How to Get Cited by ChatGPT (and Perplexity, Gemini, Claude)

Citations from ChatGPT, Perplexity, Gemini, and Claude follow a small set of on-page signals plus dedicated crawler access. I am running the playbook below on ignax.dev itself; the live experiment is documented in the growth engine case study. For the conceptual primer, see what is AEO.

I am Ignacio (IGNAX), Spain-born and Paraguay-based, a solo full-stack developer and SEO/AEO specialist. Full service page at SEO + AEO setup.

What is the short checklist?

If you ship nothing else, ship these 10 items in this order:

1. Quick Answer block (40–60 words) in the first 200 pixels of every key page.
2. All H2s phrased as questions on every long-form page.
3. FAQPage JSON-LD with 5–8 Q/As per page, validated in Google's Rich Results Test.
4. Article JSON-LD with author, datePublished, dateModified, headline.
5. Service JSON-LD on every service page with provider, areaServed, offers.
6. Person JSON-LD with a full sameAs array (LinkedIn, GitHub, X, Crunchbase, ORCID).
7. Organization JSON-LD with sameAs and logo.
8. llms.txt + llms-full.txt at the root, regenerated on every build.
9. IndexNow integration pinging Bing/Yandex on every publish.
10. robots.txt explicitly allowing AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) and referencing llms.txt.

That is the entire on-page AEO surface. The rest of this article expands each item.

How do AI engines actually find pages to cite?

Three pipelines feed citation candidates:

• Search index reuse. Gemini queries Google's index. ChatGPT browses Bing plus its own GPTBot crawler. Claude uses Brave Search plus ClaudeBot. Perplexity uses PerplexityBot plus partner data. If you are not indexed in Google and Bing, you are invisible to the entire stack.
• Direct crawl. GPTBot, ClaudeBot, PerplexityBot, and Google-Extended crawl independently of the search indexes. The robots.txt and llms.txt files control what they see.
• Trusted entity graphs. When the model is asked about a specific person or company. It disambiguates via sameAs schema linking to authoritative profiles. Without sameAs you are an undifferentiated string; with it you are an entity the model can confidently attribute.

The implication: AEO is search-engine indexing + dedicated crawler signals + entity verification. All three layers need work.

How do I write a Quick Answer block?

A Quick Answer is the single highest-leverage piece of AEO content:

• 40 to 60 words, ~150 to 450 characters.
• Stand-alone. No "this article", no "here is", no "we will explain". Lead with the topic: "AEO is…", "A SaaS MVP costs…".
• Direct factual answer to the page's target query.
• In the first 200 pixels of the rendered page.
• No promotional language. The block exists to be quoted, not to sell.

Example, from this site's seo-aeo-setup page:

AEO setup turns your site into something AI engines (ChatGPT, Perplexity, Gemini, Claude) can quote. I ship FAQPage and Service schema, Quick Answer blocks, llms.txt and llms-full.txt, IndexNow ping, and a hreflang-clean sitemap. Five to ten days for $1,800–$4,500 USD. EN/ES bilingual.

That is 53 words, stand-alone, factual, with the price band built in. A user asking ChatGPT "what does AEO setup cost" can be answered with that paragraph verbatim.

What does FAQPage schema look like?

FAQPage is the schema that AI engines lift most readily. The minimum valid block:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is AEO?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "AEO is the practice of optimizing your site to be quoted by AI engines like ChatGPT, Perplexity, Gemini, and Claude."
      }
    }
  ]
}
</script>

Constraints I apply on every page:

• 5 to 8 questions per page (more dilutes the signal).
• Each text answer ≤80 words.
• Questions phrased the way a real user would type them, not the way a copywriter would.
• Validated in the Google Rich Results Test before shipping.

The questions and answers can be authored in front-matter and rendered both as visible UI and as JSON-LD; that is the SvelteKit content registry pattern I default to.

What does Person + sameAs schema look like?

