# How do you get your website cited by AI search?

Date: 2026-08-19 · Author: Robert Scales · https://guidetoartificialintelligence.com/en/learn/get-website-cited-ai-search/

**Answer:** Make sure AI search crawlers are allowed to read your pages, keep each page indexable in ordinary search, and write pages that answer one question clearly in text, with facts a reader can check. No file or markup guarantees a citation; Google says its AI features need nothing beyond good SEO.

AI search tools such as Google's AI Overviews and AI Mode, ChatGPT search and Perplexity answer a question in their own words and add links to the pages they drew on. Those links are the **citations**, and for many sites they are becoming a real source of visitors. Not surprisingly, a small industry now sells ways to "optimize for AI". Before buying anything, it helps to read what the search providers themselves publish, because it is short and fairly clear.

The core message is reassuring for anyone who already runs a decent site. AI search still starts with ordinary search: a crawler has to be allowed in, the page has to be indexed, and the content has to answer what people ask.

## What the providers actually say

**Google.** Its Search Central guide on AI features says the usual search best practices apply to AI Overviews and AI Mode, with no extra requirements and no special optimization. A page can be shown as a supporting link only if Google has indexed it and could display it in normal results with a snippet, the short text excerpt under a result. Google adds that meeting every requirement still does not guarantee a page will be shown. It also says plainly that no special AI text file, extra machine-readable file or dedicated schema.org markup is needed for these features. Clicks from them are counted in Search Console's Performance report, under the Web search type.

Google describes a technique it calls **query fan-out**: the system runs several related searches on subtopics to build one answer. In practice, a page that answers a narrow sub-question well can be cited even if it does not rank for the main query.

**OpenAI.** ChatGPT search relies on a crawler called **OAI-SearchBot**. OpenAI says that if a site blocks it, ChatGPT search will leave that site out of its answers, apart from plain navigational links. A separate crawler, **GPTBot**, collects content that may be used to train OpenAI's models. The two are controlled independently, so you can welcome search and refuse training. OpenAI says changes to robots.txt can take about a day to take effect.

## A worked checklist for one page

Pick one important page, such as your opening hours or a guide you are proud of, and go through it:

| Check | How to verify | Fix if it fails |
|---|---|---|
| Crawlers allowed | Read `yoursite.com/robots.txt`; ask your host whether a firewall or CDN blocks bots | Allow Googlebot and OAI-SearchBot |
| Page is indexed | Search Console's URL inspection, or search `site:` plus the page address | Remove stray `noindex`, fix canonical tags |
| Answer in text | View the page with images off | Put key facts in text, not only in pictures or PDFs |
| Question answered early | Read the first two paragraphs alone | Lead with the direct answer |
| Structured data matches | Compare your schema markup with what the page shows | Correct or remove mismatched data |
| Internal links | Count links to the page from your other pages | Link it from related pages and menus |

A robots.txt that allows search crawlers but declines training could look like this:

```text
User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: *
Allow: /
```

Whether to block training is your choice; it does not affect ChatGPT search, according to OpenAI.

> **Try it:** Ask an AI search tool the question your page answers best, worded the way a customer would. Note which sites are cited, then run your page through the checklist above and fix the first failing row.

## Where llms.txt fits

You may have heard of **llms.txt**, a proposal by Jeremy Howard first published in 2024 and revised this month. It suggests a short Markdown file at `/llms.txt` with a title, a one-paragraph summary and lists of links to clean, text-only versions of your key pages. Its purpose is to help AI agents find the right information on demand, for example a coding assistant reading software documentation.

That is a different job from search ranking. Google's guidance says such files are not needed for its AI features. An llms.txt file is cheap to write and may help agents that look for one, but treat it as a courtesy to those tools, not as a way to earn citations.

## What makes a page easy for AI to quote?

A page is easy to quote when a single passage can stand on its own and still be correct. AI answers usually attach a link to a specific claim, so a passage that depends on a heading three screens up, or on a table rendered as an image, is hard to use. Write paragraphs that name their subject ("Our Montebello store opens at 9 a.m. on weekdays") instead of relying on "it" and "here".

Put the answer first and the nuance after. If your page is titled as a question, the first paragraph should answer it in one or two sentences, then explain. That is the same habit that makes a page useful to a hurried human reader, which is no coincidence.

Make facts checkable. Give dates, units, prices and names exactly, say when the page was last updated, and link to your own primary sources where you have them. A clear, specific statement is easier to cite than a vague one, and a page that contradicts itself or hides key facts in a downloadable file gives a system little to work with.

Finally, keep expectations realistic. No provider promises that following its advice leads to citations, and answers change from one query to the next. What you control is whether your page can be found, read and trusted; the rest is decided by the question someone asks.

## Sources

- AI features and your website: https://developers.google.com/search/docs/appearance/ai-features
- Overview of OpenAI Crawlers: https://platform.openai.com/docs/bots
- The /llms.txt file, v2: https://llmstxt.org/