GEO is not a new channel. It is SEO, done right.

More and more buyers no longer research on Google. They ask ChatGPT, Perplexity or Google’s own AI overviews. So the question is no longer only whether you rank, but whether the AI names you when someone asks about your category. Whoever is in that answer wins the buyer. Whoever is missing does not exist for that buyer.

The good news: you do not have to learn a new discipline for this. Google made that clear in its AI search guide, and the whole market around “GEO hacks” is currently sorting itself out. This article shows you where you stand, what actually matters, and what you can do this week.

The essentials:

  • Your AI visibility is measurable. You do not have to guess whether ChatGPT names you.
  • GEO is, to a large extent, good SEO. Google says so itself: what ranks classically feeds the AI answers.
  • The real lever is not in your own text, but in the sources the AI quotes from. That is where you need to show up.

First, find out where you stand

Before you optimize anything, work out where you stand in GEO today. I built a free AI-Visibility Check for exactly this. It asks ChatGPT, Perplexity and Google the real buyer questions in your category and checks whether you show up in the answers.

Here is how: open the check, enter your brand, your category and up to three competitors, and start it. After about a minute you get a number and an honest verdict, not 40 metrics.

GEO is not a new channel, it is good SEO

Here Google clears up the biggest misunderstanding itself. The official guide says, word for word: “Optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” The AI overviews and AI mode draw on the same ranking systems and the same index as classic search. What ranks well classically, the AI sees.

That has two practical consequences.

First, you can skip most “GEO hacks”. Google explicitly says you need no llms.txt file, no special schema markup and no content chopped into tiny pieces to show up in AI search. Anyone trying to sell you an expensive GEO specialty is, for the most part, selling you SEO with a new label. Building separate page variants per search query is actually harmful, Google files that under spam.

Second, the content itself counts. AI chatbots reward first-hand reports, opinions and authenticity. An article with the headline “Why we waived the inspection and saved money” beats “7 tips for first-time buyers”, for example. Your own point of view and real experience beat interchangeable lists that could come from any source.

Two mechanics are worth understanding, because they explain why this feels different from classic Google:

  • RAG. The AI does not invent its answer, it pulls relevant pages from the search index live and summarizes them. So ranking well classically means being seen by the AI.
  • Query fan-out. From the one question you ask, the AI generates several related queries in the background, sent at the same time. From “remove weeds” it also fires “best herbicides” and “remove without chemicals”. For you that means: one page that answers the whole question cluster beats ten thin pages that each cover one variation.

The three levers that move your AI visibility

You can pull on three levers to improve your AI visibility.

Get quoted and linked by other websites and sources. AI chatbots look for “social proof”, confirmation that other groups consider your content good. A Semrush study backed this up and found that 62% of AI citations do not lead to a brand mention at all. So the AI constantly pulls from pages that are not yours. Check how present you are in other sources. What helps: a strong G2 profile, an honest answer in the right Reddit thread, a guest post in the industry blog the AI already reads. You are not optimizing your page, you are optimizing your presence in other people’s sources.

Answer the buyer’s actual question right in the first paragraph. The AI lifts things out of context. If your first paragraph does not answer the buyers’ question on its own, it gets skipped. Write the first paragraph of your articles so that it takes up the buyers’ target questions literally and answers them immediately. Best of all, also cover the “neighbor questions” with a short FAQ at the end, so the further questions the AI asks are answered in the article too.

Deliver the signals AI reads as authority. From the research on AI answer engines: expert quotes raise the likelihood of being cited by about 41%, concrete numbers by about 30%, built-in source links by about 30%. So: name tools instead of “a tool”, write “in 8 out of 10 cases” instead of “often”, quote a real expert, and link your evidence. If you want to produce your own blog posts for this faster and in your voice, I wrote up how that works in 5 minutes here.

What is different in DACH

Three things are specific to us in the German-speaking market.

First, the language. Your buyers ask the AI in German, and they use the German terms of your category. Write for exactly those queries, not for the English translation of them.

Second, the sources. In DACH the AI answers often pull from different places than in the US market, more from German industry media, trade portals and comparison sites. The check shows you which ones those are in your category.

Third, and this is where it gets strategic: in AI answers it increasingly matters less what your brand name is, and more whether you answer the question best. t3n put it sharply, “a system does not ask for the logo”. For established brands that is a warning, for smaller ones it is the chance to pull ahead of big names.

One side note: the llms.txt file that Google ignores is actually read by Claude and Perplexity. It is not mandatory, but for those two it is a small, free bonus with no downside.

FAQ

Is GEO the same as SEO? To a large extent, yes. Google itself says AI search builds on the same ranking and index systems as classic search, and explicitly recommends putting effective SEO strategies above GEO hacks. There are a few add-ons for the non-Google engines like Claude and Perplexity, but there is no separate channel you would have to run in parallel.

Do I need an llms.txt file? For Google, no, that is stated in the official guide. But Claude and Perplexity do read it. So it is not a high-impact lever, just a small bonus with no downside if you can set it up quickly anyway.

Why does ChatGPT name me but Perplexity does not (or the other way round)? Because each engine weights differently and quotes from different sources. Perplexity, for example, puts a lot of weight on built-in source links and recency. That is exactly why the check shows you the result per engine, so you see where to act specifically.

How often should I check my AI visibility? Once a quarter as a routine, and additionally whenever you have worked on one of the quoted sources. That way you see whether your work moved your rate.