How to Get Your Website Cited in AI Overviews, ChatGPT and Perplexity
Key takeaways
- AI search tools like Google AI Overviews, ChatGPT search and Perplexity pull answers from pages they can crawl, then link to the ones they used. If they cannot reach your page, you cannot be cited.
- Google says there is no special trick for its AI features: normal SEO, a page that is indexed and eligible for a snippet, and genuinely useful content are what count.
- The biggest wins are simple: allow the right AI search crawlers, answer questions clearly, add first-hand experience, and track results in Search Console.
More people now get their first answer from an AI summary instead of a list of blue links. When that summary cites your page, you get a small but valuable link at the moment someone is actively looking for help. When it cites a competitor, you are invisible. This guide explains how AI search picks its sources and walks through practical steps to make your site a page it wants to cite.
How AI search chooses which pages to cite
Most AI search tools work in a similar way. They take the question, run one or more searches behind the scenes, read the best pages they find and write an answer based on them, with links back to the sources.
- Retrieval (grounding): Google describes its AI features as relying on its core ranking systems to fetch relevant, up-to-date pages from its index, then generating a response from them. In plain terms, you need to rank or at least be retrievable for the question first.
- Query fan-out: For complex questions, the AI runs several related searches at once. A question like “how to start a blog with AI” might also trigger searches for the best AI writing tools and how to set up WordPress. Each of those is another chance to be cited.
- Different engines, different indexes: Google uses its own Search index. ChatGPT search and Perplexity use their own crawlers plus other data sources. Being visible in all of them means being accessible to all of them.
Step 1: Make sure AI search can actually reach your pages
This is the step most sites get wrong without knowing it. A security plugin, a firewall rule or an old robots.txt line can quietly block AI crawlers and remove you from AI answers entirely.
Google AI Overviews and AI Mode
- Your page must be indexed and eligible to show a snippet in normal Google Search. A
noindexornosnippettag keeps it out. - Search Console now has a Search generative AI control under Settings. “Include” is the default, but it is worth checking that nobody on your team switched it to “Exclude”.
- Blocking
Google-Extendedin robots.txt limits AI training use. It does not remove you from AI Overviews, so you do not need to allow it to be cited.
ChatGPT, Perplexity and Claude
Each AI company runs separate bots for separate jobs. The key idea is that you can block training bots and still allow the search bots that decide whether you get cited.
| Bot | Company | What it does | Allow it to be cited? |
|---|---|---|---|
| OAI-SearchBot | OpenAI | Finds pages for ChatGPT search results | Yes |
| GPTBot | OpenAI | Collects content for model training | Optional |
| PerplexityBot | Perplexity | Builds the index behind cited answers | Yes |
| Claude-SearchBot | Anthropic | Indexes pages for Claude’s search answers | Yes |
| ClaudeBot | Anthropic | Collects content for model training | Optional |
If your robots.txt does not mention these bots at all, they are allowed by default. If you want to block training but keep AI search visibility, a setup like this works:
User-agent: GPTBot
Disallow: /
User-agent: ClaudeBot
Disallow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Claude-SearchBot
Allow: /Robots.txt is only half the check. Firewalls and CDN “bot fight” features can block these crawlers before they ever read your robots.txt, so review those settings too.
Step 2: Publish content an AI cannot write on its own
Google’s own guidance draws a line between commodity content and non-commodity content. A generic “10 tips” post restates what is already everywhere, and an AI can summarize it without needing your page. A post built on what you actually did, tested or measured gives the AI something it can only get from you.
- Share first-hand experience: “We tested three AI writing tools on the same brief” beats “here are three AI writing tools”.
- Add original numbers: Results from your own site, a small survey, before-and-after figures or real costs.
- Show real examples: Screenshots, prompts you used and the output they produced.
- Take a clear position: Say which option you recommend and why, instead of listing everything neutrally.
Using AI to help write is fine, as long as a person adds the experience and checks the facts. We cover this in more detail in Is AI content bad for SEO?
Step 3: Answer the question clearly and early
AI answers are built from passages that directly address the question. Make those passages easy to find, for readers first.
- Put a one or two sentence direct answer right under the heading, then expand on it.
