confuses you with another company | Few consistent mentions of your name and category | Use one name, category and description everywhere |
| Not trusted | Named rarely, low in the list, hedged, or next to complaints | Thin or negative reviews and coverage on other sites | Earn reviews and independent coverage; resolve visible complaints |
| Wrong or old facts | Wrong price tier, use case, country, product or price | Old, well-linked pages outnumber current ones | Update your strongest pages and listings with dated facts |
| Missing from sources | Never named, yet described correctly when asked about you by name | Absent from the lists, reviews and forums ChatGPT cites | Get onto the pages ChatGPT cites for your category |
Why is my brand never mentioned in ChatGPT responses when people ask about my industry?
Because the pages ChatGPT reads for industry questions do not name you, and its training gave it no clear link between your brand and your industry. For a question such as "best skincare brands in India", ChatGPT with web search reads a handful of roundups, marketplace pages, forum threads and directories, and lists the brands those pages mention. Without a search, it answers from what it learned in training, which stops at a cutoff date. A brand that is absent from both stays out of the answer.
Industry answers lean on a small set of sites. In Depra AI's August 2026 study of 480 shopping answers from India, reddit.com was cited in 106 answers, youtube.com in 100, nykaa.com in 93, amazon.in and flipkart.com in 79 each, and myntra.com in 47. In the October 2026 re-run of 479 answers, amazon.in was cited in 132 answers, flipkart.com in 82, hindustantimes.com in 76, nykaa.com in 65, thedermaco.com in 64 and myntra.com in 44. If your brand is not on the pages of sites like these for your category, it rarely reaches the answer.
To find which cause applies, ask ChatGPT what your company does, by name, in a fresh chat. A correct description points to Reason 5, missing from sources. A vague answer or a mix-up with another company points to Reason 2. A wrong category, price tier or country points to Reason 4.
Most-cited sites by answers: reddit.com in 106 of 480 answers in August 2026; amazon.in in 132 of 479 answers in the October 2026 re-run. (Depra AI English vs Hinglish study, led by Zia Wasi, collected from India on 14 Aug 2026 and re-run on 13 Oct 2026)
Why doesn't ChatGPT mention my brand?
ChatGPT doesn't mention your brand because one of five causes blocks it: it cannot read your site, does not know you, does not trust what it finds, has old facts, or does not see you on the sites it cites. Each cause has a quick check. Run the checks in the order below, because a blocked site makes every other fix slower to show.
| Cause | How to check | First fix |
|---|---|---|
| Site not readable | Open yourdomain.com/robots.txt and look for OAI-SearchBot and any blanket Disallow rule; search your server logs for 403 replies to OAI-SearchBot | Allow OAI-SearchBot in robots.txt and in your firewall or CDN |
| Unknown brand | Ask ChatGPT "What does [your brand] do?" in a fresh, logged-out chat | One name, category and description everywhere; Organization schema |
| Not trusted | Run one category prompt 5 times and note your position and the tone used about you | Earn reviews and independent coverage; resolve visible complaints |
| Wrong or old facts | Compare ChatGPT's description of you with your current category, price tier and products | Update your most-linked pages and listings with dated facts |
| Missing from cited sites | Copy the cited links from 5 runs and mark which pages name rivals and skip you | Get onto those pages; create or claim your marketplace listings |
How do you test why ChatGPT leaves your brand out?
Ask ChatGPT six buying questions about your category and record who it names. A prompt is the question you type into ChatGPT. Use a new chat while logged out, so your chat history and memory do not shape the answer, and ask each prompt at least 5 times.
For every answer, write down four things: whether you are named, your position in the list, the tone used about you, and the websites cited. Then ask ChatGPT what your brand does, by name, and note whether the description is correct. Run the same prompts on Gemini, Perplexity and Google AI Overviews to see whether the problem is ChatGPT alone.
For a first look without the manual runs, the free AI audit asks 5 questions on 4 AI engines and shows whether each one names you. It needs no signup and no card.
- Category shortlist: "What are the best [category] brands for [use case] under ₹[budget]?" For example, "best sulphate-free shampoo for curly hair under ₹800".
- Direct compare: "I need a [product] for [situation]. Which brands should I compare?" This shows the default shortlist ChatGPT builds for your category.
- Buyer fit: "Recommend a [category] for [specific buyer] who cares about [priority]." This checks whether ChatGPT matches you to the right buyer and price tier.
- Community sources: "What do people on Reddit recommend for [problem]?" This shows which forums and review sites ChatGPT reads for your category.
- Head to head: "Compare [your brand] and [competitor] for [use case]." This makes ChatGPT state what it thinks it knows about you.
- Hinglish version: ask the category shortlist question in Hinglish, which is Hindi and English typed in English letters, the way many Indian buyers ask. Compare its brand list with the English one.
