← All posts

How to Get Cited by ChatGPT and Perplexity (Without Being Already Famous)

The practical AU guide to LLM citations. Why most small businesses are invisible to AI, and the technical and content fixes that change it.

TL;DR Most small Australian businesses aren't invisible to AI because they're not famous. They're invisible because of fixable technical and structural problems. The technical access layer (robots.txt, llms.txt) takes an hour. The content structure (answer-first paragraphs, schema, FAQ sections that mirror real questions) takes a week. The third-party signals (Reddit, YouTube, industry directories) take longer but pay off the most. Start with the 7-day plan at the bottom of this post.

Why your customers are seeing your competitors in ChatGPT

A customer opens ChatGPT, types in your category and your city, and asks for a recommendation. The AI gives them three businesses. None of them are yours.

The reason is usually more boring than you think. The LLM can't find enough about you to confidently cite you. This article covers why that happens and how to fix it.

I'll skip the persuasion phase. If you're reading this, you already know LLM SEO matters. What you want is the mechanics.

How LLMs actually pick what to cite

LLMs work through something called retrieval-augmented generation (RAG). Here's what that means in plain English.

When a user asks a question, the model breaks the prompt into sub-queries. It searches each one across its allowed sources (the open web, Reddit, Wikipedia, news, depending on the model). It retrieves the most relevant passages, scores them for clarity and authority, and stitches together an answer with citations.

The important bit: LLMs cite passages, not pages. A single paragraph from your site might be cited even if the rest of the page is ignored. This is why pages outside Google's top 10 can still earn citations. Rank doesn't determine citation. Quotability does.

The technical access problem (the 5-minute fix that solves it for most businesses)

The most common reason a business gets zero citations is the most boring one: AI crawlers can't access the site. The site's robots.txt is blocking them.

Open your robots.txt (yoursite.com.au/robots.txt) and check whether these bots are allowed:

User-agent: GPTBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: GoogleOther
Allow: /

If any of these are blocked (either explicitly or because your robots.txt has a blanket Disallow: /), no amount of content work will help. The AI literally can't read your site.

The next layer is llms.txt. It's an emerging standard that lives at yoursite.com.au/llms.txt or yoursite.com.au/.well-known/llms.txt. It's a plain language file telling AI models who you are, what you do, the key facts about your business, and where to find your most important pages.

Think of it as a fallback document. When an AI model can't piece together a clean story about your business from scattered sources around the web, llms.txt is the source of truth it can use. It's optional, but it reduces the risk of the AI getting your business wrong, and it gives you a small lift in citation rates.

The content structure problem

Once the AI can read your site, the next question is whether your content is shaped in a way the AI can lift. Six things make a passage easy to cite.

  • Answer-first paragraphs. Under every H2 on your page, the first 1 to 2 sentences should be a standalone, quotable answer to the heading. AI extracts these passages first. Bury the answer three paragraphs in and you lose.
  • HTML tables for comparisons. AI extracts HTML tables almost verbatim. If you're comparing options, features or pricing, use a table. Don't write the comparison as prose.
  • FAQ sections that mirror real prompts. Write FAQs as the actual questions your customers would type into ChatGPT. Not "What are the benefits of our service" but "How much does X cost in Sydney" or "Is Y safe for kids under 5".
  • Clear H2/H3 hierarchy. Each section should stand alone. The AI may lift one section without context, so don't write paragraphs that depend on the section before them to make sense.
  • Specific numbers, dates and sources. Replace "many clients" with "47 clients since 2019". Replace "research shows" with "a 2024 Princeton study found". AI rewards specificity.
  • Schema markup. FAQPage, Article, Organization and Person schema that matches your on-page content. Schema tells the AI what kind of content this is and who it's about. Without it, the AI is guessing.

The signal problem (third-party mentions)

Here's the strong claim: where your brand appears across the web often matters more than how well any single page on your site is optimised. Different LLMs lean on different sources, and the citation patterns vary by platform.

PlatformWhere citations come fromWhat to do for AU businesses
ChatGPTMajor publications, Wikipedia, Reddit, your own websiteGet featured in a relevant Australian industry publication. Maintain a complete, accurate About page.
PerplexityReddit (~47% of citations), YouTube (~14%), review platforms, newsEngage in relevant Australian subreddits. Maintain complete profiles on G2, Trustpilot and Capterra for B2B, or category directories for B2C.
Google AI ModeMixed (first-party content, Reddit, YouTube, structured data)Strong technical SEO. FAQPage schema. Active Google Business Profile. Reddit presence in relevant subs.
ClaudeMajor publications, academic sources, brand-owned contentLong-form, well-cited content on your own site. Get cited in industry research or analyst reports.

