AI Overviews Cut Clicks by 58%: The Data Behind the Decline and What Actually Works Now

AI Overviews reduce your organic clicks by 58%. That is not a prediction. That is measured data from Ahrefs, confirmed across millions of search results between December 2023 and December 2025.

If you rank position one for an informational keyword and an AI Overview appears above your result, your click-through rate drops from 7.6% to 3.9%. You lose more than half your traffic overnight, and there is no SEO fix for it. The AI answer simply replaces your page in the user’s attention.

This article breaks down exactly what changed, why traditional SEO cannot solve it, and what you should do instead.

The 58% Number: Where It Comes From

Ahrefs published an updated study in early 2025 analyzing organic CTR changes since Google launched AI Overviews in May 2024. Their methodology compared position-1 click-through rates for informational keywords with and without AI Overviews present.

The findings:

The first row tells you something important. When no AI Overview appears, position-1 CTR barely changed. The decline is almost entirely caused by the AI answer itself, not by broader search trends.

A previous Ahrefs study from 2024 found a 34.5% CTR reduction. The new number is 58%. The damage nearly doubled as Google expanded AI Overviews to more queries and improved answer quality.

AI Overviews Now Trigger on 60%+ of Searches

Semrush published a separate study confirming that AI Overviews now appear on over 60% of Google searches in the United States. In January 2025, commercial and transactional queries triggered AIO only 8.7% of the time. By October 2025, that number hit 42.9%.

This means AI Overviews are no longer an informational-query problem. They are expanding into the queries where businesses make money: product comparisons, service recommendations, pricing searches, and buying guides.

Key findings from the Semrush data:

• Informational queries: AIO triggers on 74% of searches
• Commercial investigation: AIO triggers on 42.9% (up from 8.7%)
• Transactional queries: Growing rapidly, especially in SaaS, finance, and health
• Average AIO length: 361 words (longer than most featured snippets)

When an AI Overview writes 361 words answering the user’s question, there is very little reason to click anything.

Zero-Click Is No Longer a Trend. It Is the Baseline.

Separate data from Similarweb confirms the broader picture. Zero-click searches on Google grew from 56% to nearly 69% between May 2024 and May 2025. That means 7 out of 10 Google searches end without any external website receiving a visit.

Ars Technica covered this in a widely shared piece calling zero-click “the culmination of Google’s hubris.” The framing is accurate from a publisher perspective, but it understates the shift. This is not hubris. This is Google’s business model evolving. They keep users on their property, show ads inside AI answers, and monetize the attention directly.

Google confirmed at Google Marketing Live 2025 that ads are now rolling out inside AI Overviews on desktop in the US. They are testing ads in AI Mode as well. The playbook is clear: Google answers the question, Google shows an ad, the user never leaves.

Why SEO Cannot Fix This

Traditional SEO optimizes for rankings. You target a keyword, build authority, optimize your page, and climb to position one. That model assumes the user will click something on the results page.

AI Overviews break that assumption in three ways:

1. Position one is no longer position one. When an AI Overview appears, your position-1 result is pushed below the fold. Users see the AI answer first. Your blue link is secondary.

2. The AI answer synthesizes multiple sources. Even if you rank first, the AI Overview might cite your competitor instead. Google’s AI selects sources based on its own relevance signals, which are not the same as ranking signals.

3. Clicks are decoupled from visibility. Your brand can appear in an AI Overview without anyone clicking your link. You get exposure but zero traffic. That is the new paradox: visible but unclickable.

No amount of title tag optimization or internal linking fixes this. The problem is structural.

The Shift: From Rankings to Recommendations

If clicks are disappearing, what matters instead? Being the source the AI recommends.

Generative AI engines (ChatGPT, Perplexity, Gemini, Google AI Overviews) all work the same way at a high level. They synthesize information from their training data and web sources, then produce a single answer. They cite specific brands, products, and websites when those sources demonstrate clear authority.

This is fundamentally different from SEO. Instead of competing for a spot on a list of ten blue links, you are competing to be the single recommended answer. One brand wins. The other nine get nothing.

At searchless.ai, we call this Generative Engine Optimization (GEO). The goal is not to rank. The goal is to be cited.

5 GEO Strategies That Replace Lost Click Traffic

Based on our analysis of how AI engines select and cite sources, these five strategies produce measurable results.

1. Answer-First Content Structure

AI engines extract information from the first 1-2 sentences of a page 73% of the time, according to our internal citation analysis. If your answer is buried in paragraph four, the AI will miss it or paraphrase it poorly.

Structure every page and article with the direct answer in the first sentence. Then expand with supporting detail. This is the opposite of the traditional SEO advice to “build suspense” or “hook the reader.”

Example:

Bad: “In recent years, many businesses have wondered about the best CRM software. Let’s explore the options.” Good: “The best CRM for small businesses in 2026 is HubSpot Free CRM, based on ease of use, integrations, and zero-cost tier. Here is the full comparison.”

The AI engine reads the second version and has its answer immediately. You become the citation.

