Visible in ChatGPT & Co.: How AI Finds Your Brand
By Dennis Weidner · 23 July 2026
"Which agency in Berlin can help us with AI automation?" A growing share of your potential customers no longer asks that question on Google, but in ChatGPT, Perplexity, Claude or Gemini. The answer is not ten blue links but two or three concrete recommendations. Either your brand is one of them, or the conversation happens without you. This article explains how AI assistants choose their sources and what you can concretely do to appear there.
Why search behaviour is shifting right now
Classic search was a list of links from which the user assembled the answer themselves. AI assistants turn that around: they read the sources and deliver an answer directly, often with only a few references. For users this is more convenient; for brands it is a turning point. "Ranking on page one" becomes "appearing in the answer". Since 2024, studies and platform figures have pointed in the same direction: the share of research that begins in assistants is growing quickly, especially for complex, advice-heavy B2B decisions.
How AI systems decide whom to recommend
AI assistants combine two sources of knowledge: the model's training knowledge and a live web search whose hits the model reads and summarises. For both, the same holds true: these systems prefer sources they can understand with confidence and consider trustworthy. In practice, that means:
- Clear facts beat marketing prose. "Weidner & Friends builds AI agents for mid-sized businesses, based in Berlin" is machine-readable. "We reimagine the future" is not.
- Consistency counts. Name, services and location should match across your website, LinkedIn, directories and articles. Contradictions lower the models' trust.
- Citable content wins. Concrete guides, figures, definitions and honest comparisons are readily picked up and linked by assistants.
- Technical accessibility is a basic requirement. Anyone who blocks AI crawlers in robots.txt or serves content only behind scripts simply does not exist for these systems.
GEO: the new discipline alongside SEO
For this work the term Generative Engine Optimization (GEO) has become established, sometimes also called LLM-SEO or AI-SEO. The good news: GEO does not replace SEO, it builds on it. A technically clean, fast website with structured data remains the foundation. To that come a few new building blocks:
- Structured data (Schema.org): mark up organisation, offerings, articles and authors as JSON-LD so machines recognise relationships with confidence.
- llms.txt: a compact, machine-readable self-description of your website for AI systems, analogous to robots.txt.
- Fact pages instead of platitudes: every important service needs a page that answers in two sentences: what is it, for whom, roughly what does it cost, why you?
- Build authority: expert articles with real experience, named authors and cited sources. E-E-A-T (Experience, Expertise, Authoritativeness, Trust) matters even more in AI answers than at Google.
- Mentions beyond your own website: industry directories, press, podcasts, guest articles. Models weigh how often and in what context a brand appears across the web.
The central GEO question is this: if an AI were allowed to say only three sentences about your company, are those three sentences stated clearly somewhere on your website?
What you can start with tomorrow
Three steps with the best ratio of effort to impact:
- Self-test: ask ChatGPT, Perplexity and Gemini what they know about your company and whom they recommend in your category. The result is your baseline, often sobering and always instructive.
- Website check: review robots.txt (are GPTBot, ClaudeBot, PerplexityBot allowed?), structured data and whether your core services are phrased as clear facts.
- One citable piece per month: a well-founded expert article that genuinely answers a concrete question from your audience beats ten generic news posts.
Conclusion
Visibility in AI assistants is not magic but consistent craft: technically accessible content, clear facts, real expertise and consistency across all channels. Whoever starts now gains a lead that is hard to close, because models forget slowly and trust grows over time. If you want to know how your brand stands in AI answers today and what to do about it, take a look at our offering for websites and digital platforms or talk to us directly.
Dennis has been building digital companies for more than 15 years – from performance marketing and media to AI agents. With Weidner & Friends he brings agentic systems to mid-sized businesses: autonomous where it helps, with the human deciding. As the operating unit of the Weidner Ventures Group, he unites strategy, technology and delivery under one roof.