GEO optimisation for generative search engines
in Zurich
In Zurich, decisions about professional services are rarely made after the first search term — prospective customers compare, explore further and ask complex questions, increasingly turning to generative systems such as ChatGPT, Google AI features, Copilot or Perplexity. GEO — Generative Engine Optimization, or, in plain English, optimisation for generative search engines — prepares your content so that these systems can find it, put it into context and potentially use it as a source. We support you from our base in Basel: online meetings are the norm, and we arrange in - person appointments when needed. We do not claim to have an office in Zurich. Inclusion or citation in an AI answer can never be guaranteed.

Initial analysis before any measures
The initial analysis establishes which complex decision - making questions your target audiences in Zurich ask, whether your website is accessible to relevant crawlers and whether your services, target audiences and business information can be unambiguously associated by machines. This leads to prioritised measures — justified rather than generic.
GEO for businesses from Zurich
Professional services sell through clarity, not loud promises: someone who immediately understands what you offer, who it is for and the standards against which you are willing to be judged is more likely to make a relevant enquiry. Generative systems also favour unambiguous information — clear positioning helps both people and machines.
Die Zusammenarbeit läuft digital und bleibt persönlich: Sie sprechen von der Analyse bis zur Auswertung mit denselben Personen. Termine vor Ort in Zürich sind to Vereinbarung möglich, ein lokales Büro unterhalten wir nicht.
From search terms to complete questions
Traditional search starts with short terms, while generative search starts with complete questions: ‘Which agency in Zurich supports a relaunch with a redirect plan?’ rather than ‘agency Zurich’. Content that answers such complex questions in full is more useful to generative systems than keyword lists.
The demo gradually turns a short example search into a complex question and shows what information an answer requires. It does not call a real AI system.
Interactive GEO example
Clearly distinguish between services
Combining all your offerings on one page makes it harder for both visitors and generative systems to identify what belongs where. Separate pages for each service, with precise descriptions, target audiences and approaches, create clarity. This is not about multiplying pages for their own sake, but about answering different questions.
The demo creates a comparison table from service features and shows how differentiation is reflected in a machine - readable structure. It does not create a real page.
Interactive GEO example
Comparisons and decision - making aids
Comparative questions are central to generative search: ‘What is suitable for which purpose?’ Content that compares options using clear criteria can serve as a source for these answers — although selection cannot be guaranteed. Honest assessment matters more than self - praise.
The demo lets you select a target audience and shows how the same factual information changes for different groups of readers. The presentation remains a local example.
Interactive GEO example
Present expertise in a verifiable way
Fachwissen wirkt für generative Systeme überzeugender, wenn es überprüfbar ist: benannte Autorenschaft, to vollziehbare Quellen, sichtbare Aktualität. Hilfreiche, überprüfbare Inhalte sind wichtiger als künstliche Keyword Wiederholungen.
The demo combines example facts with visible sources and shows how a statement becomes more robust. It does not claim to show actual citations by an AI system.
Interactive GEO example
A search query becomes a decision - making question with sources
Example · GEO in Zurich
- 1
Eine allgemeine Suchanfrage
- 2
Wird zur komplexen Entscheidungsfrage
- 3
Ein ehrlicher Vergleich
- 4
Mehrere to vollziehbare Quellen
- 5
Possible AI-powered answer
Keep company information consistent across pages
Service pages, the contact page, the company profile and directories should state the same core details: name, services, target audiences and ways to get in touch. Contradictions make it harder for machines to identify the company correctly and establish trust.
The demo shows example company information in five example locations and lets you identify conflicting details and correct them locally. It does not change any real profiles.
Interactive GEO example
Generative Systeme unterschiedlich beobachten
ChatGPT, Google KI Funktionen, Copilot und Perplexity arbeiten unterschiedlich: Verschiedene Systeme können unterschiedliche Quellen verwenden, und Antworten können sich je to Modell, Zeitpunkt, Standort und Frage verändern. Eine Beobachtung lohnt sich deshalb pro System.
The demo switches the possible answer presentation between the four systems. The platform views are neutral explanations, not actual answers from the systems.
Interactive GEO example
Develop content in response to new questions
Generative Suche verändert sich laufend — neue Fragen entstehen, Systeme ändern ihr Verhalten. Wir entwickeln Ihre Inhalte anhand der beobachteten Fragen weiter und priorisieren Massnahmen to Klarheit, Aktualität und technischer Zugänglichkeit.
