LLM SEO: How to Get Your Site Into ChatGPT, Claude and Perplexity Answers

What LLM SEO means

A large language model (LLM) is the engine behind assistants like ChatGPT, Claude, Perplexity, Gemini and Copilot. When someone asks one of them "which tool should I use for X" or "how do I fix Y", the answer often names websites and links to them. LLM SEO, also called generative engine optimization (GEO) or answer engine optimization (AEO), is everything you do so that your site is one of those names.

It overlaps heavily with classic search engine optimization. The difference is where the reader is: inside a chat answer instead of on a results page. That changes three things: which crawlers have to reach you, how your pages need to be written to be quoted, and how you measure the result.

How AI assistants find your pages

An assistant can know about your site in two ways, and each one is fed by different crawlers (programs that download web pages).

The live paths are the ones that send visitors, because they come with links. Training is slower and indirect: it shapes what the model "remembers" about your brand. You can allow both, or allow search and block training. The companies treat them as separate switches: OpenAI says each of its robots.txt settings "is independent of the others", and Google says Google-Extended "does not impact a site's inclusion in Google Search".

The LLM SEO checklist

1. Let the AI search crawlers in

The most common reason a site never appears in AI answers is that it blocks the crawlers, often without the owner knowing. A robots.txt file written years ago, a security plugin, or a CDN (content delivery network) setting can all do it. Open yoursite.com/robots.txt and look for Disallow: / under any AI bot, or under User-agent: *.

A robots.txt that lets AI search and answer engines read the site, while keeping it out of model training, looks like this:

# Let AI search and answer engines read the site
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: Claude-SearchBot
User-agent: Claude-User
User-agent: PerplexityBot
User-agent: Perplexity-User
Allow: /

# Optional: keep the content out of AI model training
User-agent: GPTBot
User-agent: ClaudeBot
User-agent: Google-Extended
User-agent: CCBot
Disallow: /

If you are happy for your content to be used in training too, delete the second group. Changes are not instant: OpenAI and Perplexity both say it can take about 24 hours for their systems to pick up a new robots.txt.

robots.txt is only half of it. A firewall or CDN can refuse AI crawlers before robots.txt is even read, and the owner never sees it. The only way to know is to request the page the way each bot does. The free AI visibility checker does exactly that for ten AI crawlers.

2. Make sure the content is in the HTML

Many AI crawlers download the page and read the HTML without running JavaScript. If your text, prices or product names only appear after scripts run, the crawler may see an empty shell. Check by opening view-source: on your page in a browser: the words you want quoted should be there. Server-side rendering or static generation solves this for most frameworks.

3. Answer the question in the first lines

Assistants quote passages, not whole pages. A page that states the answer plainly near the top, under a heading that matches the question, is easier to quote than one that builds up to it. For each important page, write down the question a buyer would ask, put that question (or close to it) in a heading, and answer it in one or two sentences right below.

4. Be specific

Specific facts get repeated; general claims do not. Prices, plan limits, supported platforms, numbers, dates and comparisons are what an assistant needs to answer "which one should I pick". If a fact about your product only lives in a sales call or a PDF, it does not exist for an AI assistant.

5. Describe the page with structured data

Structured data is a small block of JSON in the page (JSON-LD, using the schema.org vocabulary) that says what the page is: an organization, a product with a price, an article with a date, a list of questions and answers. Search engines use it to understand pages, and the AI features built on top of those search engines inherit that understanding. Start with Organization on the homepage, Product or SoftwareApplication on product pages, Article on guides and FAQPage where you answer questions.

6. Be in the search indexes the assistants use

Live answers are only as good as the index behind them. Google AI Overviews use Googlebot, so the usual Google Search Console work applies: pages indexed, sitemap submitted, no accidental noindex. Copilot is built on Bing, so verify the site in Bing Webmaster Tools and submit your sitemap there too. ChatGPT, Claude and Perplexity run their own crawlers, which is why step 1 matters for them.

7. Add an llms.txt file

llms.txt is a proposed standard (llmstxt.org): a short Markdown file at the root of the site that tells language models what the site is and links its most useful pages. None of the large assistants has said it relies on the file, so treat it as cheap insurance rather than a ranking factor. It takes ten minutes to write and it also helps tools that build answers from your docs.

How to measure LLM SEO

Rankings in AI answers are not published anywhere, so you measure the two things you can see on your own site:

• Visitors sent by assistants. ChatGPT adds utm_source=chatgpt.com to the links it shows, and visits from Perplexity, Claude, Gemini and Copilot arrive with those sites as the referrer. Filter your analytics by them.
• AI crawler visits. Your server logs show which bots fetched which pages. A page that OAI-SearchBot or PerplexityBot never fetches cannot be quoted by them. ChatGPT-User and Perplexity-User visits mean someone asked an assistant about that page.

Watch both over months, not days. A single answer can change from one hour to the next.

Common mistakes

• Blocking GPTBot and thinking it removes you from ChatGPT search. It does not: search uses OAI-SearchBot, a separate switch.
• Blocking every AI bot with one User-agent: * rule meant for something else.
• Relying on JavaScript to show the main content.
• Writing pages that never say plainly what the product does, who it is for and what it costs.
• Treating llms.txt as a replacement for crawlable pages. It is a summary, not a substitute.

Doing this without the busywork

Every step above is a check you can repeat: are the AI crawlers allowed, does the HTML carry the content, is the structured data valid, does llms.txt exist. VisibleReady runs those checks on your site every day, writes the fix in your code as a pull request you approve, and logs which AI crawlers and assistants visit afterwards. Start with the free check below; it takes about 30 seconds.