llms.txt tells AI crawlers what your site does in plain text
Schema.org markup gets your products cited in AI-generated answers
Content structure matters more than keywords for LLM discovery
Most sites are invisible to AI search engines right now, fix takes under a day
Most Websites Are Invisible to AI Search Engines
Here is a number that should concern every business owner: over 80% of websites have zero optimization for AI-powered search. Google still matters, sure. But when someone asks ChatGPT, Perplexity, or Claude for a product recommendation, your site either shows up or it does not. There is no page two. There is no "close enough." The AI either cites you or it cites your competitor.
I run a one-person creative studio. I sell digital products, merch, and tools through my Shopify store. Six months ago, none of my products appeared in any AI search result. Today, they show up consistently across Perplexity, ChatGPT search, and Google AI Overviews. The difference was not a massive SEO overhaul. It was a few specific structural changes that took less than a day.
This is not about replacing traditional SEO. It is about adding a layer that makes your content readable by a completely different kind of search engine.
What AI Search Engines Actually Read
Traditional search engines crawl HTML, follow links, and index keywords. AI search engines do something fundamentally different. They try to understand what your site is about, what it offers, and whether that information is trustworthy enough to cite.
AI models pull from several sources when generating answers:
Training data: the massive corpus the model was trained on (you cannot control this retroactively)
Real-time retrieval: tools like Perplexity and ChatGPT with browsing actively fetch and read your pages
Structured data: Schema.org markup that gives machines explicit context about your content
llms.txt: a new standard specifically designed for AI model consumption
The retrieval layer is where you have the most control right now. When Perplexity visits your site to answer a query, it reads your HTML the same way a browser does. But it processes that content through a language model, which means clarity and structure matter far more than keyword density.
If your page is cluttered with popups, cookie banners covering content, and JavaScript-rendered text that loads three seconds after the initial request, the AI crawler gets garbage. Clean, fast, well-structured HTML wins.
llms.txt: The robots.txt for AI
The llms.txt standard is simple. You create a plain text file at yoursite.com/llms.txt that describes your site in natural language. Think of it as robots.txt, but instead of telling crawlers what not to index, you tell AI models what your site actually does.
Here is a stripped-down example:
# RAXXO Studios
> AI-powered creative studio selling digital tools, merch, and design resources.
## Products
- Claude Blueprint: Pre-configured Claude Code setup for developers (33 EUR)
- FULLMOON: Automated codebase audit tool (49 EUR)
- Statusline Builder: Free terminal status line generator
## Blog
- Lab blog covering AI tools, web development, and creative workflows
- 100+ articles on practical AI implementation
## Links
- Store: https://raxxo.shop
- Blog: https://raxxo.shop/blogs/lab
- Studio: https://raxxo.studio
That is it. No special syntax. No complex configuration. Just a clear, honest description of what you offer. AI models parse this instantly and use it as context when deciding whether to cite your site.
The key rules: be specific, include pricing where relevant, and update it when your product lineup changes. A stale llms.txt is worse than none at all because it trains AI models on wrong information about your business.
Schema.org Markup That Actually Gets Cited
Schema.org structured data has been around for years. Most SEO guides tell you to add it for rich snippets in Google. But the real value in 2026 is that AI search engines use Schema.org as a primary signal for understanding your content.
The types that matter most for AI discovery:
Product schema gives AI models everything they need to recommend your products. Price, availability, description, ratings. When someone asks Perplexity "best Claude Code setup tools," having proper Product schema on your page means the AI can confidently cite your product with accurate pricing.
Article schema on blog posts helps AI models understand authorship, publication date, and topic. This is critical for freshness signals. An AI model deciding between two sources will prefer the one with clear Article schema showing a recent publication date.
FAQ schema is arguably the most powerful for AI search. When you mark up questions and answers on your page, AI models can directly pull those answers into generated responses. I added FAQ schema to five product pages and saw Perplexity citations within two weeks.
Organization schema establishes your brand entity. Name, logo, social profiles, founding date. This helps AI models connect your various online presences into a single entity they can reference confidently.
The implementation is straightforward. Add JSON-LD script blocks to your page headers. If you are on Shopify, most themes include basic Product schema, but you will want to extend it with FAQ and Organization types manually.
