The problem
Open ChatGPT. Type "tax advisor in Zurich" or "dentist in Winterthur". Look closely at which names appear.
In almost every case, your company will not be among them. The BrightLocal study 2026 shows that 45% of surveyed consumers already use AI tools for local business recommendations. Concurrently, the SOCi Local Intelligence Report shows that only 1.2% of analysed business locations ever appear in AI-generated answers.
The trend is unmistakable: companies that do not provide structured data for AI systems will be overlooked in this rapidly growing channel.
Why Google SEO is not enough
Google presents a list of links: users click, browse, and compare. That mechanism has worked reliably for two decades.
AI systems operate fundamentally differently. ChatGPT, Claude, and Perplexity synthesize direct answers, naming specific businesses outright. Some include links and citations; others answer purely generatively. Users frequently act on the recommendation immediately or refine their query without ever visiting a traditional search engine.
Google rankings and AI recommendations are two distinct mechanisms. A website can rank number one on Google and still be ignored by ChatGPT. Conversely, a specialized SME without top Google rankings can appear in AI recommendations if its technical foundation is machine-readable.
What AI systems can evaluate
AI systems evaluate multiple layers: the visible copy on your website, structured data (JSON-LD schema markup), verified directory profiles, and third-party mentions. An llms.txt file is an effective way to deliver core corporate facts in a machine-readable format directly from your domain root.
Three core clusters of information drive accurate classification: identity (registered company name, legal form), capabilities (services, industry focus), and geographic scope (headquarters, service regions).
Most corporate websites provide this information only in unstructured body copy. Implementing structured data enables AI systems to extract, verify, and cite these facts with confidence.
Four action steps
- Publish an llms.txt file. Place a structured text file in the root directory of your website summarizing core company facts in an AI-friendly format. This emerging standard is rapidly gaining industry adoption.
- Implement JSON-LD schema markup. Deploy comprehensive Organization and Service schemas covering legal company name, physical location, industry focus, service portfolio, and contact details.
- Permit AI crawler access. Many default server configs block user agents like GPTBot and ClaudeBot in robots.txt. Ensure your robots.txt file explicitly permits indexing by authorized AI systems.
- Structure modular content entities. Define each service, regional hub, and sector specialization as its own semantically structured element rather than burying them in general paragraphs.
What it costs
A complete technical setup requires a one-time investment of CHF 2,850 to 7,350, depending on website complexity and the number of operating locations. There are no ongoing subscriptions and no monthly fees.
Everything is configured within 10 working days. When changes become visible depends on the indexing schedule of each AI model, as some refresh their training and search indices more frequently than others.
