Search is changing under your feet. A growing share of your would-be visitors now ask ChatGPT, Perplexity, Google AI Overviews, or Microsoft Copilot a question and read the answer without ever clicking a blue link. When those tools answer, they cite a handful of sources, and being one of them is the new front page. This Ultra Web Hosting guide covers what people are calling GEO or answer-engine optimization: how AI systems find and use your content, and the concrete steps to make your site the one they quote.
- What Is Actually Changing
- How AI Systems Find and Use Your Content
- Allow the Retrieval Bots in robots.txt
- Write Content AI Can Extract
- Add Structured Data With Schema Markup
- The llms.txt Convention
- E-E-A-T Signals AI Weighs
- Technical Basics That Still Matter
- Measuring Whether It Works
- The Honest Bottom Line
Good SEO, Open Doors, Clear Structure
There is no secret trick to getting cited by AI. It is mostly the same fundamentals that win at classic search, plus two things specific to answer engines: letting the retrieval bots in, and structuring your pages so a machine can lift a clean answer straight out of them.
- Allow the AI retrieval bots you want to be cited by in robots.txt
- Answer questions directly, near the top, in plain language
- Add Article, FAQPage, and HowTo schema so machines parse you
- Keep the site fast, indexed, and technically clean
You Cannot Be Cited If You Block the Bots
The number one reason a site never shows up in AI answers is that it quietly blocks the very crawlers that feed those answers. If your robots.txt disallows GPTBot, ClaudeBot, or PerplexityBot, you have opted out of being quoted. This guide is the opposite of our Allowing or Denying AI Bot Browsing of your Website article. That one is for owners who want to keep AI out. If your goal is visibility, you want these bots in.
- Training crawlers and live-retrieval bots are different things
- Live-retrieval bots are the ones that fetch you at answer time
- Blocking them removes you from the answer entirely
01. What Is Actually Changing
For twenty years the game was to rank on page one of Google so someone would click through to your site. That still matters, but a large and rising share of searches now end without a click. Someone asks ChatGPT how to fix a slow WordPress site, Perplexity to compare two products, or Google AI Overviews to summarize a topic, and they read the synthesized answer in place. Copilot does the same inside Windows and Edge.
Those answers are not pulled from thin air. Each system draws on a mix of what it learned during training and, increasingly, live web results it fetches at the moment you ask. When it fetches live results, it names its sources. Being one of the cited sources is the new equivalent of ranking on page one: it puts your name, your expertise, and a link in front of the person asking, even when they never run a traditional search.
You will see this called GEO (generative engine optimization), AEO (answer engine optimization), or LLM SEO. The labels are new and still shaking out. The work underneath is not exotic: clear writing, open crawler access, and machine-readable structure. Do not let anyone convince you it requires a secret playbook.
02. How AI Systems Find and Use Your Content
There are two very different ways your content reaches an AI answer, and confusing them leads to bad decisions.
Training Crawlers
Bots like GPTBot and ClaudeBot collect pages to help train future models. Anything they gather may inform the model months later, in a general way, with no live link back to you.
- Feeds the model's baseline knowledge
- No direct citation at answer time
- Effect is diffuse and delayed
Live-Retrieval Bots
When a user asks something current, the assistant fetches pages in real time, reads them, and cites them in the reply. This is where a citation with your link actually appears.
- Fetches your page at the moment of the question
- Names and links you in the answer
- This is the traffic you can win today
The practical takeaway: the live-retrieval bots are the ones worth courting, because they produce visible citations now. To be eligible, they have to be able to reach your pages, which is entirely a robots.txt question.
03. Allow the Retrieval Bots in robots.txt
Your robots.txt file, sitting at the root of your domain, tells crawlers what they may fetch. If you want to be cited, you must not block the AI user agents. Many site owners added blanket AI blocks in the last couple of years without realizing it also opts them out of AI answers. Check yours.
