Becoming visible as a service provider in ChatGPT: What we can learn from the “twerk button” incident

ChatGPT, Perplexity and other AI systems now regularly name specific brands and service providers. But how do these recommendations actually come about? How does an AI recognize which providers are relevant – and why are some companies taken into account while others are missing entirely?

This is exactly where GEO (Generative Engine Optimization) comes into play. Artificial intelligence is changing the way people search, find and recommend. But once again it turns out that SEO and GEO are closely connected. GEO is based on classic SEO principles, only that the “search engine” today is a language model.

Anyone who wants to be mentioned in answers from ChatGPT, Perplexity or Google Gemini has to understand how these systems recognize, evaluate and link content. What is exciting here: many of the old SEO tricks from the early 2000s suddenly work again – just on a new stage.

Why some service providers are taken into account by ChatGPT & Co. – and others are not

Whether agencies, tools or brands: some are regularly mentioned by ChatGPT and other AI systems, while others do not appear at all despite strong performance. This has less to do with the quality of the offering and more with where the models get their information from.

Many mentions come from publicly accessible sources such as industry directories, review platforms or editorial articles. Anyone who is present there has significantly better chances of also being taken into account in the answers from ChatGPT & Co.

Practical examples: How service providers appear in ChatGPT results

Current ChatGPT answers clearly show how recommendations even come about. SEO colleagues, sorry, I mean SEO service providers such as Claneo or Kosch Klink Performance are currently often named by ChatGPT. The reason lies in the way ChatGPT sources information: the model draws on publicly available websites classified as trustworthy. These primarily include industry directories, review portals, media reports and thematic best-of lists.

A particularly exciting example is OMR Reviews. Many agencies are listed there, often with ratings, service descriptions and customer testimonials. ChatGPT recognizes this site as a reputable source and draws on it as evidence for its statements. In practice this means: anyone who is visible on OMR Reviews is more likely to be identified by ChatGPT as “recommendable” – simply because this platform is structured and frequently linked.

ChatGPT itself states which sources its statements come from. This disclosure is usually shown below the answer and reveals which domains the model trusts. For Claneo, for example, omr.com appears there as a reference source – a clear indication that the mention comes directly from this platform.

This illustrates how strongly GEO (Generative Engine Optimization) leans on the logic of classic SEO signals. In the past the rule was: anyone represented on relevant, well-linked pages had higher chances of top rankings on Google. Today the same principle applies – only that it is no longer the search engine but an AI that searches through the sources, evaluates them and summarizes them in text form.

Why old-school SEO techniques are trending again

If you look more closely, the principle strongly resembles the early years of search engine optimization. Around 2000, the DMOZ directory (Open Directory Project) was one of the most important factors for good rankings on Google (source). Anyone listed there was automatically considered reputable, because the entries were editorially reviewed and were regarded by Google as a mark of quality. Numerous companies invested a lot of time and effort back then to get included in this one directory – not because of direct visitors, but because of the SEO signals it conveyed.

Today this pattern repeats itself on a new level. Instead of DMOZ, it is platforms like OMR Reviews, Clutch or Sortlist that are interpreted by AI systems like ChatGPT as trustworthy sources. The logic is identical: visibility is created through third parties, not through your own website. Anyone who appears in these ecosystems is perceived as relevant and cited more frequently.

With this, a fundamental principle from the early SEO era returns – “authority through external reference”. The only difference is that the decision about who is visible is no longer made by the Google algorithm, but by a language model. GEO thus brings mechanisms from back then into the present, only with new tools and different rules of the game.

When service providers recommend themselves: your own best-of lists as a source

But it is not only external signals that play a role. Some service providers actively help their visibility in AI searches along by publishing their own best-of lists – for example with titles like “The best SEO agencies in Germany”. At first glance such posts appear editorial, but they often come directly from agencies or marketing platforms themselves.

What is interesting: ChatGPT does not recognize these lists as self-promotion, but as a structured information source. When a page lists several agencies, adds short descriptions and presents the whole thing as a “ranking”, the AI interprets it as a trustworthy overview. The result: this content is cited and adopted into answers as if they were independent recommendations.

One example of this is the list from Keyweo:

Various SEO agencies are listed there – including Keyweo itself. Exactly such pages then show up in ChatGPT's source citations.

For companies, this can be an effective way to be present in AI results. The decisive factor here is that the page is cleanly structured, comprehensible in content and worded as neutrally as possible. Anyone who complements their own list with other credible agencies or brands increases the chance that the AI classifies the content as a reputable source – and thereby their own brand along with it.

Why all of this is far too easy – and opens the door wide to spam

The current situation shows how susceptible AI systems are to manipulation. ChatGPT and other language models adopt information from freely accessible sources without editorially checking or verifying it. Anything that is structured, frequently linked or mentioned multiple times can theoretically be classified as “trustworthy” – even if the content was created or manipulated by a company itself.

This opens the door wide to spam, automated rankings and pseudo-sources. Even now, countless websites with self-made best-of lists are circulating that are aimed solely at appearing in ChatGPT's answers. So far, AI models can barely tell the difference between genuine expertise and clever self-marketing.

An example of how easily such misinformation is adopted comes from the gaming sector: a made-up report about a supposed “twerk button” in GTA 6 was passed on by an AI search engine as fact, because it was based on a well-structured but false source. The GTA 6 incident shows how easily AI models can be influenced when content appears often enough on the web.

A YouTuber spread his made-up story across various platforms over the course of weeks – and the AI eventually adopted it as seemingly reliable information.

This clearly shows: as long as there are no clear quality filters or source evaluations for generative models, GEO remains susceptible to manipulation – and opens the same gaps that already occupied SEO almost 20 years ago.

Conclusion: When optimization becomes manipulation

Clear guidelines or technical safeguards to reliably detect such manipulation are still lacking. On the international stage, the term “black hat GEO” is therefore already appearing (https://searchengineland.com/black-hat-geo-pay-attention-463684). What is meant is the targeted influencing of AI results through manipulative tactics – for example through artificially generated content, fake directories or automated sources.

For reputable companies, this is an opportunity to position themselves early and strategically. At the same time, it carries a risk: if AI systems adopt content in an uncontrolled way, inaccurate or manipulated information can also shape a brand's image in the long term. GEO is thus at a point where transparency, authenticity and quality control are more decisive than ever before.