What happens to my website
when Google only shows AI answers?

Since May 2026, Google has used Gemini 3.5 Flash as the default model in AI Mode — worldwide, for over a billion users, in 98 languages, with no subscription required (Google Blog, 05/19/2026). In Germany, AI Mode has been available as an opt-in since October 2025. It's not yet the default view — but the direction is clear. AI search isn't coming someday. It's already here.

What's concretely changing for Google Search in Germany

AI Overviews — the AI-generated summaries that appear above the organic results — already show up on around 20 percent of all German search queries. That's what Sistrix measured in the largest published DACH study to date, covering over 100 million German keywords (Johannes Beus, Sistrix, 02/25/2026).

The consequences are drastic: the click-through rate for position 1 drops from 27 percent to 11 percent when an AI Overview appears — a decline of 59 percent. Extrapolated, that means 265 million fewer organic clicks per month in Germany alone.

Industries with high information needs are especially affected: health, home services, finance. For trades businesses that generate emergency-service or appointment inquiries through organic search, this isn't a future scenario — it's happening now.

Core answer: AI Overviews appear on 20% of all German searches. Click-through rate for position 1 drops by 59% — 265 million fewer organic clicks per month in Germany.

SEO gets you into the top 10 — GEO gets you into the AI answer

If you think SEO is done for, you're wrong. 97 percent of all AI Overviews cite at least one source from the organic top 20. SEO remains the entry ticket. But it's no longer enough to just be on the results list — you need to be cited in the AI answer.

Trades example: If someone searches "How much does pipe restoration cost?", it's no longer enough to rank on page 1 with a blog post. The AI pulls a price range, an FAQ, and a provider with structured data directly into the answer. Anyone who doesn't show up there doesn't exist for the searcher — even if they rank organically at position 3. → How AI agents help tradespeople.

Financial advisor example: Anyone searching "How do I switch insurance brokers?" gets an AI answer with process steps, legal notes, and source links. A financial advisor who has structured their §34d expertise — with FAQ schema, Person schema, and sameAs linking to their LinkedIn profile — has a chance to be cited there as a trustworthy source.

Core answer: SEO remains the entry ticket. GEO determines whether you're cited in the AI answer — or invisible at position 3.

Which GEO methods work — and which ones backfire

Researchers at Princeton University measured, in a peer-reviewed study (KDD Conference 2024, Barcelona), for the first time which optimizations actually increase visibility in AI-generated answers (Aggarwal et al., 2024).

The three most effective methods — each delivering +30 to 40 percent visibility:

1. Cite sources — verifiable evidence directly in the text, with date and publisher.
2. Include quotes — direct statements and expert voices that AI systems treat as an authority signal.
3. Integrate statistics — concrete numbers with source and as-of date instead of vague phrasing.

What doesn't work: keyword stuffing, content padding, and pure persuasion without substance. Techniques that used to gain reach in classic SEO are ineffective to counterproductive in AI search.

For trades businesses, that means concretely: instead of repeating "emergency heating repair Munich" fifteen times on a page, structure your page with real answers to real customer questions — and back them up with numbers.

For financial advisors: instead of generic industry terms, prepare your §34d expertise as an FAQ with concrete process steps and compliance notes — that's exactly the information density AI systems prefer.

Core answer: Sources + statistics + quotes = up to +40% AI visibility. Keyword stuffing is ineffective to harmful in AI search.

Schema markup: the technical lever most people overlook

The most effective technical lever for GEO is schema markup. Combining FAQPage, Article, and HowTo in the schema graph achieves 1.8 times more LLM citations than Article schema alone (Walker Sands, 2026). Structured data raises the accuracy of AI answers from 16 to 54 percent (seoscore.tools, 2026).

What that means in practice:

• FAQPage schema with 3 to 8 real questions and answers per page — too few gives no signal, too many dilutes it.
• Person schema for the author, with sameAs links to LinkedIn and other profiles. Without this verification path, LLMs can't evaluate the E-E-A-T signals.
• Article schema with publication date, author, and organization.

The priority order: FAQPage > Article > Person > HowTo.

For a trades website, that means: mapping response times, emergency prioritization, GDPR notices, and pricing logic as FAQ schema — not burying them in body text.

For a financial advisor page: §34d compliance, the §3.1 opening disclosure, GDPR data-storage locations, RDG boundaries, and client confidentiality as structured FAQ — each one a potential AI citation building block.

Core answer: FAQPage + Article + Person schema = 1.8× more LLM citations. Priority: FAQPage → Article → Person → HowTo.

How amazing-agents.com makes its own website GEO-ready

We build for others what we use ourselves. Our web-deployment agent ran a complete SEO and LLM-visibility audit on amazing-agents.com in a single heartbeat — including internal linking, noindex cleanup, and a spec-compliant llms.txt. What classic agencies deliver in one to two weeks for €1,500 to €3,000, an agent finished in minutes.

Our Knowledge Hub articles already use Article schema. The next step: consistently implementing FAQPage and Person schema with sameAs links — so AI search systems don't just find our content, but cite it. And that's exactly what our content already is — our daily AI news hub, our guides, our industry pages — sources that AI systems can already draw on. We build websites that AI understands in search too.

Core answer: We build for others what we use ourselves — and demonstrate GEO live on amazing-agents.com with FAQPage + Person schema on every page.

What you should do now

1. Check your current AI visibility: Search your most important keywords in Google AI Mode and see whether your website shows up in the AI answer.
2. Structure your content for AI: Question-and-answer format, backed-up statistics, quotable sentences.
3. Implement schema markup: FAQPage + Article + Person as the minimum standard.
4. Measure the difference: Track whether your pages get cited in AI Overviews — and adjust iteratively.

SEO isn't going away. GEO is added on top. Whoever masters both will be visible in Google Search in 2026 — in the classic results list and in the AI answer.

Frequently asked questions about GEO and Google AI search

What's the difference between SEO and GEO?

SEO optimizes your website for the organic results list — rankings, keywords, backlinks. GEO optimizes for AI-generated answers — citability, source structure, schema markup. SEO remains the foundation; GEO builds on top of it.

Is SEO obsolete now?

No. 97 percent of all AI Overviews draw their sources from the organic top 20. Without solid SEO, you won't get within reach of the AI answer at all.

Which industries are most affected?

Industries with high information needs: health, home services, finance, legal. In the trades, click-through rates for emergency-service searches drop especially sharply, because the AI prefers to give a direct answer.

How fast does GEO take effect?

Schema markup can be indexed within a few weeks. Visibility in AI Overviews depends on the overall quality of the page — source density, factual structure, authority.

Can I implement GEO myself?

The content-side optimization — question headlines, statistics, cited sources — you can start right away. For schema markup you'll need technical support or a web-deployment agent like the one from amazing-agents.com.