How to get found (and cited) by ChatGPT: a GEO guide

15 JUL 2026

How to get found (and cited) by ChatGPT: a GEO guide

A site visible on ChatGPT: GEO (Generative Engine Optimization) and an AI answer that cites your page as a source

There’s a question that a year ago was for nerds and today even your cousin asks: “did you ask ChatGPT?” More and more people look for an answer by talking to an AI assistant instead of scrolling ten blue links, and that opens a new game: being first on Google is no longer enough — you have to be cited in the generated answers. Getting found on ChatGPT, Perplexity and the like even has a name: GEO, Generative Engine Optimization. Let’s look at what it is, where it resembles SEO and where it parts ways, and above all what you can concretely do.

Why you want a site visible on ChatGPT.

When an AI model answers a question, it often isn’t making things up: it reads the web and synthesizes, and in the best answers it points to sources with clickable references. Being one of those sources means two things. The first, immediate: someone reads you at the exact moment they’re deciding. The second, subtler: appearing in a ChatGPT or Perplexity answer is a signal of authority that carries trust with it, offline too.

The mistake to avoid is thinking it’s a passing fad or something “for big brands.” The mechanism rewards whoever explains themselves clearly, not whoever has the biggest budget — and that’s exactly the ground on which a well-run small business can beat a bigger, lazier competitor.

How a model ends up citing your page.

The path is simpler than it looks, and understanding it helps you work on it. Your site is read by specialized crawlers — OpenAI’s GPTBot, ClaudeBot, PerplexityBot; that content feeds the model; and when someone asks a relevant question, the model builds the answer and, if your text is clear and trustworthy, cites you as a source.

The GEO flow: from your site to the AI crawlers (GPTBot, ClaudeBot, PerplexityBot), to the model, to the answer with the citation
The path, in four steps: your site (1) is read by AI crawlers (2), feeds the model (3) and, if the content is clear, ends up cited in the answer (4). Every step has a way to go wrong — or to work.

GEO and SEO: what changes (and what doesn’t).

Much of the work is the same as always: clear content, a clean technical structure, speed, structured data. Anyone with solid technical SEO already starts half done. But three things become more important than usual.

aOpen the door to the right crawlers: in robots.txt, don’t block GPTBot, ClaudeBot, PerplexityBot and Google-Extended if you want to appear in generated answers.
bWrite explicit facts: models cite what they understand without ambiguity. “We build e-commerce sites in Milan, delivery in 6 weeks, fixed price” is worth more than “tailored digital solutions for your business.”
cBe readable as text, not as images: a valuable fact inside a JPG poster, to a model, simply doesn’t exist.

Where to start, without getting lost.

GEO isn’t a separate service to buy in a hurry: it’s the same good technical SEO as always, aimed at a new kind of reader. The first concrete step is to check whether your site is already machine-readable — whether it exposes structured data, whether it’s blocking the AI crawlers, whether its sitemap is in order. From there you see at once what’s missing.

And if you want to understand the most talked-about piece of this world — the famous llms.txt file — we’ve covered it separately, no smoke: what it is, how to write it, and how much it really matters.

Sources.

The figures and claims in this article come from here. These are primary sources, not summaries: open them and check for yourself.

Let’s talk about your website.

Free analysis of your current website; a fixed quote within 24 hours of the call.

100% freeNo obligation
Response within 24 hoursDetailed quote
Your data, protectedFull confidentiality

Lascia un commento

Il tuo indirizzo email non sarà pubblicato. I campi obbligatori sono contrassegnati *