Language changes what AI reads. In August 2026 (480 answers), Gemini cited sources in 90.0% of English answers and 41.3% of Hinglish answers to the same buying questions; ChatGPT brand lists moved 2.2 points between languages beyond normal rerun change, and Perplexity lists moved 23.4. In the October 2026 re-run (479 answers), the Perplexity gap was 28.2 points, Gemini 2.0 and ChatGPT 0.8, and Gemini cited sources in 73.4% of English and 56.3% of Hinglish answers. (Depra AI English vs Hinglish study, 480 answers from India, 14 Aug 2026, and its re-run of 13 Oct 2026, 479 answers)
Why is my brand not showing up in ChatGPT?
Your brand is not showing up in ChatGPT because the sources behind its answer leave you out, or because ChatGPT cannot read your own site. Before you change anything, confirm the absence is real. Ask each prompt at least 5 times in fresh chats, since one missing answer proves little.
Then look at the cited links. If ChatGPT cites review sites, marketplaces and competitors for your category, and never a page on your own domain, start with Reason 1, site access. If your site is cited for other questions but your brand is still left off the shortlist, start with Reason 5, missing from sources.
If you show up in Gemini or Perplexity and not in ChatGPT, the engines are reading different pages. Each AI engine finds web pages through its own search route, which our map of which search index each AI engine uses sets out from each vendor's own documentation. ChatGPT search depends on its own crawler, OAI-SearchBot, being allowed on your site.
Reason 1: Can ChatGPT read your site?
ChatGPT cannot cite pages its search crawler cannot fetch. A crawler is a bot that visits web pages and reads them. OpenAI says sites that block OAI-SearchBot, the crawler behind ChatGPT search, are left out of ChatGPT search answers (OpenAI crawler documentation).
What you see: ChatGPT cites review sites, marketplaces and competitors for your category, and never a page on your own domain. When you ask about your brand by name, its description is thin or out of date.
Why it happens: three causes are common. First, your robots.txt, the file that tells bots which pages they may fetch, blocks OAI-SearchBot by name or blocks every bot with a "User-agent: *" line followed by "Disallow: /". A user agent is the name a bot sends with each request, and * means every bot. Second, a firewall, or the bot filter in your CDN (the network that delivers your site's pages, such as Cloudflare), refuses OAI-SearchBot with a 403 error, the "access denied" reply a server sends. The page loads for people and fails for the bot. Third, your main text loads only through JavaScript, code that runs in the browser after the page arrives, and many AI crawlers may not run it.
Blocking GPTBot is a separate choice. OpenAI uses GPTBot to collect text for training its models, so a rule that blocks only GPTBot keeps your pages out of training and leaves OAI-SearchBot free to fetch them for answers.
How to test: open yourdomain.com/robots.txt and look for OAI-SearchBot, GPTBot and any blanket Disallow rule. Then search your server or CDN logs, the record of every request your site receives, for OAI-SearchBot and check the status code on each request. A run of 403 responses is a checkable, fixable cause. Depra AI's AI crawler checker reads robots.txt for 41 AI crawlers from 26 vendors, on your site and your competitors' sites, and shows the exact rule that decides each bot. If you send it your server or CDN logs, it also shows which AI bots visited and which hit errors. Our list of AI crawlers and what each one does gives each bot's user agent.
The fix: allow OAI-SearchBot in robots.txt, add it to the allowed bots in your firewall or CDN, and make sure product names, prices and descriptions are in the HTML your server sends.
Reason 2: Does ChatGPT know who you are?
ChatGPT names brands it has a clear picture of, built from many consistent mentions across the web. If that picture is missing, you stay out of category answers.
What you see: your brand never appears in category answers. When you ask about it by name, ChatGPT hedges, says it has limited information, mixes you up with a company with a similar name, or makes up details. This is the sign that separates Reason 2 from Reason 5: there, ChatGPT describes you correctly when asked by name.
Why it happens: your footprint is thin or inconsistent. Your site says one category, your marketplace listing says another, and directories use an old name or a different spelling. With no stable pattern to follow, the model has nothing firm to repeat.
The fix: use one name, one category and one short description everywhere you control. Write an About page that says what you sell, who it is for, where you operate and when you started. Add Organization and Product schema, the structured data that labels these facts for machines. Keep listings in credible industry directories consistent with your site, and add an entry on Wikidata, the free public database of facts, if your brand meets its notability rules.
Reason 3: Does ChatGPT trust what others say about you?
ChatGPT recommends brands that other websites vouch for. If the outside evidence about you is thin or negative, it names you rarely or with caveats.
What you see: you do appear, but in a small share of answers, low in the list, wrapped in phrases like "some users report", or next to complaints.
Why it happens: you have few reviews on the marketplaces and review sites in your category, little independent coverage, or complaint threads that rank well.
The fix: earn reviews on the sites that show up in your test answers, such as Amazon, Flipkart or Nykaa for consumer brands (see the sites AI cites for shopping questions in India) and software review sites for SaaS. Reply to and resolve the complaints that keep surfacing. Get included in independent comparisons written by people outside your company.
Reason 4: Does ChatGPT have the wrong category or old facts?
