· Engineering  · 5 min read

How Will Web Pages Be Found by AI? — llms.txt and JSON-LD Guide

How to make your website discoverable by AI models like ChatGPT, Perplexity, and Gemini. A step-by-step guide to two critical tools of AI discoverability: llms.txt and JSON-LD structured data.

How to make your website discoverable by AI models like ChatGPT, Perplexity, and Gemini. A step-by-step guide to two critical tools of AI discoverability: llms.txt and JSON-LD structured data.

Since 2024, the AI world has radically changed. ChatGPT is no longer just a chatbot — it’s an assistant that browses the web and brings you the latest information. Perplexity, Google Gemini, Claude, and dozens of other AI models are constantly crawling and analyzing websites to answer users’ questions.

But how well do these systems understand your website? When a user asks “daily boat tour in Bodrum,” can ChatGPT recommend your business? Does Perplexity accurately present the pricing information on your site? The answer to these questions lies in the concept of AI discoverability.

What is AI Discoverability?

AI discoverability is the degree to which a website is findable, understandable, and usable by AI systems. While classic SEO (Search Engine Optimization) targets Google, AI discoverability targets LLMs (Large Language Models) like ChatGPT, Perplexity, and Gemini.

The difference is: when Google indexes a page, it scans all HTML, follows links, and determines results with a complex ranking algorithm. LLMs work quite differently:

  • Natural Language Understanding: LLMs “read” your page like a human and grasp its meaning. But they still need structured data.
  • Context Integrity: Information on your site needs to be consistent. If you write “10:00-18:00” on one page and “09:00-19:00” on another, the AI gets confused.
  • Summarization: LLMs extract information from your site and present it to the user as a summary. llms.txt lets you control this summary.

llms.txt — Give LLMs a Guide to Your Site

llms.txt is a text file added to the root directory of your website. Just as robots.txt guides search engines, llms.txt provides AI models with a summary of your site’s content.

What Does It Look Like?

Here’s a real llms.txt example:

# Simge Yachting | Bodrum & Yalıkavak Luxury Yacht Charter

> Simge Yachting — Luxury gulet charter in Bodrum, Yalıkavak since 1990.

## Company
- **Name:** Simge Yachting
- **Location:** Yalıkavak, Bodrum, Muğla, Türkiye
- **Phone:** +90 532 681 74 63
- **Email:** simgeyachting2@gmail.com
- **Languages:** Turkish, English, Italian

## Services
- Daily Boat Tours
- Blue Cruise (weekly)
- Dinner Cruises
- Greek Islands Tours

## Key Pages
- [Home](https://simgeyachting.com/)
- [Our Boat](https://simgeyachting.com/teknemiz)
- [Daily Boat Tours](https://simgeyachting.com/gunluk-tekne-turlari)
- [Contact](https://simgeyachting.com/iletisim)

Why Is It Important?

  • ChatGPT’s web browsing feature automatically scans llms.txt
  • Provides LLMs with the official summary of your site — instead of their own extracted summaries
  • Lists all your important pages in a single file, giving AI a roadmap
  • Takes less than 5 minutes to set up, almost zero maintenance

How to Create It?

  1. Create an llms.txt file in your website’s root directory (public/ folder)
  2. Start with business name, short description, and contact information
  3. List all your services
  4. List all important pages with links
  5. Deploy and check at yoursite.com/llms.txt

JSON-LD Structured Data — Give Context to AI

If llms.txt is your site’s table of contents, JSON-LD structured data is your site’s detailed identity card. It’s the most modern way to add structured data to web pages using Schema.org standards.

Most Critical JSON-LD Schemas

1. LocalBusiness

Contains your business name, address, phone, working hours, rating, and social media links. ChatGPT uses this data when asked about a business.

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Simge Yachting",
  "telephone": "+90******",
  "address": {
    "@type": "PostalAddress",
    "addressLocality": "Yalıkavak",
    "addressRegion": "Bodrum",
    "addressCountry": "TR"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.7",
    "reviewCount": "40",
    "bestRating": "5"
  },
  "priceRange": "€€€€"
}

2. FAQPage

Presents frequently asked questions in structured format. Perplexity and ChatGPT use FAQ schema to answer users directly.

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "How long does the daily tour last?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Our tours last 8 hours from 10:00 to 18:00."
      }
    }
  ]
}

3. BreadcrumbList

Shows the page’s position within the site. Helps AI understand the page’s context.

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://simgeyachting.com" },
    { "@type": "ListItem", "position": 2, "name": "Daily Tours", "item": "https://simgeyachting.com/gunluk-tekne-turlari" }
  ]
}

What You Can Do Today — AI Discoverability in 10 Steps

  1. Create an llms.txt file in your website’s root directory (public/)
  2. Add business name, description, and contact info to llms.txt
  3. List all important pages in llms.txt
  4. Add LocalBusiness JSON-LD schema to the homepage
  5. Add FAQPage schema to service pages
  6. Add BreadcrumbList schema to interior pages
  7. Make sure you have robots.txt and sitemap.xml
  8. Verify your site is registered in Google Search Console
  9. Test that your pages don’t return 404 errors
  10. Keep your llms.txt up to date — update it as you add new pages

Conclusion

AI-powered search is growing rapidly. ChatGPT has 400 million weekly active users, and a large portion of them now ask AI assistants instead of Google. Perplexity, Google Gemini, Claude, and many more… You need to be prepared to be visible on these platforms.

Fortunately, the preparation is very simple: one llms.txt file and a solid JSON-LD structure. You can make your site AI-friendly in under 30 minutes total.

What We Did at Simge Yachting?

As a real case study, we implemented the following on Simge Yachting:

  • ✅ Created llms.txt file — listed all services, pages, and contact info
  • ✅ Added LocalBusiness JSON-LD — address, phone, working hours, 4.7 rating and 40 reviews
  • FAQPage schemas — structured FAQs on all service pages
  • BreadcrumbList — added navigation schema on interior pages
  • Multi-language support — Turkish, English, and Italian content

Check your website too. Go to yoursite.com/llms.txt and test if the file exists. Search for JSON-LD structure in your page source. Don’t be late to be visible in the AI era!

Back to Blog

Related Posts

View All Posts »
WhatsApp ile yazin