---
title: "7 Myths About Getting Cited by ChatGPT and Perplexity"
description: "You can’t buy an organic citation, and a #1 Google ranking won’t guarantee one. Here are seven citation myths, the mechanisms behind them, and what to check instead."
url: "https://groas.com/post/7-things-people-believe-about-getting-ci"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9ea767cd-d914-4b6e-8c6d-39a2bffd9db4.png"
published: "2026-10-25T05:40:18.497Z"
modified: "2026-10-25T05:40:18.558Z"
---

October 25, 2026 · 10 min read

# 7 Myths About Getting Cited by ChatGPT and Perplexity

[Alexander PerelmanHead Of Product @ groas](https://groas.com/author/alexander-perelman)[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![Tabletop town: a crowd hears a town crier praise a rival bakery while one shopkeeper polishes his own sign and a 'SPONSORED' billboard stands ignored.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9ea767cd-d914-4b6e-8c6d-39a2bffd9db4.png)

In this article

1. [Myth 1: “You can pay to be the cited source”](#myth-1-you-can-pay-to-be-the-cited-source)
2. [Myth 2: “Rank #1 on Google and ChatGPT will cite you”](#myth-2-rank-1-on-google-and-chatgpt-will-cite-you)
3. [Myth 3: “Block GPTBot and you’re out; allow it and you’re in”](#myth-3-block-gptbot-and-youre-out-allow-it-and-youre-in)
4. [Myth 4: “Publish an llms.txt file and the models will prefer you”](#myth-4-publish-an-llmstxt-file-and-the-models-will-prefer-you)
5. [Myth 5: “Add schema markup and the models will pick you”](#myth-5-add-schema-markup-and-the-models-will-pick-you)
6. [Myth 6: “Longer, more comprehensive pages win”](#myth-6-longer-more-comprehensive-pages-win)
7. [Myth 7: “Citations are mostly about your own site”](#myth-7-citations-are-mostly-about-your-own-site)

An ad slot appears in ChatGPT, and the paid-search reflex kicks in: set a budget, buy visibility, get cited. I know the reflex. I spent years buying attention on Google. But advice about AI citations keeps borrowing the wrong instincts from both ads and SEO. **You cannot bid your way into an organic answer, and a #1 ranking does not reserve you a mention.** The mechanics are different. So are the mistakes.

## Myth 1: “You can pay to be the cited source”

People believe this because it is how paid search works. You bid, you show up. And now there really are [ads inside ChatGPT](https://latinlaunch.com/), so the leap feels logical: pay OpenAI, get mentioned in answers.

The mechanism says otherwise. Those paid placements run as clearly labeled Sponsored cards below the response, served to adult Free and Go users. OpenAI states that ads do not influence the answer itself. **Buying an ad does not buy a mention inside the organic answer.** I treat them as [two separate channels with two separate rules](https://shoutdigital.com.au/), just as I would paid and organic visibility on Google.

That distinction matters when you measure the result. A Sponsored card can put you in front of someone reading an answer. A citation puts your brand or page _inside_ that answer because it was selected as a source. Count the first as advertising, not proof that the second is working. If you want to be cited, the job is to make your material retrievable and useful to quote. More budget does not replace either step.

## Myth 2: “Rank #1 on Google and ChatGPT will cite you”

This one survives because Google rank feels like a portable measure of authority. If a page wins the search result, surely an answer engine will borrow it. What the deck calls “synergy,” I call assuming two different selection processes pick the same winner.

They do not reliably do that. One test tracking 15 brands across 1,000 queries found that brands on [Google page 1 appeared in ChatGPT only about 62% of the time](https://netranks.ai/). It reported a correlation between Google position and ChatGPT placement of 0.022 to 0.034, effectively random in that test. A check of 15,000 prompts found that [only about 12% of URLs cited](https://stevecoulter.co.uk/) by ChatGPT, Gemini, and Copilot ranked in Google’s top 10 for the same query.