To explicitly welcome the major AI crawlers, an allow-everything block looks like this:
User-agent: GPTBot
Allow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: ChatGPT-User
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Google-Extended
Allow: /
Sitemap: https://www.yourdomain.com/sitemap.xmlIf you instead want to keep AI tools out of your content, do not follow this section. Read our Allowing or Denying AI Bot Browsing of your Website guide, which walks through the Disallow rules for each bot. The two guides are two sides of one switch. Pick the side that matches your goal, and do not accidentally run both. For how robots.txt interacts with your traffic reports, see article 85 on robots.txt in web stats.
robots.txt is honored on trust, not enforced. Reputable crawlers from OpenAI, Anthropic, Google, and Perplexity respect it. A missing or default file is usually fine on its own, but if you ever added AI Disallow lines, those are what shut you out. Edit the file in cPanel File Manager at the root of your public_html and remove the blocking lines.
04. Write Content AI Can Extract
An answer engine is looking for a clean, quotable passage it can lift with confidence. The pages that get cited tend to share the same shape, and it is the same shape that helps human readers too.
- Lead with the answer. Put a direct, one or two sentence answer right under the heading, before the background. If someone asks "what port does SMTP use," the first line should say it, then explain.
- Use question-shaped headings. Headings that mirror how people actually ask ("How do I speed up WordPress?") match the query and give the machine a clear boundary for the answer that follows.
- Be specific and factual. Concrete numbers, versions, dates, and named steps get quoted. Vague marketing copy does not. State facts plainly.
- Define your terms. Spell out acronyms and jargon on first use. Machines and newcomers both benefit, and a clear definition is exactly the kind of passage that gets pulled into an answer.
- Keep passages self-contained. Write so a single paragraph makes sense on its own, without the three paragraphs before it. That is how it will be quoted.
A short list of real questions and direct answers at the bottom of a page is some of the most citable content you can write. It maps cleanly to how people query assistants, and it pairs perfectly with the FAQPage schema in the next section.
05. Add Structured Data With Schema Markup
Schema markup is a block of machine-readable data, usually written as JSON-LD in the page head, that tells crawlers exactly what your content is: an article, a set of FAQs, a step-by-step how-to, an organization. It removes guesswork. When a machine knows a block is the answer to a specific question, it can quote it with far more confidence.
The types worth knowing are Article, FAQPage, HowTo, and Organization. Here is a compact FAQPage example you can adapt:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "Do I need to allow AI bots to be cited?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes. If robots.txt blocks the retrieval bots, your pages cannot be fetched or cited in AI answers."
}
}]
}
</script>And a minimal Organization block that establishes who is behind the site:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Your Company",
"url": "https://www.yourdomain.com/",
"logo": "https://www.yourdomain.com/logo.png"
}
</script>You do not have to hand-write JSON-LD. On WordPress, both Yoast SEO and Rank Math output Article, FAQPage, HowTo, and Organization schema automatically from fields you fill in. If you have not set one up yet, see our Setting Up Yoast or Rank Math for WordPress SEO guide, then use its schema and FAQ blocks. Article 276 covers keeping that WordPress install fast.
06. The llms.txt Convention
A newer idea borrowed from the robots.txt and sitemap.xml pattern is llms.txt: a plain text (Markdown) file at the root of your domain that points AI tools at your most important, cleanest content. The goal is to give an assistant a curated map of your site instead of making it crawl everything and guess.
A sample llms.txt looks like this:
# Your Company
> One-line description of what your site is and who it serves.
## Docs
- [Getting Started](https://www.yourdomain.com/docs/start): first steps
- [Pricing](https://www.yourdomain.com/pricing): plans and costs
## Guides
- [How to Speed Up WordPress](https://www.yourdomain.com/wp-speed): performance tuning
## Optional
- [Company Background](https://www.yourdomain.com/about): history and teamTo create it in cPanel:
- Open cPanel > File Manager and go to
public_html(the root of your site). - Click + File, name it
llms.txt, and click Create New File. - Select the file, click Edit, paste your content in the Markdown format above, and save.
- Confirm it loads by visiting
https://www.yourdomain.com/llms.txtin a browser.
llms.txt is a proposed convention, not a ratified standard, and as of mid 2026 the major AI vendors have not committed to honoring it. It costs you ten minutes and cannot hurt, so it is reasonable to add. Just do not expect it to move the needle on its own, and do not let anyone sell it to you as a guaranteed ranking factor. The real work is in the sections above and below.
