Frequently asked questions

GEO is about how companies appear in the answers of AI assistants such as ChatGPT, Gemini or Perplexity – whether they are named, in what position and with which facts.

Essentially none: Generative Engine Optimization (GEO), Large Language Model Optimization (LLMO), Generative AI Optimization (GAIO), Answer Engine Optimization (AEO) and AI SEO are different names for the same discipline – visibility in AI answers. AIVISIBLE measures all of it.

Yes, AI SEO usually means the same as GEO: optimising so that AI assistants name a company and describe it correctly. It complements classic SEO rather than replacing it.

SEO remains important. But more and more questions are answered directly by AI systems that name a few specific companies instead of listing links. A good Google ranking does not guarantee a mention in those answers.

Twelve models from OpenAI, Anthropic, Google, xAI, Perplexity, Meta, Mistral AI, DeepSeek, Alibaba, Moonshot AI, Amazon and Cohere – including Perplexity Sonar, a system with live web search.

They differ in training data, architecture and whether and how they search the web live. That is why we test every model individually and repeatedly.

We query the models via their programming interfaces. The providers' apps sometimes also use web search, location and history, so individual answers may differ. The report discloses the methodology.

Less by keyword density than by context, machine readability for AI crawlers such as GPTBot or ClaudeBot, and mentions in trustworthy sources.

JSON-LD (Schema.org) describes your company in a structured, machine-readable way; the LLM file is a compact AI guidance file that summarises your company briefly and clearly. Both can help convey facts correctly – but an effect on individual models cannot be guaranteed.

Models sometimes repeat incorrect or outdated details – addresses, prices, services. The report checks the answers against your block of facts and lists discrepancies.

Systems with web search can pick up changes comparatively quickly; models without web search only with new training data. Monitoring makes changes visible over time.

Query languages: German, English, French, Italian, Spanish, Japanese, Russian, Chinese, Hindi and Korean. The report is in German or English. We test globally or for specific markets such as Germany, Switzerland, Austria, the United Kingdom and the USA.

We only process the information required for the audit. Payment is handled by Stripe; reports are provided via personal, time-limited download links.

No one can seriously guarantee that. But AI recommendations arise in an advisory context – being named there reaches prospects at a late stage of their decision.

Nothing immediate. But the more research runs through AI assistants, the more enquiries go to the providers named in the answers.

The Starter Audit costs €490 one-off, Monitoring €290 per month, Enterprise on request. All prices are exclusive of statutory VAT.

An audit across 12 AI models in one of ten languages, 20 customer questions with three runs each, a premium PDF report with action plan, drafts for an LLM file and structured data, and the report in the client portal.

Everything in the Starter Audit, plus two re-audits per month with trend comparison, a benchmark against your three main competitors and ongoing updates of the LLM file and structured data. Cancel monthly.

For groups of companies, multi-brand organisations and international presence: up to 50 questions, several brands, markets and competitors, a 90-day roadmap with implementation support and a personal results review.

Usually one to three working days after ordering and agreeing the profile. You receive the report by e-mail and in the client portal.

Securely via Stripe straight after ordering on the website. For Enterprise we get in touch personally.

A metric calculated from all neutral answers – overall, per model and per question category. It shows how often and how prominently your company is named.

Realistic customer questions in six categories: direct search, recommendation, comparison, local search, problem statement and trust. Neutral questions contain neither your name nor your domain.

AI answers vary from run to run. Several runs per question show how stable a mention is rather than capturing a one-off result.

Each answer is classified as top recommendation, mention, domain only or not mentioned – plus list position, cited sources, competitors named and a check for incorrect facts.

A reference containing your correct company details, such as services, locations and contact. The models' answers are checked against it and discrepancies are listed in the report.

Management summary, methodology, results, competitor analysis, sources and citations, brand perception with fact check, technical on-site audit, action plan, implementation package and an appendix with results per question.

Crawler rules, access for AI crawlers, sitemap, structured data and the LLM file of your website.

An instant technical check of your website: AI crawler access, LLM file, sitemap, HTTPS, structured data, Open Graph, language declaration and headings. A short automated check may also query a single AI model. We record the domain, e-mail, optional name, result, time of consent and a hash of the IP address.

No. Website, market and language are enough; optionally you can add services, competitors and city. Your web team can then adopt the drafts in the report.

Yes. When ordering you can name your main competitors. The report also shows which providers the models name of their own accord.

In your personal client portal, together with the status of ongoing audits. You also receive each report by e-mail.

No. The models are operated by third parties and are not influenced by us. We measure the current position and show which measures make sense.

No. AIVISIBLE is neither affiliated with the providers of the tested models nor commissioned or endorsed by them. Brand names serve only to identify the models.

The offer is aimed at businesses, for example SMEs, B2B providers, law firms, consultancies, agencies, local service providers and brands with international markets.

AIVISIBLE is a brand of Harmecs Lab based in Altendorf, Switzerland. Contact: aivisible@mail.ch.

A program such as ChatGPT, Claude, Gemini or Perplexity that answers questions in full sentences. It is powered by a large language model (LLM).