**A search ranking is not a citation reservation.** Google ranks pages for its results. An answer engine assembles a response from material it can retrieve and use, often at the passage level. A high-ranking page may contain the right answer, but its position alone does not make that answer easy to extract. Keep doing useful SEO work; just do not use your Google position as a stand-in for citation performance. Check the answers themselves.

## Myth 3: “Block GPTBot and you’re out; allow it and you’re in”

I understand the appeal of one bot and one switch. The names make this sound simpler than it is. OpenAI uses different fetchers for different jobs, and robots.txt can address them independently:

- **GPTBot** collects [training data](https://honeyb.ai/).
- **OAI-SearchBot** surfaces sites in ChatGPT search answers.
- **ChatGPT-User** fetches a page on demand when someone asks about it.

You can disallow GPTBot while allowing OAI-SearchBot. Opting out of training is not the same decision as blocking search visibility. Perplexity also separates the jobs: [PerplexityBot builds the index used for surfacing](https://shahidali.co/) and respects robots.txt, while Perplexity-User fetches at the moment of a human request and generally ignores robots.txt. The name on the request matters more than a blanket belief that you have either “allowed AI” or “blocked AI.”

If you blocked everything in a panic, check the rules before blaming your content. A page cannot win a citation through a fetcher you have shut out. On the other hand, allowing a fetcher only makes the page available; it does not guarantee selection. For the copy-paste version, use [the AI crawler swipe file with rules labeled by bot and engine](https://groas.com/post/the-ai-crawler-swipe-file-copy-paste-rob). Decide what you want to permit, bot by bot. **Training opt-out is not a visibility strategy.**

## Myth 4: “Publish an llms.txt file and the models will prefer you”

_llms.txt_ gets sold as robots.txt part two. Put a clean Markdown summary of your site in the root directory, and answer engines will supposedly take the hint. It sounds plausible. Developers like clean inputs; vendors like a ten-minute fix.

I would not put it on a citation-work checklist. SE Ranking checked roughly 300,000 domains and found [no statistically significant correlation between llms.txt and AI citation frequency](https://lacleo.ai/). OtterlyAI’s test found no visibility improvement either. [Google confirmed in 2025 that it does not support llms.txt](https://thehoth.com/), and its May 2026 AI guide lists it as unnecessary for AI Overviews, AI Mode, or generative features. Coding assistants may read these files; that is not the same as consumer answer engines preferring your page when they cite a source.

**A tidy file is not a citation signal.** Publish one if your developer wants tidy documentation. Just do not log the upload as a visibility win and move on while the pages you actually want quoted remain hard to fetch or hard to use.

## Myth 5: “Add schema markup and the models will pick you”

People believe this because schema helped win rich results on Google. I ran that play for years: mark up the FAQ, the product, the reviews, and watch the search result get prettier. It is tempting to treat markup as a direct vote for an AI answer too.

But richer display and source selection are different jobs. A cross-platform test of 730 citations found [schema on 43.1% of cited pages and 44.8% of non-cited pages](https://semarkglobal.com/). An Ahrefs matched test of 1,885 pages adding JSON-LD found [changes of +2.2% on ChatGPT, +2.4% on AI Mode, and -4.6% on AI Overviews](https://semarkglobal.com/). Those results do not make a convincing case that adding markup gets a page chosen.

Schema can help a parser understand what a date or price means once it reads your page. **Helping a system understand a page is not the same as making it select that page.** Keep existing schema when it is correct and useful. But if your citation report is empty, another markup type is a poor substitute for an answer worth quoting and independent places that name you. The search result can look immaculate while the answer engine quotes someone else.

## Myth 6: “Longer, more comprehensive pages win”

On Google, depth often correlated with ranking. A 4,000-word guide could cover more keywords, earn more links, and look authoritative. That history makes “write a longer page” sound like sensible citation advice.

Answer engines can lift a passage without rewarding the length around it. An analysis of roughly 2 million organic sessions and 7,500 ChatGPT referrals found that [72.4% of cited pages shared an _answer capsule_](https://previsible.com/): a concise, self-contained 40- to 80-word unit after an H2 that still makes sense when lifted away from the page. That does not mean every page should be 80 words long. It means the useful part should not require a reader, or a retrieval system, to work through nine paragraphs of setup.

