Generative Engine Optimization
Generative Engine Optimization is the discipline that gets AI systems to name your brand when someone asks them for advice. Schwung is the GEO agency that first measures whether you are named, and then builds what influences it.
What is Generative Engine Optimization?
Generative Engine Optimization (GEO) is the discipline that gets generative AI systems such as ChatGPT, Google AI Overviews, Gemini, Perplexity and Copilot to name and recommend your brand in their answer.
Where SEO aims for a place in the list of links, GEO aims for the answer itself. That answer is the moment an AI makes a choice on your customer's behalf. If you are not in it, you do not exist for that customer, and unlike the blue links, there is no second page to appear on instead.
GEO or SEO? Three differences that matter
They overlap, a fast, well-structured site helps both. But treating GEO as "SEO with AI added" means optimising the wrong things.
Page versus passage
SEO rewards a page that is about something. GEO rewards a passage that can be lifted out verbatim: the definition first, a question as a heading, a figure with its source alongside. An excellent page without an extractable passage simply gets skipped.
Links versus mentions
SEO leans heavily on links to your site. GEO leans more on what is written about you elsewhere: trade press, directories, reviews and a brand entity that reads the same everywhere. Your own site is the confirmation, not the source.
Position versus mention
SEO is measured in positions, and tools exist for that. GEO is measured in whether you are named, which you only see by putting a fixed set of questions to the systems and repeating them. Without repetition you have a snapshot, not a measurement.
GEO is not geography
GEO here stands for Generative Engine Optimization and concerns AI answers. It has nothing to do with geotargeting, geofencing or aiming advertising at a location.
The confusion is common, because the same three letters mean different things in two different fields. If you are looking for location-based advertising, this is not the page you need. If you want to know whether ChatGPT names your brand, it is.
Measure first, build second
Baseline: are you named?
We put a fixed set of realistic questions from your market to the major AI systems, in a fresh conversation with no history, and record what happens for each one. We repeat that same set later, word for word, only then is the comparison valid.
- Which brands are named, and in what order?
- Which sources are linked?
- Is the picture the AI paints of you accurate?
Extractable, fact-dense content
AI systems quote what they can lift out. That means the answer at the top rather than after the introduction, questions as headings, and figures with a source in the same sentence. We usually start with pages you already have, rewriting pays off faster than adding.
- Answer first, introduction later or not at all
- Question-led headings and FAQ
- Your own figures with sources, not generalities
Brand entity and off-site reputation
Ask an AI for a recommendation and the answer rarely rests on one website. It rests on directories, reviews, trade press and a brand entity that is named consistently everywhere. That is the part outside your own control, and precisely why most agencies skip it.
- One brand entity, the same details everywhere
- Business profile and reviews in order
- Mentions and cases as external proof
Technical hygiene, without the hype
Schema markup, a fast accessible site, a robots.txt that admits AI crawlers: all of it belongs, and none of it costs much. But we do not sell it as leverage. Google says it does not use llms.txt, and research shows no measurable uplift from schema on citations.
- AI crawlers explicitly admitted
- Correct Organization and Service schema
- Fast, accessible, semantic structure
We measure the field, not just ourselves
In September 2026 we measured fifteen Dutch agency websites on exactly the signals GEO turns on. Seven had a working llms.txt. Two named AI crawlers explicitly in their robots.txt. And the share of pages opening with a direct answer rather than an introduction ranged from 3 to 67 per cent.
That last figure matters most, because it is the cheapest to change and carries the most weight. It is also exactly why a GEO engagement with us starts with rewriting rather than adding.
Frequently asked questions about Generative Engine Optimization
What is Generative Engine Optimization?
Generative Engine Optimization (GEO) is the discipline that gets generative AI systems such as ChatGPT, Google AI Overviews, Gemini, Perplexity and Copilot to name and recommend your brand in their answer. SEO aims for a place in a list of links; GEO aims for the answer itself, the moment an AI makes a choice on the customer's behalf.
What is the difference between GEO and SEO?
SEO optimises for a position in a list, GEO for inclusion in an answer. Three practical differences. SEO rewards a page that is about a topic; GEO rewards a passage that can be quoted verbatim. SEO leans heavily on links; GEO leans more on mentions and reputation beyond your own site. And SEO is measured in positions, GEO in whether you are named, an outcome you only see by asking the systems themselves.
Is GEO the same as geographic optimisation?
No, and the confusion is common. GEO here is short for Generative Engine Optimization and concerns AI answers. It has nothing to do with geotargeting, geofencing or aiming advertising at a location. If that is what you are looking for, this is not the page you need.
What does a GEO agency do?
A GEO agency first measures whether and how AI systems name you, then works on the three things that influence it: content that can be extracted, a brand entity that is consistent, and reputation beyond your own site. Note what is not included: nobody can guarantee a place in an AI answer, and anyone promising that is selling something they cannot deliver.
Do llms.txt or schema markup help with GEO?
Less than is often promised. Google says it does not use llms.txt, and controlled research shows no measurable uplift from schema on AI citations. We handle that hygiene properly because it costs little, but we do not sell it as leverage. The gains sit in extractable content and in what can be found about you elsewhere.
How do you measure whether GEO is working?
By asking the systems themselves, using a fixed set of questions you repeat over time. For each question you record which brands are named, in what order, and which sources are linked. Without that repetition you measure nothing: a single answer is a snapshot, and only the second measurement shows movement.
How long does GEO take to show results?
Expect weeks rather than days. Retrieval engines typically pick up new or changed content after four to eight weeks, and reputation beyond your site takes longer to build. GEO is not a switch but an accumulation, and the first gains usually come from rewriting pages you already have.
Who is GEO worth it for right now?
For brands whose customers decide after asking a question rather than after scanning a list of links: professional services, healthcare, education, B2B and anything where advice precedes the purchase. And for organisations in a field where nobody claims this yet, because that is where the head start is largest and cheapest.
Do you work with clients outside the Netherlands?
Yes. Distance matters little for GEO: an AI answer does not consider travel time, it considers what can be found about you. We work on this for brands across the Netherlands and beyond, from our office in Tilburg and remotely.
Ask Nick.
Curious whether AI names your brand, and what it would take to change that? Give us a call and we will run the baseline and see where the gains are.