"Make it easy for the AI!"

Ronny, I recently read: "SEO is dead - GEO is the future." - Is that true?

No, SEO - Search Engine Optimization - is of course not dead and never will be: The optimization of websites for AI searches, i.e. GEO (Generative Engine Optimization), is ultimately an extension of SEO, but not a replacement for it.

Without SEO, there is no GEO. Good search engine optimization is a perfect basis for a functioning GEO, because both overlap greatly in their basic principles - only the target system has evolved from classic search engines to AI systems.

Search engines still exist and provide data, while AI embeds the results in a thematic context, summarises them and personalises them according to the respective question.

So the two types of optimization complement each other?

Yes, exactly. With GEO, we are responding to the fact that many people now prefer to rely on a practical AI answer instead of clicking through lists of search results. Google now offers such an answer with Google AI Overviews (Google AIO), even before the actual search results. Search behaviour has changed.
With SEO, we had the goal of getting websites into the search engines, appearing as high up as possible in the search engine results and thereby attracting visitors to our websites. With GEO, we are now bringing the websites to AI in order to reach new users.

What does this mean for website operators, what should they do?

They should use both search forms. SEO includes the technical optimization of websites as well as good content - in other words, good content that uses relevant keywords and shows expertise in the respective topic.

GEO makes content easy to understand for LLMs - which stands for Large Language Models. The aim here is for the AI to generate answers to search queries from the website content. And this is easiest if it can quote the content. If you are cited by the AI, this in turn leads to traffic on your own website.

Are there companies or sectors for which GEO is particularly relevant?

In general, everyone benefits from it. I would particularly recommend GEO to companies in e-commerce, local businesses and service providers as well as industries in the fields of health, finance and law.

Search behaviour has changed as a result of AI - many people now ask precise and detailed questions to the AI assistant. Instead of "Pizza Kreuzberg", people now search for "Best pizzeria with vegan options on Oranienstraße with dishes for under 12 euros". The winners here are those who are cited as a credible source in the AI answer.

So the search engine landscape is changing - probably with consequences ...

Yes, many website operators have seen a drop in traffic due to the use of ChatGPT & Co. because users no longer click on the actual search engine results. Those who generate revenue through advertising on their own website, for example, are of course particularly affected by this.

Ok, so you should make sure that you also appear in the AI responses. How can you design content so that it is favoured as a source by AI systems?

By making the AI's work as easy as possible. After all, the aim is to be cited - so you have to design the content in such a way that it is citable.

I also address these top topics in my GEO training courses:

• Clear language and precise answers: Technical jargon should be avoided or explained wherever possible so that the context of the page content is clear to the AI and ultimately the user.
• Direct naming of entities: Clear language also applies in particular to the naming of people, things, concepts or events. Technical terms or marketing language should also be avoided here.
• Focus on topical authority: Those who cover topics comprehensively and competently on their website distinguish themselves through their expertise.
• Use of abstracts: A brief, generally understandable overview of a topic in just a few sentences - also available as a bullet point list - offers added value for both website visitors and AIs.
• FAQ: If the "important questions" on the respective topic are formulated and compiled as W-questions, they correspond (with a bit of luck) to the questions asked to the AI - answers can then be transferred directly.
• Chunking: "Chunking" refers to the division of longer texts into individual sections. If possible, this is done in such a way that each section of text can be understood and quoted individually.
• Use of structured data: Structured data according to Schema.org helps to clearly understand content on a website. It improves machine interpretability so that AI systems can categorise content more precisely and use it more frequently in their responses.

REQUEST GEO COUNSELLING NOW

Clear language for entities ... can you explain this in more detail? Or do you have an example?

A restaurant has a menu on its website with a special choice of words to make the dishes look more interesting. So instead of "calzone", it says "thermally sealed pastry shell with multi-component filling" ... For example, if someone asks ChatGPT for "best calzone in Kreuzberg", what are the chances of being quoted by the AI?
Entities - and this also includes offers from a menu - should therefore be clearly named.

Speaking of which: what role do user-generated content and social media play for GEO?

User-generated content - everything that others write about a product or service - and social media are becoming even more important in the wake of GEO: the data that an AI finds in product reviews, forum discussions or social media is valued as an authentic user experience.

Helpful entries can even be used as AI answers. Appearing in high-quality external sources therefore signals relevance. This in turn increases the chance that your own website will be cited.

You should still keep an overview here: Negative reviews or critical discussions are also considered authentic! You should therefore be active in reputation management and respond to criticism.

What advice would you give to companies or organisations starting out with GEO today?

First of all: just as with search engine optimisation, they should also be patient when it comes to "optimising for AI" - it doesn't work overnight. However, small changes are a good foundation on which to build. A first step could be, for example, to insert a quotable FAQ into the pages, another to back the whole thing up with structured data.

Existing content should be revised for clear language and precise answers, and abstracts should highlight the key points of the entire content - this is how you work your way through the possible optimisations.

Nevertheless, don't forget SEO! Those who cannot be found in search engines are generally less likely to end up in the training data of LLMs, i.e. AI applications. This is because SEO is an essential building block for GEO.

Ronny, a look into the future: How do you see the role of search engines in five years' time? Will AI assistants replace Google Search?

I assume that we will completely outsource many of the more precise search queries to AI chats. Search engines will continue to be used for simple facts and quick searches (searching for the weather, a brand, service, etc.), while AI will be used for larger, more complex searches, programming assistance or more detailed product searches.

AI assistants will therefore not replace search engines. But they will change how and, above all, what we search for.