How AI Overviews Are Cutting Google CTR by 61% (Recovery Plan 2026)

Google AI Overviews are collapsing organic click-through rates by 34.5% to 61% for top-ranking pages, even when those rankings hold perfectly steady. This is the defining search crisis of 2026: your keyword positions look fine in the dashboard, but the traffic is quietly disappearing.

This matters right now because the damage is accelerating. Early Ahrefs data from March 2024 showed a 34.5% CTR drop for position-one results when an AI Overview appeared. By December 2025, that figure had climbed to 58%. The longer you wait, the more compounding ground your competitors gain.

In this post, you'll get the full picture: what the CTR impact studies actually show, a before-and-after breakdown by query type, the five implementation mistakes costing brands the most visibility, and a concrete four-step framework for protecting and growing your click share in an AI-first search environment.

Quick Answer: How AI Overviews Are Cutting Google CTR

1. Audit your AI visibility gap — query top 20–30 prompts in ChatGPT, Perplexity, Gemini, Claude; cross-reference with GSC for impression-up/click-down pages
2. Deploy AI-native infrastructure — llms.txt, JSON-LD schema (Organization, FAQPage, Product), unblock AI crawlers
3. Reformat content for AI extraction — direct answer in first 100 words, semantic H2/H3 as questions, HTML tables (not images), proprietary data
4. Close the feedback loop — track AI referrals in GA4, identify which content earns citations, retroactively apply patterns to under-performing pages

Key Takeaways

• Seer Interactive's longitudinal study of 25.1 million impressions found organic CTR dropped 61% on queries triggering AI Overviews (from 1.76% to 0.61%), while paid CTR fell 68%.
• Brands cited directly inside AI Overviews achieved 35% higher organic CTR and 91% higher paid CTR than brands that ranked organically but were not cited, according to the same Seer Interactive research.
• Amsive's analysis found non-branded queries saw a 19.98% CTR decline on average, worsening to -37% when an AI Overview appeared alongside a Featured Snippet.
• Gartner predicts traditional search engine volume will drop 25% by 2026 as queries shift to AI chatbots and virtual agents.
• AI-referred traffic converts 4.4x better than standard organic search, meaning citation capture is now a higher-value KPI than raw click volume for most B2B and SaaS brands.
• BrightEdge data shows total clicks fell 30% while impressions rose 49% over the same period, confirming that GSC impression data now actively masks traffic losses.

What the CTR Data Actually Shows (And Why Your Dashboard Is Lying to You)

Here is what the major independent studies found when they isolated the AI Overview effect specifically:

The Seer Interactive Study: 61% Organic CTR Collapse

Seer Interactive analyzed 3,119 informational queries across 42 organizations, covering 25.1 million organic impressions and 1.1 million paid impressions between June 2024 and September 2025. The findings are the most comprehensive available:

• Organic CTR: Dropped from 1.76% to 0.61%, a 61% decline, on queries where an AI Overview appeared.
• Paid CTR: Crashed from 19.7% to 6.34%, a 68% drop on the same queries.
• The citation advantage: Brands that were cited directly inside the AI Overview achieved 0.70% organic CTR versus 0.52% for brands that ranked organically but were not cited. That is a 35% difference in clicks from the same SERP position.

The implication is stark: being in position one is no longer enough. Being cited inside the AI summary is the new position one.

The Ahrefs Longitudinal Data: Getting Worse Over Time

Ahrefs ran two studies using aggregated GSC data across 300,000 informational keywords. Their first study (March 2024 to March 2025) found a 34.5% CTR decline for the top-ranking result when an AI Overview was present. Position-one CTR dropped from 7.3% to 2.6%.

The Amsive Breakdown: Branded vs. Non-Branded Divergence

Not all queries are hit equally. Amsive's research surfaced a critical distinction that most marketers are missing:

Non-branded, top-of-funnel queries are taking the worst hit. When an AI Overview appears alongside a Featured Snippet, the entire above-the-fold SERP is consumed by zero-click answers. Branded queries, interestingly, see a CTR increase, because high-intent buyers use the AI summary to validate a vendor they already know before clicking through to book a demo or buy.

The strategic implication: your brand needs to appear in AI answers on non-branded category queries, where the damage is worst and where evaluation decisions are actually forming.

Before and After CTR by Query Type

This table aggregates data across the Seer Interactive, Ahrefs, and Amsive studies to show the full picture of how AI Overviews are reshaping CTR by intent type:

Pew Research Center data from July 2025 provides the user-behavior explanation. When an AI summary appeared, users clicked a traditional organic result in only 8% of searches, compared to 15% when no summary was present. And only 1% of users clicked the citation links inside the AI summary itself, preferring to read the synthesis and stop searching entirely.

Why This Is Happening: The Root Causes

Google's AI Overviews run on a Retrieval-Augmented Generation (RAG) pipeline. The system fetches relevant documents from the index, synthesizes a conversational answer, and presents it at the top of the SERP. The user gets their answer without visiting any site. For informational and top-of-funnel queries, this is catastrophically effective at eliminating clicks.

The zero-click trend predates AI Overviews but has been massively accelerated by them. Approximately 58.5% to 60% of all Google searches now end without a click to a third-party website. On mobile, that figure reaches 77%. Gartner predicts traditional search engine volume will drop 25% by 2026 as query share shifts to AI chatbots and virtual agents.

