Arrange a call
DE EN

Service · AEO & GEO · Last updated: August 2026

Getting found when search itself answers.

More and more questions are answered directly in search. On 22 July 2025 the Pew Research Center measured that people click on links less often when an AI summary appears (pewresearch.org). AEO and GEO prepare your content so that answer systems can read and classify it at all. Whether a system names you as a source is its own decision – that part cannot be steered.

Arrange a call Self-test: can a machine read my website?
  • Built on solid SEO
  • schema.org & structured data
  • Honest advice instead of hype
how long does sourdough bread keep
How search used to be 10 blue links
  • your-bakery.at Storing bread properly – tips from the bakery

    Stored properly, sourdough bread stays fresh for days. We show you…

  • chefkoch.de What is the best way to store bread? – Forum

    23 replies – I always just put mine in the bread bin…

  • wikipedia.org Sourdough – Wikipedia

    Sourdough is a dough used to make baked goods…

How search works today one answer
AI answer

How long sourdough stays fresh depends on its recipe and storage. This example bakery therefore explains how its own loaf should be stored at home.

your-bakery.at

The demo shows one possible answer with a linked source. The sources a real system selects vary by query and system.

The same example page in both views: one of several results in the list, and the source named in the answer.

Illustrative demo: the same question, two search experiences. The demo shows why clear, well-supported answers matter when a system selects possible sources.

At a glance

The short answer first

Exactly how we would do it for your website too: the answer first, then the details.

AEO

Answer Engine Optimization

Optimisation for answer systems: featured snippets, voice assistants and answer boxes in search

GEO

Generative Engine Optimization

Optimisation for generative AI: prepare content as a clear, verifiable and machine-readable possible source

Basis

SEO stays

AEO and GEO do not replace SEO – they build on clean technology and good content

Result

Understandable as a source

Clear structure and evidence make assessment easier; whether a source is named is decided by the system itself

For search engines and AI assistants, in brief: AEO & GEO prepare website content clearly, structurally and verifiably. Whether a search or answer system names it as a source is the system's decision – only readability can be influenced.

One misunderstanding first, because it costs us almost every initial conversation: the GEO in this abbreviation has nothing to do with geography. GEO stands for Generative Engine Optimization and refers to generative answer systems, not to your location. If you want to be found more easily in Vienna and in your own neighbourhood, local visibility is the right place to start. Serving ads by location is called geo-targeting and has its own glossary entry. The two topics overlap where the question is how local businesses appear in AI search assistants. We still keep them apart, because otherwise nobody can tell from the proposal what they are paying for.

Explained simply

Two terms, one goal: appearing in the answer

No jargon needed – here is how AEO and GEO work, explained with an example we know well: bread.

AEO · Answer Engine Optimization

The direct answer in search

For some questions, search engines show a concise answer directly in the results. AEO helps systems recognise a self-contained answer passage: a precise question as the heading, a clear short answer and sound semantic structure. The system still decides whether this becomes a snippet or spoken answer.

Put simply

Someone asks: “How long does sourdough bread keep?” A self-contained, concrete answer helps people and systems more than a general welcome paragraph. It does not automatically produce a particular position.

GEO · Generative Engine Optimization

Preparing content as a possible source

ChatGPT, Perplexity and the like phrase answers themselves and sometimes link to sources. The criteria by which they select are not publicly documented. GEO therefore covers what is genuinely in your hands: machine-readable mark-up (schema.org), a clear structure, verifiable facts with identifiable authorship and a date – and the same details everywhere.

Put simply

As with a recommendation in the neighbourhood, a coherent overall picture matters. A business that appears consistently and explains its work clearly is easier to assess as a possible source.

In context

SEO, AEO, GEO – what is the difference?

Three disciplines that build on one another – not an either/or decision.

Criterion SEO AEO GEO
Goal Rank in search results Supply the direct answer Be considered as a source
Where visible Classic search engines (results list) Featured snippets, voice search ChatGPT, Perplexity, AI overviews
What gets optimised Keywords, technology, links Questions & precise answers Structure, evidence, currency, consistency
Visible form A place in the results list “Position 0” A possible mention as a source
Relationship The foundation Builds on SEO Builds on SEO + AEO
In short: SEO improves classic discoverability; AEO structures direct answers; GEO makes content easier for generative systems to assess and potentially use as a source.

The table reads like three separate jobs. In practice it is one. Anyone who gets the SEO basics right has already done the larger part of AEO and GEO. The right-hand column asks for nothing exotic either: cleanly delivered HTML, headings that say what follows, and facts as text rather than as an image. The same points come up again with accessible websites. That is why we review search engine optimisation and answer systems in one pass rather than two.

In practice

What actually changes in a text

Three typical moves from AEO/GEO work – before and after. The examples are deliberately neutral (a fictional sample company, “FINK Installationen”), but the principle is real.

