Skip to content
Get in touch

GEO and AI search

GEO (Generative Engine Optimization) works on the answers written by ChatGPT, Gemini, Perplexity, Copilot, AI Overviews and AI Mode: there, what counts is whether your site is read, understood and cited. It is for anyone who sees customers asking an AI assistant what they used to search for on Google, and does not know whether their brand appears in the answers. I do not promise citations: I measure what stops AI search engines from citing you and fix it with your team.

What it solves

A search engine or an AI assistant cites content only if it can read it, understand who it is about and attribute it to whoever wrote it. Each of these three steps can break without anyone noticing: the site works, people see it, AI bots do not.

  • They do not read you. GPTBot, ClaudeBot and PerplexityBot download the HTML but do not run JavaScript: if the text only appears in the browser, the page is almost empty for them. Or robots.txt excludes a search crawler, and the site risks appearing less, or no longer appearing, in the answers of that service.
  • They do not understand who you are. The brand has different names across the title tag, structured data and profiles, or speaks from several domains.
  • They do not attribute the content to you. The author, the sources and the profiles linked with sameAs are missing.

How I work

I follow the MRC method and its three pillars: readable, understandable, citable.

I measure what AI bots read in the raw HTML, what robots.txt allows them, which of them actually visit in the server logs, and how the brand presents itself as an entity. Then I look at the answers ChatGPT, Gemini, Perplexity, Copilot, AI Overviews and AI Mode give to the questions of your market: whether they cite you, and whom they cite instead.

I fix things with your team: structured data, rules for bots and entities.

After release I check AI bot visits in the logs, citations in answers, and traffic.

What I deliver

  • The GEO audit: what AI bots read on every template, with the evidence.
  • The rules for AI bots in robots.txt, separated into search, user requests and training. You decide whom to exclude: I tell you what each choice means.
  • The fixes to the brand as an entity: name, structured data, sameAs.
  • The check after release: AI bots in the logs, citations in answers, traffic.

An example

GEO and AI search

Take an online shop with 2,140 product pages. The product description arrives through JavaScript: Googlebot runs it and reads the description, GPTBot and PerplexityBot do not. The server logs show that they visit, but the page they receive has no description.

Then there is robots.txt, and here the choice is yours. GPTBot is OpenAI’s training crawler: excluding it does not remove the shop from the answers. OAI-SearchBot, on the other hand, must stay allowed: it is the crawler that serves ChatGPT’s answers.

The fix is in the template: the description in the HTML the server sends. The check after release has two items: whether OAI-SearchBot reads the product pages, and whether the answers of ChatGPT and Perplexity cite the shop.

Frequently asked questions about GEO and AI search

Does GEO replace SEO?

No. SEO works on the results of Google and Bing, GEO on the answers of AI search engines, and the two overlap to a large extent: a site that bots read and understand, and whose content they can attribute, serves both. That is why I measure them together.

Can you guarantee that ChatGPT will cite my site?

No. Nobody decides what an AI assistant writes. I can remove what stops it from citing you: text that AI bots do not read, search crawlers excluded in robots.txt, a brand that presents itself under different names. Then I check in the logs and in the answers whether anything has changed.

Should I block AI bots?

It depends on which ones. Training crawlers, such as GPTBot or Google-Extended, collect content to train models: excluding them is your choice and does not remove the site from the answers. Search crawlers, such as OAI-SearchBot or PerplexityBot, serve the answers: if you exclude them, the site risks appearing less, or no longer appearing, in those answers. You can see what your robots.txt allows today with AI Robots & llms.txt.

Do I need an llms.txt file?

llms.txt is a useful index for the agents and tools that read it. Google states that its search ignores it; OpenAI, Anthropic and Perplexity do not state that they use it (checked on 21 November 2026). If it is missing, it is not a serious problem: readable HTML and a robots.txt in order come first.

How do you measure whether an AI assistant cites me?

With three measurements. The server logs say which AI bots read the site, and which pages. The assistants’ answers to the questions of your market say whether they cite you, and for what. Traffic says how many people arrive from those answers.

See what AI bots read on your site

MRC Check and Entity Check are free and give you a first measurement. Then get in touch with the result: I’ll reply personally and, if needed, we’ll arrange a call.

Get in touch