**Make the answer stand on its own.** Lead with it, give each section one clear job, then add the detail that earns its place. If your definition hides in paragraph nine, a page that gives the same answer in paragraph one is easier to quote. Length can support an answer. It cannot rescue one you have buried.

## Myth 7: “Citations are mostly about your own site”

This is the one I find hardest to kill because it feels like control. In paid search, your landing page is the asset. You write it, test it, and watch what happens. So founders keep asking for one more homepage rewrite while their category gets explained by someone else. I get the impulse. I like an asset I can edit on Tuesday.

Answer engines do not share that loyalty. An AirOps analysis of [21,311 brand mentions across ChatGPT, Claude, and Perplexity found that 85% came from third-party domains, versus 13.2% from brand domains](https://brandmentions.com/). For top-funnel commercial queries, brands were 6.5 times more likely to surface via a third party than via their own site. Perplexity leaned particularly hard on community material in the cited research: Profound tracked 680 million citations with [Reddit at 46.7% of Perplexity’s top-10 share](https://geoaura.world/), while a later January 2026 check put Reddit at 24%, still the highest of any engine.

Those numbers do not make your website irrelevant. It still needs to be fetchable and quotable. They do mean **your site is not the whole citation campaign**. Review profiles, best-X listicles, and forum threads can carry the description of your business into an answer you never get to edit. If 43.8% of chatbot citations come from listicles, being left off three relevant ones deserves more attention than another pass at your H1.

![Small brand website podium overlooked beside a crowd of review sites and forums](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/27debf2c-ad04-4f9b-9f13-db47706c12cf.png)

### Five checks for Monday morning

Say you are spending $20k a month and a competitor appears in Perplexity while you do not. You do not need a new theory yet. You need to learn whether your page can be fetched, whether it offers a passage worth lifting, and whether anyone else vouches for you. This is the PPC-audit version of citation work: less folklore, more checks you can repeat.

1. **Allow the live fetchers you need.** Review robots.txt for OAI-SearchBot and ChatGPT-User even if you block GPTBot. Check PerplexityBot separately from Perplexity-User. Do not let a training decision silently become a search-visibility decision.
2. **Check whether AI can read your site.** Run [the fetch diagnostic before you rewrite a word](https://groas.com/post/before-you-write-another-page-for-chatgp). If a page blocks the fetcher, loads its answer through JavaScript the fetch does not execute, or returns junk to a text browser, better phrasing will not fix the access problem.
3. **Give each money page one quotable capsule.** Put a 40- to 80-word answer after an H2: definition, scope, then limit. It should make sense even when the surrounding page disappears.
4. **Work on three kinds of third-party proof.** Clean up a review profile, seek a place on a relevant best-X listicle, and participate where real buyers compare options. Your own copy cannot stand in for what those pages say about you.
5. **Track prompts, not just ranks.** Ask the five buyer questions you want to win and record who gets cited. For another take on the same problem, read the [sibling myth-by-myth debunking](https://groas.com/post/7-things-businesses-believe-about-gettin).

![Infographic with five numbered Monday-morning citation checks](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/05a9e8bc-a46a-4aac-ab9c-805e30ae7813.png)

If you come from paid search, this feels unfair. In Google Ads, you set a bid, win an auction, see an ad, and get a report on what it cost. Citation work has a different sequence: a fetcher has to reach material, retrieval has to find a useful passage, and the answer has to choose what to cite. Much of the material that can influence that choice sits on sites you do not own. **There is no organic-citation auction where a higher bid skips those steps.**

That lack of control is why this myth sticks. I could rewrite a landing-page headline and watch the result. I cannot change a listicle with a deploy or A/B test a Reddit thread. So it is easy to keep polishing the part you own and put off the harder work of being named elsewhere. The answer is not to neglect your site. Make it accessible, give it a clear passage worth quoting, then spend time on the places you cannot edit directly.

