Book a call

Daniel & Friends / Blog / GEO & AI

How to get into ChatGPT's answers: a practical GEO guide

More and more purchases start with a question to an AI assistant rather than a Google query. When a customer asks which agency you would recommend, you are either in the answer or you do not exist for them. Here is how to be in it.

Daniel Votruba · 6 July 2026 · 3 min read


GEO in brief: generative engine optimization is not about your position in a list of ten links, it is about being cited in the single synthesised AI answer, where you either appear or do not exist. Five things decide the citation: accessibility to AI crawlers, quotable answers in the text, structured data in JSON-LD, consistency of facts across the web, and authority off your own site.

What GEO is and why it should matter to you

Generative engine optimization (GEO), sometimes also called AEO or LLMO, is content optimisation for AI search engines: ChatGPT, Perplexity, Claude, Google AI Overviews. It differs from SEO on one fundamental point. SEO fights for a position in a list of ten links; GEO fights for a citation in a single synthesised answer. You can finish eighth in a list and still get a click. In an answer, you either appear or you do not.

And the reality on the ground? While preparing this site we went through sixteen leading agency websites. None of them was optimised for AI search engines. One of the largest e-commerce agencies we looked at actively blocks AI crawlers, so as far as ChatGPT and Perplexity are concerned it does not exist. That is a gap in the market of the kind SEO last offered fifteen years ago.

How an LLM picks who to cite

An AI search engine assembles its answer from sources it understands and trusts. In practice five things decide it:

  • Accessibility. GPTBot, ClaudeBot and PerplexityBot need a green light in robots.txt, and the content has to sit in the HTML rather than hidden behind JavaScript.
  • Citable answers. Paragraphs that answer the question on their own. The definition belongs right under the heading.
  • Structured data. JSON-LD (Organization, Service, FAQPage) serves a machine who you are, what you do and at what price.
  • Entity consistency. The same facts about the company on the site, in maps, on LinkedIn and in directories. Where there are contradictions, distrust follows, and with it no citation.
  • Authority beyond your own site. Mentions in the media, reviews, links. The model weighs who talks about you.

A checklist for the first month

1) Allow AI crawlers in robots.txt and add llms.txt, a structured company summary for language models. 2) Give every key page a citable definition and an FAQ section with FAQPage schema. 3) Align the facts everywhere the internet knows about you: Google Business Profile, LinkedIn, local business directories. 4) Write content that answers customer questions. Nobody asked for PR about how great you are. 5) Measure the baseline. Ask ChatGPT, Perplexity and Gemini the questions your customers ask and note who shows up in the answers. Then repeat it every month.

Incidentally, the site you are reading right now is built to exactly this checklist. We can build you one too, or check it yourself: in a month, ask ChatGPT about a growth agency in Prague.

Key takeaways (TL;DR)

  • GEO is not about a position in a list of ten links. It is about a citation in a single synthesised AI answer, where you either appear or you do not exist.
  • Of sixteen leading Czech agency websites, none was optimised for AI search engines. A gap in the market like SEO fifteen years ago.
  • Five things decide a citation: accessibility to AI crawlers, citable answers, structured data (JSON-LD), consistent facts across the internet, and authority beyond your own site.
  • Start practically: allow the crawlers in robots.txt, add llms.txt and an FAQ with schema. Then measure monthly who ChatGPT, Perplexity and Gemini mention in their answers.

FAQ

What is GEO?

Content optimisation for AI search engines and assistants (ChatGPT, Perplexity, Google AI Overviews). The goal is not a position in a list of links but a citation in the generated answer.

How does GEO differ from SEO?

SEO optimises for a list of links, GEO for a single synthesised answer. The fundamentals overlap, but GEO additionally requires citable definitions, FAQs, consistent company facts and accessibility for AI crawlers.

How do I know whether AI is citing me?

Ask the assistants the questions your customers ask and record the results monthly. It is the new share of voice, and almost nobody measures it yet.

How do you start with GEO in practice?

Four steps: allow AI crawlers in robots.txt, add structured data in JSON-LD, write answers that work when lifted out of context, and make the facts about your company consistent across your site and directories. Only then does off-site authority become worth chasing.

How many agencies are actually optimised for AI search?

In an audit of sixteen leading Czech agency websites in July 2026, none was. That is a situation comparable to SEO fifteen years ago: low-hanging fruit for anyone who starts before the competition does.

Sources: Our own audit of 16 Czech agency websites (July 2026) · GPTBot, ClaudeBot and PerplexityBot documentation · llmstxt.org.