2. Entity Authority Building Across Multiple Domains

AI engines measure entity authority by counting brand mentions across independent domains. Our benchmark data shows that brands cited by ChatGPT are mentioned on an average of 6.2 unique domains. Brands that are invisible to AI average 1.4 domains.

This is not about backlinks for PageRank. This is about brand mentions, whether linked or not, across diverse and credible sources. Guest posts, PR coverage, podcast transcripts, Reddit threads, industry reports. The AI picks up on the pattern of independent corroboration.

Actionable steps:

• Get mentioned in 3+ industry publications per quarter
• Appear in comparison articles on review sites (G2, Capterra, independent blogs)
• Have your founders quoted in news articles and roundups
• Build presence on forums where your audience asks questions (Reddit, Quora, niche communities)

3. Implement llms.txt

llms.txt is a structured file that tells AI engines how to read your website. Think of it as robots.txt for the AI era. As we documented in our llms.txt implementation guide, sites with llms.txt see faster and more accurate AI citations because the engine can parse your content structure without guessing.

The format is simple. Place a llms.txt file in your root directory with structured sections for your company, products, and key content. Include direct links to your most important pages with brief descriptions.

Over 95% of websites do not have llms.txt. Implementing it gives you a structural advantage.

4. FAQ Schema and Structured Data

Google AI Overviews, ChatGPT, and Perplexity all read JSON-LD structured data. Your FAQ schema literally becomes the AI’s answer source. When you mark up your content with Question/Answer schema, you are handing the AI a pre-formatted citation.

For every key page on your site:

• Add FAQ schema with 5-10 questions and concise answers
• Use HowTo schema for process content
• Implement Product schema with detailed descriptions
• Add Organization schema with complete brand information

The AI does not have to infer or paraphrase. It lifts your structured answer directly.

5. Multi-Platform AI Visibility Monitoring

You cannot optimize what you do not measure. Track whether AI engines mention your brand across ChatGPT, Perplexity, Gemini, and Google AI Overviews. Our cross-platform AI visibility analysis found that 89% of brands are invisible on at least one major AI platform.

Set up regular checks:

• Ask ChatGPT about your product category. Are you mentioned?
• Search Perplexity for your key topics. Does your brand appear?
• Check Google AI Overviews for your target keywords. Is your content cited?
• Run a free AI visibility score at audit.searchless.ai to get a baseline

If the AI does not know you exist, no amount of content will fix it. Visibility monitoring tells you where you stand.

The Data Is Clear: Adapt or Lose Traffic

Let us recap the numbers:

• 58% CTR reduction when AI Overviews appear on your keyword (Ahrefs)
• 69% of Google searches now end without a click (Similarweb)
• 60%+ of US searches trigger AI Overviews (Semrush)
• 42.9% of commercial queries now show AI Overviews, up from 8.7% in January (Semrush)
• 89% of brands are invisible on at least one AI platform (Searchless.ai benchmark)

Every metric points in the same direction. Click traffic from traditional search is declining structurally, not cyclically. AI engines answer questions directly. Users click less. Google monetizes the attention internally.

The brands that will thrive in this environment are not the ones with the best keyword rankings. They are the ones that AI engines recommend by name. That requires a different strategy, different metrics, and a different set of optimizations.

GEO is that strategy. It focuses on being cited, not being clicked. On being recommended, not being ranked. On building the signals that make AI engines choose you as the answer.

If you have not checked your AI visibility yet, that is the single highest-leverage action you can take today. Find out whether the AI knows you exist, then build from there.

FAQ

Do AI Overviews affect all industries equally?

No. Informational queries are hit hardest (74% trigger rate), but commercial and transactional queries are catching up fast. SaaS, finance, health, and e-commerce see the fastest growth in AI Overview presence.

Can I opt out of AI Overviews?

Google does not offer a direct opt-out for AI Overviews in the same way you can block a featured snippet. You can use nosnippet meta tags, but this may also remove your standard snippet, reducing visibility further. The better approach is to optimize for being cited inside the AI Overview.

Is SEO completely dead?

No. SEO still matters for queries without AI Overviews, for local search, and for building the foundational authority that AI engines also consider. But SEO alone is no longer sufficient. You need GEO on top of it.

How quickly can I improve my AI visibility?

Most brands see measurable changes within 8-12 weeks of implementing answer-first content, entity authority building, and llms.txt. The brands in our benchmark that went from invisible to cited invested in consistent weekly publishing and backlink acquisition.

What is the difference between GEO and AEO?

GEO (Generative Engine Optimization) covers all AI engines: ChatGPT, Perplexity, Gemini, Google AI Overviews. AEO (Answer Engine Optimization) is a subset focused specifically on voice assistants and answer engines. They share tactics but differ in scope.

How do I measure AI visibility?

Track how often AI engines mention your brand when users ask about your product category or key topics. You can do this manually by testing prompts across platforms, or use a free tool like audit.searchless.ai to get an automated AI visibility score in 60 seconds.

Want to know if AI engines recommend you? Get your free AI Visibility Score in 60 seconds at audit.searchless.ai.