Website GEO: How to Get Your Pages Cited by AI Search Engines

GEO for Websites: How to Get Your Pages Cited by AI and LLMs

Key takeaways

• Client-side rendering is killing website visibility. Most AI crawlers don’t run JavaScript, so if your website platform doesn’t support server-side-rendering, you’re almost invisible to AI search engines.
• A page can rank #1 on Google and still never get cited by AI. Google renders JavaScript. The AI crawlers don’t. Server-side rendering closes that gap.
• Generative engine optimization in 2026 starts as a technical job, not a writing one. If the crawler can’t read the page, no amount of answer engine optimization copy saves it.
• To get quoted, give engines something quotable. Clear headings, stats, and JSON-LD in the initial HTML all raise your odds of showing up in AI Overviews and chat answers.

AI has made many things simpler. Getting traffic to your website, though? AI threw a spanner in the works there.

A page might be number one on Google, but get zero traffic from any LLM like ChatGPT or Claude. And with over millions of people now searching for products and services through AI, that’s a big problem.

Here’s what’s going on in a nutshell.

Firstly, the ways in which AI crawlers “grade” a website or web page is still being decided by the AI companies themselves. Reddit, YouTube, Wikipedia, Medium, and mentions of your company’s name across the internet in general all play a greater role in GEO than they do in traditional SEO. But the exact science behind it is still up in the air.

But one thing we know for sure is, many websites are being ignored by LLMs right now, especially “vibe coded” websites. Because ionically, AI-built websites aren’t very AI-crawler-friendly out of the box.

Most AI crawlers don’t run JavaScript, they fetch your raw HTML, which means a vibe coded site is un-crawable.

In fact, Vercel and MERJ tracked more than 500 million GPTBot fetches and found zero evidence of JavaScript execution (Vercel). GPTBot downloads JS files about 11.5% of the time, then never runs them. Same story for Anthropic’s ClaudeBot, PerplexityBot, and Meta’s crawler. Only Google’s Gemini renders JavaScript, and that’s because it borrows Googlebot’s infrastructure.

Okay, so there are some obstacles to overcome when it comes to website GEO, but it’s not all bad news. Let’s get into it.

What is GEO?: How AI website crawlers work

Generative engine optimization (GEO) is about getting your website pages mentioned by AI. So when somebody asks ChatGPT about the best burgers in Houston, Texas, your restaurant is on the list. Answer engine optimization (AEO) is used to describe the same strategy.

Traditional SEO aims for a spot on Google’s page one for the same search term.

Both strategies overlap quite a lot. You need your website to be full of fresh and valuable content. You need other websites to link to you. You also need your brand to be mentioned and discussed across the web (think Reddit, X, Medium, Wikipedia, etc).

Why ChatGPT and Claude are ignoring your website right now

Here’s a quick test to see if AI search engines can crawl your site.

Go give Claude a link to your homepage, product pages, and pricing page. Ask it to tell you what it sees. If it reports back that it’s being blocked in some way, your website could be essentially invisible to AI.

The pages that break this way are often the ones you want quoted most, such as product pages, comparison tables, pricing, documentation, and FAQ sections. These are usually built with frameworks like React, Vue, or Angular using client-side rendering or as single-page apps.

The solution is simple. Use server-side rendering.

When your server sends complete HTML immediately, the crawler can access your headings, text, links, and schema without running any code. In one case study, AI bots made up nearly half of all page requests after pre-rendering was enabled. The difference is clear: your content is either in the first HTML response or it is not. There is no partial credit and no waiting list.

First, the AI reads your page. Then, it decides what to quote. AI engines break your page into sections, turn each section into data, and only use the parts that answer a question. Pages that get cited usually answer one question clearly in each section, with a heading that matches what people are asking.

A 2024 peer-reviewed study on generative engine optimization (Aggarwal et al., presented at KDD) found that adding citations, direct quotes, and statistics to a page increased its visibility in AI answers by up to 40% (GEO study). This makes sense because engines quote pages that already sound like reliable sources.

If you want to rank or appear in AI Overviews, follow the same habits. Put the answer at the top, keep facts as plain HTML text instead of images, and include your JSON-LD structured data in the first response so the crawler can see it. Google’s own advice for AI features supports this approach.

Website GEO checklist: A short checklist to enhance your GEO

You can make real progress on your GEO and AI search engine optimization with a few enhancements you can make today.