Die Demo priorisiert Beispielmassnahmen to diesen drei Kriterien. Eine einzelne Abfrage ist keine belastbare Erfolgsmessung; alle Werte sind Demodaten.
Interactive GEO example
Was GEO bedeutet
GEO stands for Generative Engine Optimization — in plain language, optimisation for generative search engines. This page covers the potential visibility of companies, websites and content in generative search systems and AI-assisted answers.
Traditional SEO fundamentals remain important: content must be crawlable, indexable and understandable, and visible content must match structured data. GEO complements these foundations with clear answers, verifiable facts, sources and unambiguous machine identification of your company information.
Four generative systems at a glance
- ChatGPT Suche
- ChatGPT's search functions; sources may be displayed.
- Google AI features
- AI Overviews and AI Mode; the usual SEO fundamentals apply.
- Microsoft Copilot
- Answers based on Bing; citations are not rankings.
- Perplexity
- Web search with citations; the system makes the selection.
The systems shown are examples of generative search. Inclusion or citation cannot be guaranteed. There is no partnership between Widmer & Kunz and the providers named.
GEO service areas at a glance
These areas are based on the confirmed services in the web design and marketing quotations — from topic research and technical accessibility to reporting and training. Technical statements follow the official documentation from Google Search Central, OpenAI, Bing and Perplexity.
- Initial analysis
- Review of existing visibility, the questions being asked and the relevant systems.
- Crawler and robots.txt check
- Separate settings for search and training crawlers, based on official documentation.
- Indexierbarkeit
- Content must be crawlable, indexable and understandable — the shared foundation with SEO.
- Sitemap and canonicals
- Unambiguous addresses and clean sitemaps without duplication.
- Eindeutige Unternehmensinformationen
- Consistent name, address, services and contact methods across all channels.
- Services and target audiences
- Each service clearly described and assigned to a target audience.
- Questions and clear answers
- Specific questions answered fully and without promotional language.
- Fachinhalte
- SEO copy, blog articles and specialist articles that answer real questions.
- Sources and evidence
- Verifiable evidence for key statements.
- Authorship and responsibility
- Visible details of the author, organisation and responsibility.
- Up-to-dateness
- Truthful review dates and regular updates.
- Structured data
- Markup for visible content — without fabricated details and without any guarantee.
- Internal linking
- Related content linked in an understandable way.
- External mentions
- Cannot be artificially forced; well-maintained content of your own is the foundation.
- Regionale Informationen
- Ehrliche Gebietsangaben ohne erfundene Standorte.
- Multilingualism
- Complete, well-maintained language versions rather than automatic translations.
- Bildbeschreibungen
- Alternative text and graphics that meaningfully complement the content.
- Tables and comparisons
- Machine-readable comparisons using truthful criteria.
- ChatGPT Suchzugriff
- OAI SearchBot must not be blocked if content is to be eligible to appear in ChatGPT search answers.
- Google AI features
- Normal SEO fundamentals apply; no special GEO file or AI markup is needed.
- Bing and Copilot
- Crawling, indexing, clear canonicals and IndexNow as technical foundations.
- Perplexity
- PerplexityBot respects robots.txt; source selection is up to the system.
- Quellenbeobachtung
- Structured monitoring of defined questions over periods of time.
- Referral Traffic
- Measurable visits from generative systems as a supplementary indicator.
- Regular ongoing development
- Content and monitoring as an ongoing process, not a one-off project.
Content and sources
Helpful, verifiable content is more important than artificial keyword repetition: clear company information can support machine identification, and understandable, original content can support eligibility for citation. External mentions and sources cannot be artificially guaranteed.
Technical accessibility alone does not guarantee a mention. Generative systems can misinterpret information — results must be checked regularly, and a single query is not a reliable measure of success.
Crawlers and technical accessibility
OpenAI distinguishes three types of access: OAI SearchBot serves ChatGPT's search functions — for a website's content to be eligible to appear in ChatGPT search answers, OAI SearchBot must not be blocked. GPTBot concerns the possible use of publicly accessible content to train generative models. ChatGPT User can access a page in response to a specific user action; this is not the same as automatic crawling. OAI SearchBot and GPTBot have separate settings: a website can allow OAI SearchBot and block GPTBot. Allowing OAI SearchBot guarantees neither inclusion nor citation.
The normal SEO fundamentals still apply to Google's AI features: pages must generally be indexable and eligible for a search result with a text snippet. Google does not require a special GEO file or special AI markup, and there is no special GEO schema that guarantees inclusion. Structured data must match visible content, important content should be available as readable text, internal links help discovery, and images and videos can usefully complement text. Data from AI features may appear in Search Console within Web Search. Inclusion, indexing and display are not guaranteed.
