How to get cited by ChatGPT, Claude and Perplexity in 2026: the complete method
Generative engines select sources; they do not rank pages. Four conditions for entering their answers, and how to measure your citation share.
Read the article →
GEO for real estate: being recommended by AI
Buyers ask AI where to buy, at what price and with which agency. Real prompts, citable local pages, entity and reviews: how to get into their answers.
Building a representative prompt panel
A prompt panel is a sample of the questions your buyers really ask: sources, split by intent and language, minimum size, and maintenance without breaking the series.
Comparison sites and sector directories: the pages that decide
For “best X” questions, engines cite comparison sites and directories before brands. How to identify them, get listed, and correct what they say about you.
Press relations and AI citation: from coverage to source
A press article can become a recurring source for generative engines. Angles that get cited, wording journalists keep, consistency with your site.
Customer reviews and ratings in AI recommendations
To recommend, a generative engine reads customer reviews: platforms cited, volume, recency, content, business replies and permitted markup. Method and mistakes.
Product structured data and e-commerce
Product, Offer, reviews, availability, delivery: what shopping assistants read in a product page, through which channels, and the full JSON-LD of a readable product page.
Schema.org for generative search
Four schema.org types are enough for a generative engine to identify who publishes, who writes, what and when. Useful properties, complete JSON-LD and validation.
Multilingual content and generative engines
A generative engine answers in the prompt's language but picks sources from several. Adaptation by market, hreflang, one entity and per-language measurement.
Figures, dates and sources: verifiable content for an AI
A model does not verify a claim, it cross-checks it. Attribute, date, describe the method: the markers that make a figure stick, and those that get it discarded.
Writing citable passages
A citable passage answers a question on its own, in forty to eighty words, with evidence and a caveat. Method, review checklist and rewrites by page type.
Sitemaps, dates and freshness for generative indexes
An honest lastmod, a visible and consistent date, an update feed: the freshness signals that indexes read, and which content to refresh first.
JavaScript rendering and generative engines
ChatGPT, Claude and Perplexity crawlers read the server's HTML, not the rendered page. A ten-minute test, the patterns that empty a page, SSR, pre-rendering and islands.
Claude and web search: source selection and citation
Claude-SearchBot, Claude-User, accessibility requirements, citations anchored to an exact passage, differences from ChatGPT: what Anthropic documents and what we measure.
Perplexity: anatomy of an answer and the levers for appearing in it
Perplexity searches on every question and numbers its sources. Zones of an answer, PerplexityBot and Perplexity-User, weight of third parties, winning formats and B2B cases.
How generative search works
Prompt decomposition, passage retrieval, reranking, cited synthesis: the four steps of a generative engine, and what each one concretely demands of a page.