The practical version for an Australian SME:

  • One well-written comment in a relevant subreddit (genuine answer, no promotion) can move the needle on Perplexity citations
  • A feature in a local industry publication moves ChatGPT and Google AI Mode
  • A HARO or Source of Sources quote in a major outlet lifts visibility across all four platforms
  • Complete profiles on G2, Trustpilot and Capterra (B2B) or local directories (B2C) are baseline hygiene

None of this requires being famous. It requires being provable.

Why YouTube matters more than you'd think

Recent research suggests YouTube mentions correlate more strongly with AI visibility than backlinks do. The reason is structural, not magical: YouTube content comes with transcripts, descriptions and chapter markers. That's dense, citable text in a format LLMs love.

For a small Australian business, the practical implication is simple. One short tutorial video on your most-asked customer question, transcribed and uploaded with a clean description, can move your citation rate more than another blog post.

If you've never used video before, start with one. Five minutes long. You speaking to camera. Answer one specific question your customers always ask. Upload it. Make sure auto-captions are clean. That's the whole thing.

A 7-day starter plan

Each step is one focused action. None should take more than 60 minutes.

  • Day 1: Audit robots.txt. Check whether GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot and GoogleOther are allowed. Unblock any that aren't.
  • Day 2: Write and publish llms.txt. Cover who you are, what you do, your key facts, your service areas, and links to your most important pages.
  • Day 3: Rewrite your homepage and top service page with answer-first paragraphs. The first 1 to 2 sentences under every H2 should be a quotable, standalone answer.
  • Day 4: Add Organization, Person and FAQPage schema sitewide. If you're using WordPress, plugins like Rank Math or Yoast handle most of this. Otherwise, use JSON-LD directly.
  • Day 5: Post one substantive comment in a relevant Australian subreddit. Genuine answer to a real question. No promotion. No link to your site. Just expertise.
  • Day 6: Claim or refresh your G2, Trustpilot or industry directory profiles. Make sure your name, address, phone number and description are consistent across all of them.
  • Day 7: Run five customer-style prompts through ChatGPT, Perplexity and Google AI Mode. Log whether you're mentioned, what's said, and what's missing. This is your baseline.

A week from now, you'll have done more for your AI visibility than 90% of your competitors.

So where does this leave you?

The brands earning AI citations now aren't always the biggest. They're the most provable. Clear content. Clean technical setup. Real signals from third-party sources. The work isn't glamorous and it isn't fast, but it compounds.

If you'd like a free AI visibility audit where I'll run your business through ChatGPT, Perplexity and Google AI Mode and tell you what's working and what's missing, send me a message. No charge, no pitch.

Frequently asked questions

How do LLMs decide which sites to cite?

LLMs break a user's prompt into sub-queries, retrieve relevant passages from their allowed sources, score those passages for clarity and authority, and synthesise an answer with citations. The key thing to understand is that they cite passages, not whole pages, so quotable, answer-first content matters more than overall page rank.

Do I need to be famous or have lots of backlinks to get cited?

No. Most small businesses are missing citations because of fixable technical and structural problems, not because of brand size. Clean robots.txt, llms.txt, answer-first content, schema markup, and a handful of authentic third-party mentions (one Reddit comment, one industry directory, one HARO quote) move the needle for businesses with little to no name recognition.

What is an llms.txt file and do I need one?

llms.txt is an emerging standard. It's a plain language file at yoursite.com.au/llms.txt that tells AI models who you are, what you do and where to find your important pages. It's optional but useful. It reduces the chance of AI getting your business wrong and gives you a small lift in citation accuracy. Worth setting up if you're serious about AI visibility.

Does Reddit really matter for AI citations?

Yes, especially for Perplexity. Around 47% of Perplexity citations come from Reddit, and ChatGPT pulls from it heavily too. One well-written, genuinely helpful comment in a relevant subreddit (no promotion, no link) can move the needle. Multiple comments over time build a footprint LLMs can lean on.

Why does Perplexity cite me but ChatGPT does not?

Different platforms use different source weightings. Perplexity leans heavily on Reddit and YouTube. ChatGPT leans more on major publications, Wikipedia and your own website. If you're appearing in one but not the other, it usually means your third-party signals are skewed toward one platform's preferred sources.

How long does it take to start appearing in AI answers?

The technical fixes (robots.txt, llms.txt, schema) can lift citation rates within weeks. The content rewrites take 4 to 8 weeks to fully take effect as the AI re-crawls and re-scores your pages. Third-party signals (Reddit, YouTube, industry features) compound over months. Realistic expectation: visible movement in 30 to 60 days, meaningful movement at 3 to 6 months.

Can I pay to be cited by ChatGPT or Perplexity?

No. None of the major LLMs offer paid citation placement. Anyone selling "guaranteed ChatGPT citations" is selling air. The only ways to earn citations are technical access, content structure and real third-party signals.