The sameAs array is how AI engines verify you are a real entity and not a name collision:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Person",
  "name": "Ignacio (IGNAX)",
  "url": "https://ignax.dev/about",
  "jobTitle": "Freelance Full-Stack Developer & SEO/AEO Specialist",
  "sameAs": [
    "https://www.linkedin.com/in/ignaxdev",
    "https://github.com/ignax",
    "https://x.com/ignaxdev"
  ]
}
</script>

The more sameAs URLs the better, LinkedIn, GitHub, X, Crunchbase, ORCID, Mastodon, personal blogs, podcast guest pages, conference speaker pages all count. The entity becomes more confidently identifiable as the graph thickens.

What goes in robots.txt for AEO?

The minimum robots.txt for AEO in 2026:

User-agent: *
Allow: /

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Google-Extended
Allow: /

Sitemap: https://yourdomain.com/sitemap.xml

Add llms.txt reference if your aggregator supports it (the spec is still evolving on this). Block bots you do not want via explicit Disallow rules; the default User-agent: * covers the long tail.

For the IndexNow piece see IndexNow setup tutorial: it is the publish-time complement to robots.txt.

What does llms.txt look like?

A working llms.txt for a small site:

# Ignax.dev

> Solo full-stack developer and SEO/AEO specialist. SaaS MVPs, RAG chatbots, multilingual SEO. Spain-born, Paraguay-based.

## Services

- [SaaS MVP Development](https://ignax.dev/services/saas-mvp): Production SaaS MVPs in 2–6 weeks
- [SEO + AEO Setup](https://ignax.dev/services/seo-aeo-setup): Get cited by ChatGPT, Perplexity, Gemini, Claude
- [RAG Chatbots](https://ignax.dev/services/rag-chatbots): Production retrieval-augmented chatbots

## Articles

- [What is AEO?](https://ignax.dev/articles/what-is-aeo): Answer Engine Optimization explained
- [How to Get Cited by ChatGPT](https://ignax.dev/articles/how-to-get-cited-by-chatgpt): The concrete checklist
- [llms.txt Explained](https://ignax.dev/articles/llms-txt-explained): What it is and how to ship it

## Case Studies

- [Growth Engine](https://ignax.dev/case-studies/ignax-dev-growth-engine): The live AEO experiment

The companion file llms-full.txt contains the full markdown content of every linked page concatenated together, so an AI agent can read the whole site without hitting your sitemap. Both files are auto-generated from the content registry on every build.

For the deeper explainer see llms.txt explained.

How do I get the first citation in 90 days?

From a clean install, the realistic timeline:

Day	Milestone
0	AEO setup shipped (all 10 checklist items)
1–7	Indexing in Google + Bing complete via IndexNow
14–30	Rankings appear on long-tail keywords
30–60	First AI-engine citations on niche queries
60–90	Citations on competitive buyer-intent queries
90+	Compounding, citations beget more citations

The 30–60 day citation window is real. I have seen it across multiple bilingual SvelteKit sites I have shipped. The variable is how niche the target prompt is. A prompt like "best Paraguay-based bilingual SaaS developer" cites within weeks because the topical authority graph is sparse. A prompt like "best SaaS developer" takes much longer because the graph is dense.

How do I track AI citations?

There is no GSC equivalent. The manual probe sheet I maintain:

• 10 to 20 target prompts in EN and ES, written as a real buyer would type them.
• Run weekly across ChatGPT (with browsing on), Perplexity, Gemini, and Claude.
• Log each result: which engines cited me, which URL, which competitor was cited instead.
• Track citation deltas month over month.

I update this against my own site monthly and publish it in the growth engine case study. The full checklist for site-readiness lives in AEO citation checklist.

What are the best external references?

• Google Search Central, structured data documentation: canonical JSON-LD reference.
• Schema.org FAQPage type; the spec for FAQ schema.
• llmstxt.org; the llms.txt proposed standard.
• IndexNow protocol; the publish-ping API spec.
• OpenAI GPTBot documentation: how ChatGPT's crawler behaves.

Ship the 10-item checklist, run the probe weekly, and you will see citations inside 90 days. There is no magic, there is only the work.

Ready to ship the playbook on your site? Email hello@ignax.dev with your URL. I crawl the site before our call.