Content Architecture for LLM Readability
Keywords still matter for traditional SEO. For AI search, the architecture of your content matters more. AI models process your page top to bottom and build understanding incrementally. If your most important information is buried below three promotional banners and a newsletter signup, the AI might never reach it.
Rules I follow for every page:
Lead with the answer. If your page is about "best AI coding tools," the first paragraph should contain your actual recommendations. AI models weight early content more heavily when extracting answers.
Use clear heading hierarchy. H1 for the page topic, H2 for major sections, H3 for subsections. Never skip levels. AI models use heading structure to understand content relationships.
Write in complete, factual sentences. AI models struggle to cite content that relies on context from surrounding paragraphs. Each paragraph should be independently meaningful. "This tool costs 33 EUR and includes 6 pre-built skills" is better than "It is affordable and comes with several features."
Include specific numbers. AI models prefer citing sources with concrete data. "Reduced build time by 47%" gets cited. "Significantly improved performance" does not.
Avoid walls of text. Short paragraphs, bullet lists, and clear formatting help AI models extract discrete facts. A 500-word paragraph about your product features is harder to cite than a bulleted list of those same features.
Technical Optimization Checklist
Beyond content, several technical factors determine whether AI search engines can even access your content:
Server-side rendering. If your site relies on client-side JavaScript to render content, many AI crawlers will see an empty page. Next.js, Nuxt, and Astro all support server-side rendering out of the box. If you are using a SPA framework, add SSR or pre-rendering.
Fast response times. AI crawlers are impatient. If your page takes more than three seconds to return content, some crawlers will skip it entirely. I keep all my pages under 1.5 seconds using Vercel's edge network and aggressive caching.
Clean URLs. /products/claude-blueprint is better than /products?id=47&variant=default&ref=home. AI models include URLs in citations, and clean URLs build trust.
No content gating. If your content is behind a login wall, email capture, or JavaScript modal, AI crawlers cannot access it. The content you want cited must be freely accessible.
XML sitemap. Make sure your sitemap is current and submitted to Google Search Console. Perplexity and other AI search tools use Google's index as a starting point for their own crawling.
Meta descriptions that read like answers. Write meta descriptions as if they are the answer to a question. "Claude Blueprint is a 33 EUR pre-configured Claude Code setup with 6 skills, 6 commands, and 3 hooks" beats "Learn more about our amazing product."
Monitoring AI Search Presence
You cannot optimize what you do not measure. Here is how I track AI search visibility:
Manual testing. Once a week, I search for my products and topics on Perplexity, ChatGPT, and Google AI Overviews. I note which queries return my content and which do not. Low-tech, but effective.
Referral traffic. Check your analytics for traffic from perplexity.ai, chatgpt.com, and similar AI search sources. This traffic is growing fast. I went from zero AI referrals in January to consistent daily visits by March.
Citation tracking. When an AI cites your site, it usually includes a link. Track these inbound links separately from traditional backlinks. They indicate which pages AI models consider authoritative.
Competitor monitoring. Search for your core topics on AI search engines. If a competitor shows up and you do not, study their page structure. What structured data do they have? Is their content more specific? Do they have an llms.txt file?
The tools I use for scheduling content around these findings: Buffer for social distribution that drives the initial traffic signals AI search engines look for, and Shopify analytics for tracking which product pages convert from AI referral traffic.
The Compounding Effect
AI search optimization compounds faster than traditional SEO. Once an AI model cites your content in a response, that response gets seen by thousands of users. Some of those users visit your site, increasing your traffic signals. The next time the AI crawls your site, it sees more engagement data confirming your content is valuable. More citations follow.
I added llms.txt, expanded my Schema.org markup, and restructured five key pages in a single weekend. Within three weeks, my products started appearing in Perplexity results for relevant queries. The traffic from AI search is still a fraction of organic Google traffic, but it converts at nearly double the rate because AI search users arrive with high intent.
The window to get ahead of this is closing. Right now, most sites have not optimized for AI search at all. That means even basic optimization gives you a significant advantage. In a year, every SEO guide will include this. The early movers will already have established authority.
Start with llms.txt. Add Product and FAQ schema to your top pages. Restructure your content to lead with answers. It is less work than a typical SEO sprint, and the results show up faster than you expect.
Top comments (0)