For Bing and Copilot, crawling and indexing remain the technical foundation: clear canonicals and redirects help establish unambiguous URLs, sitemaps should contain only canonical URLs, and IndexNow can notify Bing about new, updated or removed URLs. Bing Webmaster Tools can analyse visible citations in supported AI systems — citations are not rankings or guaranteed visitors. Perplexity uses web search and citations; PerplexityBot respects robots.txt directives, and allowing crawling does not guarantee a citation.
llms.txt is not a standard required by all generative search systems, and Google does not require an llms.txt file for its AI features. An llms.txt file must not be sold as a guaranteed GEO advantage; if used, its benefit for the specific project must be assessed. It does not replace robots.txt, sitemaps, canonicals, indexing or helpful content.
Potential benefits for your visibility in Zurich
- Clear positioning for complex decision - making questions
- Service pages that machines can classify unambiguously
- Comparisons and decision - making aids with honest criteria
- Verifiable expertise with sources and up - to - date information
- Consistent company information across all pages
- Beobachtung verschiedener Systeme statt einer Einzelabfrage
- A better foundation for traditional and generative search
- Prioritised development based on clarity and freshness
Lokale Informationen in Zurich
Widmer & Kunz is based in Basel and supports businesses in Zurich online: meetings, analyses and evaluations are conducted remotely, and in - person appointments are possible by arrangement. For your content, this means that we highlight how your services are geared towards Zurich without claiming locations that do not exist — clear, honest business information is also the more reliable approach for generative systems.
What GEO cannot honestly promise
Mentions, citations and positions in AI answers cannot be guaranteed — especially in a competitive environment. Different systems may use different sources, answers may vary depending on the model, time, location and question, and systems are constantly changing. Crawler access does not guarantee use, structured data does not guarantee a mention, and llms.txt does not guarantee an advantage. AI systems can make mistakes and misinterpret information. A single query is not a reliable measurement, not every citation generates a click, and external mentions cannot be artificially forced. GEO does not replace SEO, content requires regular maintenance, platforms offer different levels of control, and personal or confidential data does not belong in public content. Success must be monitored over a longer period and across multiple questions.
Monitoring and evaluation
We monitor selected decision - making questions over defined periods and for each system: is a source displayed, is one of your own pages mentioned, is measurable referral traffic generated? A citation is not a ranking or a guaranteed visitor — monitoring provides context, rather than grounds for claiming success.
Project process for your GEO work in Zurich
Distinguishing GEO from SEO
SEO optimises for traditional search engines and organic search results. GEO prepares information for potential use, summarisation or citation in generative search systems. GEO often builds on SEO fundamentals, but does not replace SEO: the same technical accessibility helps both areas, while GEO focuses more strongly on clear answers, verifiable facts, sources and automated association. Each area has its own pages, with links between them.
SEO Suchmaschinenoptimierung im DetailYour questions about GEO in Zurich
Was bedeutet GEO?
GEO stands for Generative Engine Optimization — in plain English: optimisation for generative search engines. It concerns the potential visibility of your content in AI - powered answers, for example to complex comparison questions.
Do you have an office in Zurich?
No. Widmer & Kunz is based in Basel. We work together digitally and arrange in - person meetings when needed.
Can a citation in Google's AI answers be guaranteed?
No. The usual SEO fundamentals apply to Google's AI features; there is no special GEO file, specific AI markup or special GEO schema that guarantees inclusion. Inclusion, indexing and display are not guaranteed.
What content is suitable for generative search?
Content that answers specific questions fully and verifiably: clear service descriptions, honest comparisons, traceable sources and clear signs that it is up to date.
What is the difference between OAI SearchBot and GPTBot?
OAI SearchBot serves ChatGPT's search features, while GPTBot serves the potential use of public content to train generative models. The two are controlled by separate settings.
Brauche ich eine llms.txt Datei?
llms.txt is not a required standard; Google does not require an llms.txt file for its AI features. It must not be sold as a guaranteed GEO advantage — we assess its usefulness on a project - by - project basis.
Which platforms are monitored?
ChatGPT search, Google's AI features, Microsoft Copilot and Perplexity — with the caveat that each system selects sources differently and no selection can be guaranteed.
What does a GEO analysis involve?
We examine your target audiences' questions, technical accessibility, company information and existing content, and use the findings to identify prioritised measures — without promising success.