Claus Siebeneicher
DE

Your Website Is Invisible — AI Ate Your Traffic

Published 1 Mar 2026 · 5 min read

The search engine you optimised for no longer exists

For twenty years, website visibility meant one thing: rank on page one of Google. That model is over. Not declining — structurally over.

Google’s AI Overviews now appear in roughly half of all searches. In the US, the figure is over 60%. When they do appear, organic click-through rates drop by 61%. In Google’s AI Mode — launched mid-2025 — 93% of sessions end without a single click to an external website.

The user asks a question. The AI answers it. Your website provided the source material. The user never visits.

The numbers that matter

The shift accelerated faster than most predictions suggested:

  • Google AI Overviews appear in ~50% of searches globally, 60%+ in the US. They cite an average of 13 sources per response — but 40% of those sources rank below position 10 in traditional results
  • ChatGPT has 800 million weekly active users and processes over 2 billion queries daily. It reached $10 billion in annual recurring revenue
  • Perplexity AI serves 30-45 million monthly active users with 170 million monthly visits, processing over a billion queries per month
  • Zero-click searches run at 43% with AI Overviews present, up from 34% without — and 93% in Google’s AI Mode

Gartner predicted in 2024 that traditional search volume would drop 25% by 2026. The prediction was directionally correct but structurally wrong — search volume has not dropped much, but what happens after the search has changed completely. Users still search. They just do not click.

What AI search engines actually cite

This is where it gets interesting for anyone trying to be visible. AI Overviews do not simply cite the top-ranked page. They synthesise from multiple sources and strongly favour:

  • Reddit (cited in 21% of AI Overviews) and YouTube (18.8%) — user-generated, authentic content
  • Structured data and schema markup that AI can parse programmatically
  • Clear authorship with demonstrated expertise (Google’s E-E-A-T framework)
  • Direct, factual answers rather than keyword-optimised filler

40% of sources cited in AI Overviews rank below position 10 in traditional search. That means a well-structured page on a small site can outperform a large competitor’s SEO-optimised content — if the AI considers it authoritative.

GEO: optimising for the machine that answers

The industry now calls this Generative Engine Optimization (GEO). It is not a replacement for SEO — it is a parallel discipline that optimises content for AI systems that generate answers rather than rank links.

The core question shifts from “How do I rank higher?” to “How do I get cited when the AI answers?”

What works:

  1. Structured, machine-readable content. Schema markup, clear headings, factual statements that AI can extract and attribute. If your content cannot be parsed by a language model, it does not exist for AI search.

  2. Authoritative sourcing. AI systems weight content from sources that are cited by other sources. Third-party mentions, industry publications, and verifiable credentials matter more than backlink counts.

  3. Entity-based identity. AI models think in entities, not keywords. Your brand, your people, your expertise need to exist as structured entities that AI can recognise across contexts. This means consistent schema markup, clear author profiles, and presence across authoritative platforms.

  4. Conversational depth. AI search answers questions. Content that directly answers real questions — with nuance, not just surface-level summaries — is more likely to be cited. Long-tail, intent-driven content wins.

  5. Multi-platform presence. AI models train on and cite diverse sources. Being visible on LinkedIn, GitHub, industry forums, and in published articles creates citation surface area that a website alone cannot.

The antitrust wildcard

In December 2025, a US federal court imposed a six-year remedies framework on Google after finding it a monopolist in search. The ruling prevents Google from using exclusivity deals to lock in default status and limits default search agreements to one year. The judge explicitly noted that generative AI — particularly ChatGPT — constitutes a “disruptive force” that changes the competitive landscape.

The DOJ and 35 states are appealing, pushing for stronger remedies. The Supreme Court will likely decide by 2027-2028. Meanwhile, the ruling has already opened distribution channels that were previously locked to Google, creating opportunities for alternative AI search providers.

What this means for businesses

If your digital strategy still measures success by Google rankings alone, you are measuring the wrong thing. The metrics that matter now are:

  • AI citation rate — how often your content is cited in AI-generated answers
  • Share of voice across AI platforms — presence in ChatGPT, Perplexity, Gemini, not just Google
  • Entity recognition — whether AI models recognise your brand and expertise as authoritative
  • Direct traffic and brand search — the only metrics AI cannot intermediate

The organisations that will maintain visibility are those that treat AI search as a first-class channel — not an afterthought bolted onto an existing SEO strategy.

The parallel to what I do

I work with organisations where technology decisions have operational consequences. AI visibility is one of those decisions. Most businesses are still optimising for a search paradigm that is being replaced in real time. The ones that adapt early will not just maintain their traffic — they will capture share from competitors who are still chasing page-one rankings that increasingly lead nowhere.

The technology has changed. The strategy needs to change with it.

This article was originally published in April 2025 and has been substantially updated in March 2026 with current statistics and developments. The AI search landscape is moving fast — the numbers cited here reflect the state of play as of early 2026.