Before we talk strategy, here is the raw data. Four snapshots from the actual audit that show where Panorama stands today. Read these, and the rest of the proposal becomes obvious.
What panoramaresort.ch actually talks about right now. The site covers Genuss, Wellness, Hotelübernachtung, Paare, Aussicht. Real topics, but the cluster structure is built for a brochure, not for AI retrieval. No anchor on “wellness weekend Switzerland”, no “spa resort near Zurich”, no comparison cluster against named competitors. AI retrieval engines map queries to topic clusters. This is what they see.
How the new topic map will be built. Core drives revenue (Wellness & Spa, Rooms & Suites, Culinary, Events, Long-term Living). Outer builds authority (Lake Zurich activities, Local attractions, Wedding planning, Sustainability, Tatar Festival).
The audit tells us which side is currently empty and where the quick wins are. The Core card is small today (Genuss, Wellness, Hotelübernachtung, Paare, Aussicht) and not structured around how AI retrieval engines map queries to topic clusters. The Outer card barely exists. That's the gap.
Note: this is a sample set - 20 representative titles across 6 clusters, used to demonstrate structure and the type of coverage we'd build. The full topical map (50+ query intents, hundreds of mapped titles, per-AI corpus study, German priority market) is created when we begin the engagement. Each article is engineered to be retrieved, parsed, and cited by AI, not just rank on Google.
Which is what SEO has always been about. The acronyms are new. The discipline is not.
In plain English: an entity's prominence is how loudly it shows up across the web when AI looks for an answer in a topic. The knowledge graph is the web of "this thing relates to that thing" AI uses to decide what's authoritative. AIO + GEO + AEO all push the same engine: more mentions, more associations, more trust signals - from more places AI is reading from.
Visibility across the full model landscape: ChatGPT, Gemini, Claude, Perplexity, DeepSeek, Meta, plus the open-source wave (Kimi, Minimax, Qwen, GLM). Map each model's semantic borders and produce the content each one needs to surface your brand.
Each AI has its own preferences shaped by training data - travel verticals, encyclopedic sources, news outlets, niche authorities. Strategies are nuanced per target: the corpus that influences ChatGPT is not the corpus that influences Qwen or Gemini. We tailor outreach and PR by model.
The search index each model actually pulls from: Claude→Brave, Grok→Bing, ChatGPT→Bing + proprietary search crawler, Qwen→Quark Search, Kimi→Moonshot search, OpenClaw / Hermes Agent → configurable (Tavily, SerperAPI, SERPAPI, Firecrawl). We optimize the questions those indexes get asked.