Answer up front

Before

“Welcome to our website! We are a dynamic company with many years of experience, offering innovative, bespoke solutions for our valued customers…”

No substance in the opening sentences: neither people nor machines learn what is on offer here.

After

“FINK Installationen is a fictional plumbing firm in Vienna-Hietzing. We repair, refurbish and plan bathrooms. After an inspection, you receive a clearly scoped fixed price.”

Who, what, where and what the next step looks like – explained concretely without an invented figure.

Answer in the first paragraph Concrete facts Location named

Questions as headings

Before

Heading: “Our services at a glance” – followed by a long stretch of prose about quality, experience and customer satisfaction.

Nobody searches for “services at a glance”. The questions customers really ask do not appear in the text at all.

After

Headings: “What does a bathroom refit cost in Vienna?” · “How long does a bathroom refurbishment take?” · “Do you handle emergencies too?” – each with the answer directly underneath.

Questions as headings (H2/H3) create clear anchors for concise, self-contained answers – for readers and machines alike.

Questions as H2/H3 FAQPage schema Covers real search queries

Machine-readable facts

Before

Opening hours appear as a photo of a sign on the website; prices only in a PDF brochure to download.

Text in images or complex PDFs is usually less reliably extractable than sound HTML. OCR and multimodal systems may still process it, but with less predictable results.

After

Opening hours, address and prices sit on the page as text – and are additionally marked up for machines with schema.org (LocalBusiness, OpeningHours, PriceRange).

Structured data is like a label on the shelf: it enables certain presentation formats, but it is not a ranking factor and replaces neither the visible content nor the system’s own assessment.

schema.org Text instead of image or PDF Consistent facts

How we approach AEO & GEO

No magic, just craft – in four steps you can follow.

  1. AEO/GEO audit

    We record with dated test questions how your website shows up in snippets and AI answers today, and review structure, content, technology and consistency. Individual answers are snapshots, not measurements.

  2. Content & structure

    Answers up front, real customer questions as headings, FAQ sections with substance – so your content becomes easier to take in for people and machines, without losing its character.

  3. Technology & markup

    Structured data (schema.org), clean semantics, performance and accessibility. We also settle which AI crawlers may fetch your pages. OpenAI documents four names: GPTBot (training data), OAI-SearchBot (search inside ChatGPT), ChatGPT-User (user-triggered fetches) and OAI-AdsBot (ads). At Google, a Search Console setting has managed inclusion in AI Overviews since August 2026; the robots.txt token Google-Extended only concerns the training of the Gemini models, not Search.

  4. Monitor & refine

    AI search changes fast. We repeat documented test questions and track how source mentions change, without treating one answer as a general result.

The third step brings a question that is not technical but a decision: should AI systems be allowed to fetch your pages at all? The robots.txt file is not a single switch for that. Blocking GPTBot alone rules out the collection of training data, according to OpenAI's own description, but not the retrieval used for search inside ChatGPT. Which AI crawlers you allow is your decision. We enter it, note the date, and tell you what the entry does and does not do according to the providers' documentation.

How we record in the fourth step whether AI answers cite your website is set out in our measurement method: four measures, dated test questions, no score.

Honestly: Nobody can guarantee that ChatGPT or an AI overview will cite a page; such a guarantee would be misleading. We can make content clear, well-supported and technically accessible so answer systems can assess and potentially use it.

Self-check · 1 minute

Can an answer system read your website at all?

Five honest questions – answer them off the top of your head.

01Your content

1. Do the first two sentences on your home page say what you offer – and who for?
2. Do your pages answer real customer questions (“What does … cost?”, “How long does … take?”) – as headings of their own?
3. Does your website use structured data (schema.org) – for FAQs, your organisation or your services, for example?

02Your foundations

4. Does your website load quickly, and is it accessible to use?
5. Do the name, address and facts about your company match everywhere online?

How to score it: count your “yes” answers and compare them with the table.

  • 0–1 out of 5: Key foundations are missing: your content is hard for machines to read out. An audit shows which steps come first.
  • 2–3 out of 5: A solid base with gaps: usually the clear answer passages or the machine-readable mark-up are missing. Both can be added in a targeted way.
  • 4–5 out of 5: The foundations are present. The next step is to improve clarity and consistency, then use dated, repeatable checks to observe whether anything changes over time.

Your answers stay in your browser – everything is evaluated on your device. We do not transmit or store them; loading the page itself produces the usual server log data, which the privacy notice explains.