ChatGPT repeats how other pages describe you, even when that description is wrong or out of date. You end up named for questions your buyers do not ask.
What you see: ChatGPT lists you as a budget option when you are premium, matches you to the wrong use case or country, names a discontinued product, quotes an old price, or uses your name from before a rebrand.
Why it happens: your oldest pages and listings often have the most links, so they outrank the current ones. What the model learned in training stops at a cutoff date, and search results tend to favour established pages.
The fix: state your category, price tier and main use case plainly on your key pages. Update the pages that other sites link to most instead of only adding new ones, and put a date on facts like prices and launches. Ask directories and roundups that describe you wrongly to correct their entries. Pages that include sources, quotes and numbers also give AI engines more to cite.
Adding citations, quotations and statistics raised a page's visibility in AI answers by up to about 40% in a GPT-3.5 test setup. (Aggarwal et al., KDD 2024)
Reason 5: Are you missing from the sites ChatGPT cites?
For buying questions, ChatGPT with web search reads a handful of pages and names the brands those pages mention. If the lists, reviews and forum threads it reads leave you out, the answer leaves you out too.
What you see: your brand is absent from recommendation answers, yet ChatGPT describes you correctly when you ask about you by name. The competitors that do appear are named in the "best of" list articles, comparison posts, forum threads and video reviews that the answers cite.
Why it happens: pages that compare options answer a buying question more directly than a brand homepage does. ChatGPT also reads a different set of pages on each run. In Depra AI's August 2026 study of 10 shopping questions asked from India (480 answers), two ChatGPT answers to the same question shared 33.4% of their cited websites on average, so being named on many of the cited pages matters more than being named on one.
The fix: collect the cited links from your test answers and list the pages that name competitors and skip you. Ask those authors to review your product, get into comparisons, and answer relevant forum threads openly as the brand. Our guide on how to get your brand cited in AI answers walks through outreach for each page type.
Depra AI's citation gap analysis lists the websites that recommend your competitors but never mention you, by domain, subdomain and page, with your competitors' own sites filtered out.
Why does ChatGPT mention my competitors but not my brand for category searches?
Your competitor is named on more of the pages ChatGPT reads for that category, or ChatGPT has a clearer picture of what they sell. You can find out which in four steps.
For consumer brands in India, many of those pages are marketplaces. Depra AI labels every cited website by site type, marketplace included, so a brand sees which marketplaces AI answers cite in its category, such as Amazon.in, Flipkart, Nykaa, Myntra and Meesho, and where its rivals are cited and it is not. For ChatGPT, it also lists the stores that ChatGPT's product cards point to. Depra AI does not track rankings inside a marketplace's own search, marketplace sales, Buy Box, ads or Amazon Rufus. Where the gap is a marketplace page, the fix is to create or claim your listing so the page exists for an engine to find and cite.
- Collect the sources: run one category prompt 5 times and copy every cited link.
- Mark each page: note whether it names your competitor, you, both or neither.
- Read the gap: pages that name only your competitor are your outreach list, which is Reason 5.
- Check shared pages: if a page names you both and ChatGPT still picks them, the gap is Reason 2 or Reason 3. Run the head-to-head prompt to see what ChatGPT believes about each of you. Then see what to do when a competitor appears in the AI answer.
How to know if a fix worked
Measure how often ChatGPT names you across many runs, before and after the change. A single answer cannot tell you, because answers vary. In Depra AI's August 2026 study of 10 shopping questions asked from India (480 answers), two ChatGPT answers to the same question, asked within one hour, shared 54.7% of the brands they named on average. Read more on why ChatGPT gives different answers to the same question.
Record a baseline before you change anything. Ask each question at least 5 times per engine, log whether you appear, your position, the tone and the cited sites, and repeat on the same questions every week.
Judge the rate over weeks, and allow for noise. A 50% rate measured over 45 answers could sit anywhere from about 36% to 64%, so a move from 55% to 60% on that sample may mean nothing. Our guide to how to measure AI visibility covers the metrics and how many answers you need. If nothing moves after several weeks, run the six test prompts again: you may be fixing the wrong reason.
Depra AI runs this for you. It asks your buyers' questions to ChatGPT with web search, Gemini and Google AI Overviews every day and to Perplexity every week (daily on Enterprise), from inside your chosen country. It stores every answer and records whether you were named, in what order, in what tone and which pages were cited. Your report shows a confidence range, the band your true rate most likely sits in. After you mark an action done, Depra AI shows your visibility on the day you marked it next to your visibility now. That view shows the change. It does not prove your fix caused it. See AI visibility tracking for how it works.
Depra AI's action plan, the list of pages to publish or update and sites to get cited on, comes from a proprietary method developed by Zia Wasi, Depra AI's co-founder and a GEO (generative engine optimization) researcher. Depra AI used the same method on its own brand and got into ChatGPT's answers for AI-visibility questions in India, built without paid backlinks or paid marketing. The Depra AI case study gives the numbers and what has not moved yet.