1. Crawl access: check this before anything else

• Confirm GPTBot, ClaudeBot, PerplexityBot, and Google-Extended aren't blocked in robots.txt. A lot of sites blocked these by default back in 2023 and 2024 and never revisited it.
• Add an llms.txt file at your root. It's a plain markdown file listing your key pages with short descriptions, basically a curated map instead of forcing AI crawlers to guess from your sitemap. Not every model reads it yet, but it costs nothing and several do respect it.
• Check your server logs for AI crawler hits. If GPTBot or ClaudeBot aren't showing up at all, something upstream is blocking them, usually a CDN rule or Cloudflare's bot fight mode.
• For rendering, open your homepage, right-click and click “view source”. If your website’s HTML is not there, you don't have a content problem, you have a rendering problem, and no amount of better copy fixes it. If you vibe-coded your website, you can easily add Fimo into the mix to make your website crawlable and unlock autonomous SEO and GEO agents to grow your website on autopilot.

2. Structured data

JSON-LD needs to sit in the initial HTML response, not get injected by JavaScript after the page loads. It's the exact same rendering issue, just applied to schema instead of body copy.

• Use Article, FAQPage, HowTo, and Organization schema where they actually fit. Don't bolt FAQPage schema onto content that isn't really Q&A shaped.
• Add author and sameAs properties linking to real profiles. Entity clarity, meaning can the system tell who said this and whether they're a real person, matters more for AI trust signals than it ever did for classic SEO.

3. Content structure

• When writing content, stick to one clear, complete answer per section, with the heading phrased the way someone would actually ask it out loud. "How much does this cost" beats "Pricing Overview", for instance.
• Front-load the answer. AI systems tend to grab the first clear claim under a heading, not the best one three paragraphs down.
• Put the number, date, or named source directly in the sentence making the claim. Not in a linked PDF, not in a chart image. Text that's easy to lift is text that gets lifted.

4. Citations and authority

LLMs want to make sure you’re trustworthy. How do you appear trustworthy? By citing statistics and facts from authoratative websites and institutions. Mention them and link to them, and your reputation will improve.

A few habits worth locking in:

• Keep stats under 12 months old where you can. Stale numbers get skipped in favor of a competitor citing something from last quarter.
• If you've got a named case study, put the company name and the actual metric in the text. Not "a leading brand saw significant improvement." That sentence could belong to anyone.

5. Content freshness and the boring technical wins

• Show a real "last updated" date on pages you're actively maintaining, and then actually update them. A stale date is worse than no date at all.
• Keep sitemap lastmod values accurate, not just bumped on every deploy. Some AI crawlers use this to decide what's worth re-fetching.
• Fix canonical tags, especially around parameter variations or old redirect chains. AI crawlers seem to give up on redirect chains faster than Google does.
• Page speed still matters. A slow first paint means some crawlers time out before they see any content at all.

Most of this is a few hours of cleanup, not a rebuild. The rendering piece is the one that takes actual engineering time, which is exactly why most sites skip it and wonder later why they're invisible everywhere except Google.

GEO Automation: How Fimo handles GEO for you

Fimo is built to solve the problem of GEO.

Every page on a Fimo site is served as clean Markdown that AI crawlers can read without running JavaScript, and it includes JSON-LD structured data in the initial HTML. Server-side rendering is enabled by default, so your page is readable by ChatGPT, Claude, and Perplexity as soon as it goes live. You do not need to add anything later.

But it doesn’t end there. Fimo was built to get your website seen, cited, and visited. It has internal SEO, GEO, and even translation agents that will automatically refresh your content, optimize pages, and ensure LLMs can crawl every sentence.

Try Fimo’s automated GEO agents.

FAQ

Is generative engine optimization the same as SEO?

There is an overlap, but no. SEO gets you a ranked link on Google’s results page. Generative engine optimization gets you a citation inside an AI answer. You can succeed at one and not the other on the same page.

Do AI crawlers read JavaScript?

Mostly no. GPTBot, ClaudeBot, and PerplexityBot fetch raw HTML and do not run JavaScript. Only Google’s Gemini processes JavaScript, since it uses Googlebot’s infrastructure.

What is answer engine optimization (AEO)?

Answer engine optimization means shaping your content to be the direct answer an engine provides, using clear structure, direct language, and quotable facts. It is closely related to GEO.

How do I check if AI can see my page?

Turn off JavaScript in your browser and reload the page. Anything that disappears is what the crawler cannot see. To confirm, check your server logs for AI crawler user agents, since there is no Search Console equivalent yet.

How do I show up in AI Overviews?

Serve your content in plain HTML, answer one question clearly in each section, and keep structured data in the initial response. Google creates AI Overviews from content it can read and trust. That is the short version of how to appear in AI Overviews.

Does LLM SEO need different content than normal SEO?

The writing is similar, but LLM SEO relies more on machine-readable structure. This means using server-rendered text, self-contained sections, and cited facts that an engine can easily use.

Will blocking AI crawlers hurt me?

It depends on which crawlers you block. Training crawlers like GPTBot and ClaudeBot influence long-term model knowledge. Search crawlers like OAI-SearchBot and PerplexityBot decide real-time citations. If you block the search crawlers, you will not appear in live answers.