What is GEO, and how do you get named by ChatGPT?

How do these tools pick their sources?

They draw on what is published on the web, and they favour content that is quotable, clear and organised around questions. Google is explicit about its own AI features. There is no separate trick. The site has to be indexed and the content has to be allowed to appear. GEO is built on top of solid SEO, it does not replace it.

What kind of writing gets used in an AI answer?

Three qualities stand out. Definition first. The page defines the subject in its opening sentence, which is where the machine takes its answer from. Question-shaped headings. Sections written the way people ask ("what decides the price?") match the question itself. Self-contained passages. Each section has to make sense without the rest of the page, because these tools quote passages rather than whole pages.

On the technical side the content has to exist in the raw HTML. Text that only appears after JavaScript runs in a browser is invisible to many of these crawlers.

What should you check on your own site?

Look at whether your robots file is blocking AI crawlers. A surprising number of sites block them without knowing. Make sure your structured data says exactly what the visible text says. And check that the basics about your business (what you do, where you serve, how to reach you) are written out in plain sentences. These tools repeat what is written rather than inferring it.

Should you allow AI crawlers on your site?

It is a choice, and it should be a deliberate one. The robots file in your site root tells crawlers what they may access, and the crawlers behind AI tools read it. Block them and your content is not available to be cited in their answers.

If visibility is the goal, allowing access is the starting point. Some publishers block deliberately and have good reasons, usually because original research is their product. For a local service business the calculation runs the other way, since there is little to protect and something to gain.

Here is what we do on our own site. Our robots.txt allows 15 AI crawlers by name, among them GPTBot, ClaudeBot, PerplexityBot and Google-Extended. No page is blocked. The page a bot sees and the page a visitor sees are identical. Serving a separate version to bots is a policy violation, and it is needless work.

What does schema markup do for AI visibility?

Schema restates what is on the page in a form a machine reads without guessing. Business name, address, opening hours, the list of services and the questions you answer can all be marked up this way.

One rule governs it. The markup has to match what a visitor can actually see on the page. Marking up information that is not there, or giving a different value, breaks search engine policy and costs trust. Schema describes the page, it does not decorate it.

Does being mentioned in an AI answer bring traffic?

The honest answer is that a mention brings less direct traffic than a click from a search result, because the user often stops at the answer. The value shows up elsewhere. When your name appears in the answer someone gets while researching a service, recognition is being built, and that person searches for you directly later.

So it is better understood as an extension of search visibility than as a separate channel. The prerequisite is the same as well. A page that has not been indexed by a search engine does not get cited in an AI answer either.