# SEO for LLMs: how to get cited by AI assistants

> Large language models answer from two places: what they learned in training and what they fetch live. You influence the second far more than the first. This guide covers the practical parts: access, wording, and proof.

- Page: https://liftup.tech/guides/seo-for-llms/
- Type: Guide
- Updated: 2026-11-20

## In short

SEO for LLMs means making your pages reachable by AI crawlers, easy to quote, and credible. Allow the bots you want in robots.txt, serve readable text without relying on scripts, put clear answers in self-contained passages, state facts with sources, and earn mentions on trusted sites. Nobody can promise a citation, so measure by testing prompts.

## How models find your content

There are three routes. Training: the model learned from a snapshot of the web, and you cannot change that quickly. Live retrieval: assistants such as search-connected chatbots fetch pages when a question needs fresh facts, usually through a search index plus their own crawler. User-triggered fetches: a person pastes a link, or asks the assistant to open a page.

Most of what you can do concerns the last two. Pages that are indexed, crawlable and easy to extract answers from have a better chance of being retrieved. Pages that are blocked, slow, thin or hidden behind scripts usually do not.

## The crawlers to know

Names and behavior change, so confirm each one in the vendor's own documentation before editing rules. These are the common ones at the time of writing.

| User agent | Run by | Typical purpose | Blocking it affects |
|---|---|---|---|
| GPTBot | OpenAI | Collects content that may be used to train models | Training use, not live search results |
| OAI-SearchBot | OpenAI | Finds pages for search features in ChatGPT | Appearing in ChatGPT search |
| ChatGPT-User | OpenAI | Fetches a page when a user asks | Pages opened on request |
| ClaudeBot | Anthropic | Collects content that may be used for training | Training use |
| PerplexityBot | Perplexity | Indexes pages for its answer engine | Appearing in Perplexity results |
| Google-Extended | Google | A control token for use of content in Gemini models; it does not change Search | Gemini training and grounding use |
| Googlebot | Google | Crawls for Search, which also feeds AI Overviews | Search and AI Overviews |
| Bingbot | Microsoft | Crawls for Bing, which feeds Copilot | Bing and Copilot |

Block training crawlers and you can still be cited in live answers, because those use different agents. Block the search agents and you remove yourself from those answers. Decide per bot, deliberately.

## Steps to make a page citable

1. **Check access**: Read robots.txt and test as each bot. Also check the firewall, CDN bot rules and rate limits. Many sites block AI crawlers by accident through a security setting.
2. **Serve the text in the HTML**: Fetch the page without running scripts. If the answer is not in the raw HTML, many crawlers will not see it. Server-side rendering fixes this on JavaScript sites.
3. **Write self-contained passages**: Models lift paragraphs, not pages. Each section should make sense alone: state the subject, then the answer, in two or three sentences.
4. **State facts plainly with a source**: Dates, numbers and definitions should be exact and attributed. Vague promotional text is rarely quoted.
5. **Name the entity clearly**: Use the same business name, address details and descriptions on your site and profiles. Add Organization or LocalBusiness schema so the facts are explicit.
6. **Earn mentions elsewhere**: AI answers lean on sources they trust: industry publications, reviews, directories and well-known reference sites.
7. **Refresh dated pages**: Models that retrieve live content prefer current facts. Update the date only when the content really changed.

## What about llms.txt?

An llms.txt file is a plain text file at the root of a site that lists key pages in a simple format for language models. It is a proposal, not a standard. Google has said it does not use the file for Search, and the major assistants have not confirmed that they rely on it. Adding one is cheap and does no harm, but it is not a substitute for crawlable pages. Our [llms.txt generator](https://liftup.tech/tools/llms-txt-generator/) creates a draft.

## How to test whether it is working

- Write twenty questions your customers ask. Run them in the main assistants each month.
- Record whether your brand is named, whether your page is a cited source, and whether the facts are correct.
- Check analytics for referral visits from assistants.
- Look at server logs for the bots above to confirm they fetch your pages.
- Compare with Search Console. Pages that gain impressions often gain citations later.

Answers vary between runs and users, so look for patterns across several runs, not a single result. See the [AI visibility checker](https://liftup.tech/tools/ai-visibility-checker/) and our [GEO checklist](https://liftup.tech/checklists/geo-checklist/) for a fuller routine.

## Pitfalls

- Blocking every AI bot out of caution, then wondering why you never appear.
- Allowing all of them without thought about training use of your content.
- Hiding key facts in images, tabs or pop-ups.
- Keyword-stuffed pages written for machines. Models and search systems read for meaning.
- Believing a service that promises mentions in ChatGPT.

We cover this as part of [generative engine optimization](https://liftup.tech/services/generative-engine-optimization/). You get a fixed quote after a short call on the [contact page](https://liftup.tech/contact/).

## Questions, answered

### Can I make ChatGPT recommend my business?

You cannot force it. You can allow its search crawler, publish clear pages, keep your facts consistent and earn mentions on sources it trusts, which raises the chance of appearing.

### Should I block GPTBot?

That is a choice about training use of your content. Blocking GPTBot does not stop live search by OAI-SearchBot, which is a separate agent. Check OpenAI's current documentation.

### Does llms.txt help SEO?

Google has said it does not use it for Search. It may help some tools find key pages, but it is optional and unproven.

### Is SEO for LLMs different from normal SEO?

Much of it is the same: crawlability, quality, trust. The differences are an extra set of bots, a stress on quotable passages and a focus on mentions.

### How do I know if an AI assistant visited my site?

Check server or CDN logs for the user agent names. Analytics show only visitors who clicked through.

## Related

- [GEO checklist: earn citations in AI answers](https://liftup.tech/checklists/geo-checklist/)
- [llms.txt generator](https://liftup.tech/tools/llms-txt-generator/)
- [Generative engine optimization (GEO): be the name models recommend](https://liftup.tech/services/generative-engine-optimization/)

---

Get a fixed quote: https://liftup.tech/contact/ · email rohit@liftup.tech · WhatsApp +91 77689 10234 (https://wa.me/917768910234) · reply within one working day.
Source page: https://liftup.tech/guides/seo-for-llms/ (Liftup Technologies)
