Ranking in AI
Schema.org, JSON-LD, llms.txt and clean semantics: the foundation without which AI cannot read your business correctly.
ChatGPT, Gemini and Perplexity read a site differently from Google: they rely on structured data, semantic hierarchy and the llms.txt file. Without the right technical foundation, AI systems ignore or distort your company information in their answers. We set up the markup turnkey (from FAQPage to LocalBusiness) and make your site readable for all modern AI search engines.
Technical AI search setup means implementing Schema.org (JSON-LD), structured data, an llms.txt file and clean HTML semantics so that ChatGPT, Gemini and Perplexity understand and cite your site correctly. Without this foundation, AI systems ignore or distort your business information. Request an audit and we reply within 24 hours.
Here we cover technical AI search setup for businesses across Spain. We explain what Schema.org, JSON-LD and structured data are, why you need an llms.txt file and how clean semantics affect the way ChatGPT, Gemini and Perplexity read and cite your site. By technical setup we mean implementing Schema.org markup of the right types (FAQPage, LocalBusiness, Article, BreadcrumbList, Product, Service), writing valid JSON-LD, creating and optimising llms.txt, cleaning up HTML structure and heading hierarchy, and configuring sitemap and robots.txt for LLM crawlers. Below you will find a step-by-step audit and implementation plan, a table of key setup elements, Google Search Central guidance on structured data and answers to common questions. At the end, our "Ranking in AI" packages with real prices and related services.
Most sites either have no Schema.org at all, use the outdated Microdata format instead of JSON-LD, or contain validation errors. Before adding anything, you need to understand what already exists and where the critical gaps are. The right sequence is: audit → prioritise markup types → write JSON-LD → validate → llms.txt → check semantics. After implementation, AI systems start correctly interpreting pages on the very next crawl.
Not all technical elements affect AI equally. Schema.org in JSON-LD format and the llms.txt file are direct signals to language models. Clean HTML semantics and heading hierarchy help LLM parsers split the page into meaningful blocks correctly. Sitemap and robots.txt determine what actually enters AI crawler indexes. The table below shows each element with its role and the expected effect on AI search.
| Element | Why it matters | Effect on AI |
|---|---|---|
| Schema.org JSON-LD | Passes structured data about the business, services and FAQ directly to the parser | AI knows the exact type of object, its attributes and relationships |
| FAQPage schema | Marks questions and answers as separate semantic blocks | Direct answer output in ChatGPT, Perplexity and voice assistant chats |
| llms.txt | Controls LLM crawler access to content and sets indexing priorities | AI systems read the right sections and do not index service content |
| HTML semantics and H1-H6 | Forms the logical page structure for parsers | LLM correctly chunks content and cites the right block |
| Sitemap.xml + robots.txt | Tells crawlers which pages exist and are accessible | AI crawlers traverse the full site and do not waste budget on blocked URLs |
AI systems, including ChatGPT, Gemini and Perplexity, use structured data in two scenarios: during indexing, to classify the page precisely, and during answer generation, to extract a specific attribute (address, price, FAQ answer) without the risk of hallucination. Without Schema.org, the model is forced to "guess" from the text, which sharply reduces citation accuracy. Google Search Central confirms that structured data is a direct signal for search and AI systems, enabling content to be displayed in rich result formats.
«Structured data is code that you add to your website pages to describe your content so that search engines can better understand it. Search engines use that knowledge to display your content in useful (and visually engaging) formats in search results.»
Technical AI search setup means Schema.org (JSON-LD), an llms.txt file, clean HTML semantics and a correct sitemap. Without this foundation, ChatGPT, Gemini and Perplexity do not read your site correctly: they ignore services, mix up the address and do not cite your FAQ. With valid markup, AI systems receive precise data about your business and use it in their answers, which is the direct path to being cited in AI search. Start with an audit of the current markup, then implement in priority order: LocalBusiness → FAQPage → Article → llms.txt. The work takes 1-2 weeks and delivers a long-term effect without ongoing investment. Request a quote and we will prepare your audit and "Ranking in AI" plan within 24 hours.
AI Promotion
So ChatGPT, Gemini, Perplexity and Google AI Overview recommend your company when a customer asks “who’s the best in Valencia”. From €490 a month.
Ranking inside AI models works differently from classic SEO: what matters is how facts about your company are structured, how quotable they are, and whether you show up in the sources the models read. Every month we check what ChatGPT, Gemini and Perplexity say about you and fix whatever keeps you out of the answer. The field is young, so we work on subscription: models keep changing and a set-it-once setup doesn’t hold here. In Spain these services run €1,500-5,000 a month; we come in three times cheaper.
A short brief, and we’ll come back with a plan and quote within 24-48 h. No pressure.
Schema.org is the structured data standard that describes objects on a page (business, services, FAQ, articles) in machine-readable JSON-LD format. It is a direct signal to ChatGPT, Gemini and Perplexity: instead of "guessing" the meaning of text, the AI receives precise facts and attributes.
llms.txt is a file in the site root that controls LLM crawler access (GPTBot, Google-Extended, ClaudeBot and others) to content. It specifies which sections may be indexed and which may not, and contains a brief site description for the AI. Without it, crawlers may read service pages and form incorrect context.
In priority order: LocalBusiness (or specialised type) for core company data; FAQPage for getting Q&A into AI answers; Article for content pages; BreadcrumbList for site structure; Service for service pages. Every type must be valid JSON-LD with no errors. We start with the type that covers the core company data and add the rest page by page.
Yes, positively. Valid structured data produces rich snippets in Google, raises CTR and helps with ranking. In other words, the same work improves both AI visibility and classic SEO at the same time. So the work is budgeted once, not twice.
Auditing the current state takes 24 hours. Full implementation of Schema.org, llms.txt and semantics cleanup takes 1-2 weeks depending on site size and the number of page types. The result is verified in Rich Results Test immediately after publishing. What takes longest is agreeing the company data that goes into the markup.
AI systems will read your site "as best they can": ignore services, mix up contact details and not use your FAQ in answers. Competitors with correct markup will be cited instead of you in ChatGPT and Perplexity. You can check this yourself: ask the assistant about your services and compare with the site.
Yes. Structured data must match the current content of the page. If prices, services, opening hours or FAQ change, the markup is updated in sync. A mismatch between JSON-LD and page text reduces trust from search engines and AI systems.
Use Google Rich Results Test (search.google.com/test/rich-results) and Schema Markup Validator (validator.schema.org). Both tools show errors and warnings for each markup type. The goal is zero errors and minimum warnings. The check is repeated after every major site update. The validation results are attached to the report.
Yes. We work with WordPress, Webflow, Tilda, Next.js, React and custom CMS platforms. JSON-LD can be implemented on any platform: in the template, via a plugin or through a component. We choose the optimal method for each project. The site access we need is minimal: only the templates or the markup settings.
The price includes: an audit of the current markup and semantics, writing JSON-LD for all page types, creating llms.txt, checking HTML hierarchy and alt texts, updating sitemap.xml and robots.txt, validation in Rich Results Test and a maintenance guide. Exact packages and prices are in the "Ranking in AI" section on the Prices page.