Nor is this a one-time audit. Fetch permissions drift when someone edits robots.txt. Capsules rot when marketing rewrites pages. Listicles update and drop you without telling you. groas monitors that kind of search visibility continuously, logs changes and their reasons, and pairs the execution with a human strategist who can judge what the pattern means for pipeline. If you want to buy attention, buy the Sponsored card and measure it honestly. If you want to be the source an answer cites, stop treating your homepage as the only place your reputation lives.

## Frequently Asked Questions

### If I buy an ad in ChatGPT, will my brand get cited in the answers?

No. Paid placements in ChatGPT run as clearly labeled Sponsored cards below the response, and OpenAI states that ads do not influence the answer itself. If you want to be cited organically, you need material the engine can retrieve and quote, and more budget does not replace that.

### Does ranking first on Google guarantee ChatGPT will cite my page?

No. In one test of 15 brands across 1,000 queries, Google page 1 brands appeared in ChatGPT only about 62% of the time, and a check of 15,000 prompts found only about 12% of cited URLs ranked in Google's top 10. Keep doing useful SEO work, but check the AI answers themselves rather than using your Google position as a stand-in.

### If I block GPTBot in robots.txt, does that remove me from ChatGPT search results?

Not necessarily. OpenAI uses separate fetchers: GPTBot collects training data, OAI-SearchBot surfaces sites in ChatGPT search answers, and ChatGPT-User fetches pages on demand. You can disallow GPTBot while allowing OAI-SearchBot, so decide what to permit bot by bot instead of treating it as one switch.

### Does publishing an llms.txt file improve my chances of being cited by AI answer engines?

There is no evidence it does. SE Ranking checked roughly 300,000 domains and found no statistically significant correlation between llms.txt and AI citation frequency, OtterlyAI's test found no visibility improvement, and Google confirmed in 2025 that it does not support llms.txt. Publish one if you want tidy documentation, but do not count it as a visibility win.

### Will adding schema markup make AI answer engines cite my pages?

Test results do not support that. A cross-platform test of 730 citations found schema on 43.1% of cited pages and 44.8% of non-cited pages, and an Ahrefs matched test of 1,885 pages adding JSON-LD showed changes from +2.4% to -4.6%. Keep correct, useful schema, but markup is not a substitute for an answer worth quoting.

### Do longer, more comprehensive pages get cited more by AI answer engines?

Not for citations. Answer engines can lift a single passage without rewarding the length around it, and an analysis of roughly 2 million sessions and 7,500 ChatGPT referrals found 72.4% of cited pages shared an answer capsule: a self-contained 40- to 80-word unit after an H2. Lead with the answer and give each section one clear job.

### Where do AI answer engines get most of the sources they cite about brands?

Mostly from sites other than the brand's own. An AirOps analysis of 21,311 brand mentions across ChatGPT, Claude, and Perplexity found 85% came from third-party domains versus 13.2% from brand domains, and Reddit accounted for a large share of Perplexity's top-10 citations. Review profiles, listicles, and forum threads deserve attention alongside your own site.

## Paid Ads & SEO, Run For You

Businesses pay one flat monthly fee per service, agencies resell it under their own brand, and groas does the work either way.

