How do you do SEO in the age of AI?

What is AI SEO – and what changes now?

AI SEO is the evolution of search optimization for a search that increasingly returns answers, not just links. Google's AI Overviews show AI-generated summaries above the classic results, and many people now research directly in ChatGPT, Perplexity or Gemini. The goal shifts: ranking first among blue links is no longer enough – your content has to be the source a machine cites.

Crucially, GEO does not replace classic SEO – it builds on it. AI systems and AI Overviews draw from the same web that Google can crawl well and judge trustworthy. Publish in a technically clean, topically deep and credible way, and you win both – rankings and mentions in AI answers.

• SEO goal: a position among the blue links – GEO goal: being named and linked as a source inside the AI answer.
• SEO metric: rankings and clicks – GEO: mentions in AI answers and referral traffic from Perplexity or ChatGPT.
• SEO levers: keywords, backlinks, technique – GEO adds: answer-first, clearly named entities, quotable facts.
• Shared basis: crawlable, trustworthy, well-structured content a machine can reliably understand.

Technical SEO as the foundation

Technical SEO is not becoming less important but more so: a machine must reliably crawl, render and extract the facts on your page. Without that foundation the best copy is wasted – the AI simply won't find it or will misread it. Work through these points in order.

• Crawlability and indexing: clean robots.txt, up-to-date XML sitemap, fast server response, no accidental noindex tags.
• Core Web Vitals and mobile-first: fast load times, visual stability, flawless rendering on smartphones.
• Structured data with schema.org: Article, FAQPage, Organization, Person and Breadcrumb help machines extract entities and facts cleanly.
• Govern AI crawlers deliberately: GPTBot, ClaudeBot, Google-Extended and PerplexityBot can be allowed or blocked in robots.txt – a conscious choice about whether your content may feed AI answers.
• Semantic HTML and a clean heading hierarchy (H1–H3), so passages remain individually understandable and extractable.
• HTTPS, readable URLs and hreflang for multilingual sites – especially relevant in Switzerland with German, French, Italian and English.

Content for people and AI: answer-first and depth

AI systems extract individual passages from a page. Each section must therefore stand on its own and be quotable: the clear answer first, then the depth. Text that answers a question precisely and then explains it soundly convinces readers and machines alike.

• Answer-first: a precise 40–80-word answer right under the heading that an AI can lift verbatim.
• One question, one clear statement per section – with consistent terminology throughout.
• Topic clusters instead of isolated pages: a pillar page plus deep spoke articles cover a topic fully.
• Present facts in structure: tables, lists and clear steps are easy for machines to extract.
• Signal freshness: a visible date and genuine updates – AI systems favour fresh, consistent sources.
• Natural language over keyword stuffing: write for the reader, but name entities explicitly.

E-E-A-T: make trust provable

E-E-A-T stands for Experience, Expertise, Authoritativeness and Trust – signals Google and increasingly AI models use to weigh a source's reliability. Especially for health, legal or financial topics (YMYL), trustworthiness decides whether you get cited. Make these signals visible instead of assuming them.

• Make authorship visible: named authors with real expertise, Person schema and a meaningful short bio.
• Show experience: your own examples, hands-on knowledge and originality instead of reworded third-party content.
• Build authority: consistent mention as an entity, references and backlinks from topically relevant, credible sources.
• Secure trust: a complete imprint, real contact details, transparent sources, HTTPS and no misleading claims.
• Swiss perspective: revDSG-compliant data handling and transparent privacy notices build trust – and are themselves an E-E-A-T signal.

Internal linking and entity architecture

Internal linking distributes authority across your site and shows machines how your topics relate. Consistently named and linked entities form a semantic web that AI systems read like a small knowledge graph – and that co-decides whether your site counts as a competent source on a topic.

• Link pillar pages and spoke articles both ways so the cluster is recognisable as a unit.
• Use meaningful, descriptive anchor text – not "click here" but the target topic named.
• Flat structure: important pages reachable in a few clicks, no orphan pages without internal links.
• Name entities consistently: the same terms for the same concepts, so the semantic web stays coherent.

Step by step: your AI SEO playbook for SMEs

Here is how Swiss SMEs put AI SEO into a workable order. Start with the topic, secure the technical layer, then build content and trust – consistently in four languages where your market demands it.

• 1. Topic audit: what real questions do your customers ask? Derive topic clusters with a pillar and spokes.
• 2. Technical check: crawlability, load time, schema.org and your AI-crawler policy set in robots.txt.
• 3. Build one answer-first page per question with real depth – clear answer on top, sound explanation below.
• 4. Add E-E-A-T signals: named authors, transparent sources, a complete imprint.
• 5. Link internally: weave the cluster, descriptive anchors, no orphan pages.
• 6. Publish multilingually (DE/FR/IT/EN) with hreflang and true localisation – often decisive in the Swiss context.
• 7. Measure: track classic rankings plus AI visibility – ask your core questions directly in ChatGPT, Perplexity and Google, and watch whether and how your brand is named.
• If you would rather not build this architecture yourself, Weissmann AI helps Swiss SMEs set up the technical layer, content structure and multilingual publishing for AI SEO.