- Use headings that match how people ask: “How long should a meta description be?” rather than “Length considerations”.
- Define terms the first time you use them.
- Keep one main idea per section so each part makes sense on its own.
This does not mean chopping your article into tiny fragments. Google has said there is no need to “chunk” content for AI. Clear writing helps humans and machines equally.
Step 4: Structure pages so they are easy to scan
Tables, numbered steps and short lists are easy for readers to scan and easy for AI to quote accurately. Use them where the content naturally fits that shape: comparisons, processes, prices, specifications and pros and cons. Add relevant images or video too, since AI results can show those as well as text links.
Step 5: Show who is behind the content
AI features rely on the same quality systems as regular search, so trust signals still matter:
- A named author with a short bio that explains their experience with the topic.
- A clear About page that says who runs the site and why.
- Links to the original sources for any facts and statistics you use.
- A visible “last updated” date, and genuine updates when things change.
Step 6: Earn real mentions across the web
AI answers often reflect what people say about a topic in reviews, forums, videos and other blogs. Being genuinely recommended in those places helps. Buying or faking mentions does not: Google specifically warns that chasing inauthentic mentions is not as helpful as it looks, and its spam systems filter a lot of it out. Focus on things people naturally share, like free tools, templates, original research and helpful answers in communities you already take part in.
Step 7: Measure what is working
- Google: Search Console has a Generative AI performance report showing how your content performs in AI Overviews, AI Mode and Discover.
- ChatGPT and Perplexity: In Google Analytics, look at referral traffic from sources such as chatgpt.com and perplexity.ai. Some AI traffic arrives without a referrer, so treat these numbers as a minimum.
- Manual checks: Once a month, ask your main questions in each tool and note who gets cited. It is a rough but useful way to spot gaps.
What you can safely skip
| Popular tactic | What the evidence says |
|---|---|
| Adding an llms.txt file | Google Search ignores it for rankings and AI features. Some AI tools may read it, but it will not boost Google visibility. |
| “Chunking” content for AI | Not needed. Write clear sections for readers. |
| Special AI schema | There is no special markup for AI features. Normal structured data is still useful for rich results. |
| Rewriting pages “for AI” | AI understands synonyms and meaning. Write for your audience. |
| Creating a page for every question variation | Mass-produced pages can break Google’s scaled content abuse policy. |
A quick checklist for WordPress sites
- In Settings > Reading, make sure “Discourage search engines from indexing this site” is unchecked.
- In your SEO plugin (such as Rank Math), check that important posts are not set to noindex or nosnippet.
- Open
yoursite.com/robots.txtand confirm the AI search bots above are not disallowed. Rank Math lets you edit it under General Settings. - Review security, caching and CDN settings for rules that block bots by default.
- Add an author box and keep your About page up to date.
- Check the Search generative AI control and the Generative AI performance report in Search Console.
For more ways to put AI to work on a WordPress site, see How to Use AI With WordPress.
Tip: Start with your five most important pages. Search their main question in Google, ChatGPT and Perplexity, note who is cited, then compare those pages with yours. The gap is usually missing first-hand detail or a buried answer, not a technical trick.
Frequently asked questions
How do I get my website cited in Google AI Overviews?
Make sure the page is indexed and eligible for a snippet, keep the Search generative AI control set to include your site, and focus on useful, original content that ranks for the question. Google says there are no extra requirements beyond normal SEO.
Should I block GPTBot to protect my content?
That is your choice. Blocking GPTBot limits training use, while allowing OAI-SearchBot keeps you eligible for citations in ChatGPT search. The two settings are independent.
Does llms.txt help me appear in AI answers?
Not in Google. Google has confirmed its Search and AI features ignore llms.txt. Some other AI tools may read it, so it does no harm, but do not expect it to improve rankings.
How long does it take to show up in AI search?
There is no fixed timeline. Changes to robots.txt are usually picked up within about a day, but being cited depends on your page being found, trusted and relevant, which often takes weeks or months.
Is GEO different from SEO?
Google treats optimizing for its AI features as part of SEO. For other AI tools, the main extra step is making sure their search crawlers can access your site.