[See Pricing](https://groas.com/pricing)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","description":"groas is a done-for-you marketing service, not software: elite people working with proprietary technology systems run search ads, Meta ads, SEO and AI search for businesses, and white label for agencies, with a dedicated senior strategist owning every account.","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#webpage","url":"https://groas.com/post/7-things-people-believe-about-getting-ci","name":"7 Myths About Getting Cited by ChatGPT and Perplexity","description":"You can’t buy an organic citation, and a #1 Google ranking won’t guarantee one. Here are seven citation myths, the mechanisms behind them, and what to check instead.","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9ea767cd-d914-4b6e-8c6d-39a2bffd9db4.png"},"breadcrumb":{"@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#breadcrumb"},"dateModified":"2026-10-25T05:40:18.558Z","mainEntity":{"@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"7 Myths About Getting Cited by ChatGPT and Perplexity","item":"https://groas.com/post/7-things-people-believe-about-getting-ci"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#article","url":"https://groas.com/post/7-things-people-believe-about-getting-ci","mainEntityOfPage":{"@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"7 Myths About Getting Cited by ChatGPT and Perplexity","description":"You can’t buy an organic citation, and a #1 Google ranking won’t guarantee one. Here are seven citation myths, the mechanisms behind them, and what to check instead.","datePublished":"2026-10-25T05:40:18.497Z","dateModified":"2026-10-25T05:40:18.558Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/9ea767cd-d914-4b6e-8c6d-39a2bffd9db4.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"answer, page, chatgpt, citation, cited, google, about, perplexity","wordCount":1999,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"My background includes Goldman Sachs and Stanford Computer Science.","image":"https://groas.com/media/blog/4f17cac3a81acc224cc1b2cfab8f94bbc295a3089d13e8421125771ef8d4064f.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"},"sameAs":["https://www.linkedin.com/in/alexander-433793253/"]},{"@type":"FAQPage","@id":"https://groas.com/post/7-things-people-believe-about-getting-ci#faq","mainEntity":[{"@type":"Question","name":"If I buy an ad in ChatGPT, will my brand get cited in the answers?","acceptedAnswer":{"@type":"Answer","text":"No. Paid placements in ChatGPT run as clearly labeled Sponsored cards below the response, and OpenAI states that ads do not influence the answer itself. If you want to be cited organically, you need material the engine can retrieve and quote, and more budget does not replace that."}},{"@type":"Question","name":"Does ranking first on Google guarantee ChatGPT will cite my page?","acceptedAnswer":{"@type":"Answer","text":"No. In one test of 15 brands across 1,000 queries, Google page 1 brands appeared in ChatGPT only about 62% of the time, and a check of 15,000 prompts found only about 12% of cited URLs ranked in Google's top 10. Keep doing useful SEO work, but check the AI answers themselves rather than using your Google position as a stand-in."}},{"@type":"Question","name":"If I block GPTBot in robots.txt, does that remove me from ChatGPT search results?","acceptedAnswer":{"@type":"Answer","text":"Not necessarily. OpenAI uses separate fetchers: GPTBot collects training data, OAI-SearchBot surfaces sites in ChatGPT search answers, and ChatGPT-User fetches pages on demand. You can disallow GPTBot while allowing OAI-SearchBot, so decide what to permit bot by bot instead of treating it as one switch."}},{"@type":"Question","name":"Does publishing an llms.txt file improve my chances of being cited by AI answer engines?","acceptedAnswer":{"@type":"Answer","text":"There is no evidence it does. SE Ranking checked roughly 300,000 domains and found no statistically significant correlation between llms.txt and AI citation frequency, OtterlyAI's test found no visibility improvement, and Google confirmed in 2025 that it does not support llms.txt. Publish one if you want tidy documentation, but do not count it as a visibility win."}},{"@type":"Question","name":"Will adding schema markup make AI answer engines cite my pages?","acceptedAnswer":{"@type":"Answer","text":"Test results do not support that. A cross-platform test of 730 citations found schema on 43.1% of cited pages and 44.8% of non-cited pages, and an Ahrefs matched test of 1,885 pages adding JSON-LD showed changes from +2.4% to -4.6%. Keep correct, useful schema, but markup is not a substitute for an answer worth quoting."}},{"@type":"Question","name":"Do longer, more comprehensive pages get cited more by AI answer engines?","acceptedAnswer":{"@type":"Answer","text":"Not for citations. Answer engines can lift a single passage without rewarding the length around it, and an analysis of roughly 2 million sessions and 7,500 ChatGPT referrals found 72.4% of cited pages shared an answer capsule: a self-contained 40- to 80-word unit after an H2. Lead with the answer and give each section one clear job."}},{"@type":"Question","name":"Where do AI answer engines get most of the sources they cite about brands?","acceptedAnswer":{"@type":"Answer","text":"Mostly from sites other than the brand's own. An AirOps analysis of 21,311 brand mentions across ChatGPT, Claude, and Perplexity found 85% came from third-party domains versus 13.2% from brand domains, and Reddit accounted for a large share of Perplexity's top-10 citations. Review profiles, listicles, and forum threads deserve attention alongside your own site."}}]}]}
```