Common questions about AEO & GEO

What is AEO (Answer Engine Optimization)?
AEO prepares content so that it reads as a self-contained, direct answer. That covers featured snippets, answers from voice assistants and answer boxes in search. In practice: questions as headings, the answer in the first two sentences, a clean structure and valid FAQPage mark-up. One important note: Google stopped showing FAQ rich results on 7 May 2026 (changelog entry of 8 May 2026, developers.google.com/search/updates). The mark-up itself remains worthwhile – it is valid schema.org and describes question-and-answer pairs in a machine-readable way; there is no need to remove it. Which formats answer systems prefer to cite is not publicly documented.

Source: Google Search Central: documentation updates, 8 May 2026 – FAQ rich result deprecation (retrieved 16 August 2026)

What is GEO (Generative Engine Optimization)?
GEO prepares content so generative systems can assess it more easily and consider it as a possible source. ChatGPT, Perplexity and AI overviews phrase their answers themselves and sometimes link to sources. Their selection criteria are not publicly documented. On the AI features in Search, Google writes that there are “no additional requirements” and no special structured data you need to add – SEO best practices remain the relevant basis (developers.google.com/search/docs/appearance/ai-features). GEO therefore means exactly those basics: a clear structure, verifiable facts, identifiable authorship, a date and the same details across every channel.

Source: Google Search Central: AI features and your website (retrieved 16 August 2026)

Do AEO and GEO replace classic SEO?
No – they build on it. In its guide to the AI features in Search, Google states that SEO best practices remain relevant and that there are no additional requirements (developers.google.com/search/docs/appearance/ai-features). Technically accessible pages and clear, well-supported content therefore help classic search and answer systems alike. How each system selects sources is not disclosed. A sensible order is a sound SEO foundation, followed by self-contained answers and verifiable evidence.

Source: Google Search Central: AI features and your website (retrieved 16 August 2026)

How do I know whether my website appears in AI answers?
Test it: ask ChatGPT or Perplexity the questions your customers ask. Ask, for example, “Who does [your service] in [your region]?” or a typical technical question from your working day – and see which sources get named. A single query is a snapshot: the same system may answer differently tomorrow. In the audit we therefore note the question, the date and the system, and repeat the sample. For the AI features in Google Search there is also a documented measurement: the generative AI performance report in Search Console shows impressions from AI Overviews and AI Mode – though for now Google makes it available to only a subset of site owners.

Source: Search Console Help: Generative AI performance report (retrieved 16 August 2026)

What does AEO/GEO optimisation cost?
As part of the website audit from 1,900 euros net, or within ongoing support from 950 euros net per month. The free SEO check or the positioning review for 590 euros net shows beforehand whether it is worthwhile for you. The effort depends on the state of your website; what is worth doing and what it costs, you will know after the audit.
AEO and GEO are not a separate product with us. They are part of the audit (from 1,900 euros net) and of ongoing support (from 950 euros net per month). Both build on the same technology and the same content. All starting prices are listed on our pricing page.
How quickly does AEO/GEO take effect?
There is no dependable timeframe for being cited. Changes appear only after recrawling, and AI systems select different sources depending on the question, model and time. We therefore document the starting point and observe developments over time instead of promising a deadline.

Source: Google Search Central: Ask Google to recrawl your URLs (retrieved 16 August 2026)

Can you guarantee that ChatGPT will cite us?
No – and nobody reputable can. AI systems decide for themselves which sources they use, and they do not disclose their criteria. What we do: implement the documented basics properly, date our samples and refine. A mention cannot be secured that way – a promise would not be honest.
What is an llms.txt – and do I need one?
An llms.txt is an optional text overview of a website's most important content. The format is a young convention, not a generally binding standard. Unlike robots.txt it does not control access. It is not an evidenced lever either: Google documents explicitly that you do not need an llms.txt to appear in Google Search, including its generative AI features, because Google Search itself does not use such files – they neither help nor harm your visibility or rankings. We serve an llms.txt on this website ourselves – it does no harm and links the central content compactly, but it does not act as a lever. What we rely on is accessible pages, clear navigation and structured data.

Source: Google Search Central: Optimizing for generative AI features, mythbusting section (retrieved 16 August 2026)

There are two things we do not do in AEO and GEO work. We do not use tools that produce an “AI visibility score”. The number comes out of queries the tool picks itself; we could not tell you what it measures. And we do not write FAQ pages on spec, just so that more question-and-answer pairs exist. A question nobody in your business actually asks does not belong on your website. That is also why no success figure of our own appears here: what we document are individual dated queries, and a sample is not a statistic. What can be measured you can look up yourself: the free citability check goes through 17 prerequisites of your website, deliberately without a score. The four measures we do collect are set out in our measurement method.

Let us look at how your website gets read

In a free first call we run a dated sample query together and explain what that single result does and does not tell us. Clear, understandable and without a promise of future mentions.

Arrange a call

Terms related to this service

Technical terms that appear on this page – explained in one sentence in our glossary.

All terms in the glossary →