07. E-E-A-T Signals AI Weighs
AI systems, like Google before them, try to gauge whether a source is trustworthy before leaning on it. The shorthand is E-E-A-T: experience, expertise, authoritativeness, and trust. You cannot fake these into existence, but you can make the signals you already have legible.
- Real author bylines. Put a named author on articles, with a short bio establishing why they know the subject. Anonymous content is easier to discount.
- Clear about and contact pages. A findable about page, a real contact method, and a physical or business address tell a machine there is an accountable entity behind the site.
- Citations and sources. Link out to the primary sources you rely on. Content that shows its work reads as more reliable.
- Being referenced elsewhere. When other reputable sites, directories, and profiles mention or link you, assistants pick up on that corroboration. This is the slow, earned part, and there is no shortcut.
08. Technical Basics That Still Matter
None of the above helps if a bot cannot cleanly fetch and parse your pages. The old technical fundamentals carry straight over to the AI era.
- Fast load and clean HTML. A retrieval bot has a limited budget of patience. Slow, bloated pages get skipped, and messy markup makes extraction harder. See article 276 on optimizing WordPress performance and article 466 on speeding up your website.
- A valid sitemap and real indexing. If a page is not discoverable and indexed, it will not be retrieved. Submit your sitemap and confirm coverage per article 178 on submitting your site to search engines.
- Do not accidentally block yourself. A stray
Disallowin robots.txt or anoindextag can silently remove pages from everything, AI included. Article 85 explains how robots.txt shows up in your web stats so you can spot what is being blocked.
Notice that speed, a clean sitemap, and correct crawler rules are exactly what you would do for classic SEO. That is the point. The traditional groundwork is most of what earns AI citations too, which means the effort is not wasted whichever way search evolves.
09. Measuring Whether It Works
Analytics for AI citations are immature, so measurement here is more manual than you are used to. A few methods that work today:
- Ask the tools directly. Search your brand name and your key topics in ChatGPT, Perplexity, Google AI Overviews, and Copilot. See whether you appear, and note who gets cited instead when you do not.
- Watch your referral traffic. In your analytics, look for referrers such as
chat.openai.comandperplexity.ai. Rising visits from those sources mean your citations are getting clicked. - Track over time, not day to day. Answers vary run to run, so a single check tells you little. Repeat the same queries monthly and look for a trend.
There is no AI equivalent of a rank-tracking dashboard that gives you a clean position number yet. Anyone who claims a precise AI ranking score is selling you certainty that does not exist. Manual checks plus referral traffic are the honest state of the art for now.
10. The Honest Bottom Line
Getting cited by AI is far less mysterious than the marketing around it suggests. Solid traditional SEO, open access for the retrieval bots, content structured as clear questions and answers, and a few schema blocks cover most of the game. Do those and you are eligible for the citations that matter.
What does not work is paying for guaranteed AI ranking. No one controls what ChatGPT or Perplexity cites, the models change constantly, and there is no placement to buy. If a vendor promises a spot in AI answers, keep your money. Spend the effort on the fundamentals instead, because they pay off in classic search regardless of where AI goes next.
Want a Second Set of Eyes on Your Setup?
If you would like us to check your robots.txt for accidental AI blocks, confirm your sitemap is submitted and indexed, or look at your site speed, open a ticket and we will take a look. We support these checks on shared, VPS, and dedicated plans.
Open a Support TicketQuick Recap: Getting Cited in Six Moves
If you only do six things from this guide, do these:
- Check robots.txt and remove any lines that block
GPTBot,ClaudeBot,PerplexityBot, or the other retrieval bots you want to be cited by. - Lead every page with a direct answer under a question-shaped heading, written to be quoted on its own.
- Add schema markup (Article, FAQPage, HowTo, Organization), letting Yoast or Rank Math generate it on WordPress.
- Consider an llms.txt file at your domain root, knowing it is a proposed standard and not yet widely honored.
- Keep the site fast, indexed, and cleanly built so bots can actually fetch and parse you.
- Measure by asking the tools and watching referral traffic, and ignore anyone selling guaranteed AI ranking.
Last updated July 2026 · Browse all Website Tools and SEO articles