The brands that are not being hurt are the ones cited inside the AI Overview itself. That citation advantage is not accidental: it comes from deploying specific content formats and technical infrastructure that AI models can extract and reference. Without that, you get ranked and invisible simultaneously.

How to Protect Your Click Share: A 4-Step Framework

The sequence below matters. You cannot execute Step 3 effectively without the infrastructure from Step 2, and Step 4 only generates actionable signal once Step 3 has produced content that can be tested. Follow the order.

Step 1: Audit Your AI Visibility Gap Before Touching Anything Else

Before you optimize a single page, you need to know exactly where you are cited and where you are not.

Run your top 20 to 30 target queries through ChatGPT, Perplexity, Gemini, and Google AI Overviews. Document which competitors are cited, what format the AI prefers (tables, numbered lists, paragraph answers), and where your brand is absent. Cross-reference this with your GSC data to identify pages where impressions are rising while clicks are falling, which is the signature of an AI Overview absorbing your traffic.

This audit becomes your prompt map: the specific conversational questions your buyers are feeding AI during active vendor evaluation. You are not targeting "what is payroll software." You are targeting "what is the best payroll platform for a 50-person remote team with contractors in multiple countries."

Step 2: Deploy the AI-Native Technical Infrastructure

Most brands have a complete blind spot here. AI agents — GPTBot, ClaudeBot, Claude-SearchBot, PerplexityBot, Google-Extended — encounter your website built for humans: marketing language, JavaScript-rendered content, image-based data. They struggle to extract a clean understanding of what you do.

• Location: domain root, Markdown format
• Function: AI-specific sitemap directing models to your most important content
• llms-full.txt: complete text extraction for deeper parsing

Go beyond basic website schema. Deploy nested:

• Organization schema with sameAs linking to social profiles + Google Knowledge Graph
• FAQPage, HowTo, Product for content-type signals
• Offer for pricing in machine-readable format

Step 3: Reformat Existing Content for AI Extraction

Once your infrastructure allows crawlers in, you need something worth extracting.

AI models rely on RAG to synthesize answers. They look for content that is modular, direct, and structured. The content patterns that earn citations:

• Lead with the direct answer. Place a concise, definitive response to the core query within the first 100 to 200 words. No preamble, no marketing language. AI models reward bottom-line-up-front structure.
• Use semantic HTML heading hierarchy. H2 and H3 tags phrased as natural language questions. Do not use bolded text to simulate headings. The structural signal has to be in the markup.
• Use HTML tables for comparison data. If your comparison data sits in a JPEG or a Canva graphic, it is invisible to AI crawlers. It needs to be in a <table> element.
• Inject information gain. Proprietary statistics, original case study data, named expert quotes with clear author bios. AI systems have a strong preference for content that contains factual claims they cannot easily source elsewhere.

"Google rewards demonstrated expertise across its E-E-A-T framework," says the Search Engine Journal analysis of AI Overview ranking patterns, noting that content structured for human readability and factual extractability performs best in both traditional and generative rankings.

Step 4: Build the Closed-Loop Feedback System

Once Step 3 is in place, track which content earns citations and use that signal to iterate.

1. GSC pages with AI Overview signature — impression spike + click drop on the same query
2. GA4 referral traffic from chatgpt.com, perplexity.ai, gemini.google.com, claude.ai

When a post starts earning AI referrals, analyze its structure:

• Heading patterns
• Answer depth
• Data format
• Entity density

Apply those patterns retroactively to posts that are visible but not earning citations.

• Step 2 (infrastructure) → prerequisite for Step 3, because beautifully formatted content AI crawlers can't parse is wasted effort
• Step 3 (content) → prerequisite for Step 4, because you need a body of work tested in production before signal accumulates
• Skipping straight to iteration is one of the most common reasons GEO programs stall

When DIY Fails: The Execution Gap Is Real

Here's the honest version of what this 4-step framework requires to execute internally.

1. A GEO strategist who understands how LLMs select sources and can build prompt maps from sales call data
2. Engineering bandwidth to deploy llms.txt, audit crawler access, implement nested JSON-LD schema
3. Continuous content capacity — GEO is a living system that decays without maintenance, not a one-time batch
4. Data analyst to connect GSC + GA4 + AI referral tracking into a feedback loop that actually informs editorial decisions

A company buys a monitoring tool (Profound, Evertune, Scrunch) → gets a detailed report on AI visibility gaps → stalls.

• The dashboard becomes an expensive report nobody acts on
• Hiring someone qualified takes 3–6 months
• Content teams are already overcommitted
• Engineers have a 6-month backlog
• Nothing moves

The Managed Path: How a Done-for-You Service Handles This

Mersel AI's two-layer architecture

• 100+ high-intent pages + 20 backlinks delivered over 6 months — built from buyers' actual prompts (not keyword research assumptions)
• Prompt maps from sales call recordings + competitor citation patterns + existing AI answer landscape
• Publish-ready, delivered directly to your CMS at continuous cadence
• Connected to GSC + GA4 + AI referral data → posts get smarter over time, not decay
• 20 backlinks specifically targeting authoritative third-party sources AI engines cite (review sites, industry publications, niche directories)

Currently the one piece of the GEO stack that no other managed service runs in production:

• Clean entity definitions
• Extractable product descriptions
• Nested JSON-LD Schema markup
• llms.txt configuration
• Internal linking that maps relationships AI systems need

Real client outcomes

These aren't traffic spikes — they're compounding citation footprints that get harder for competitors to displace every passing week.

Pricing & honest limitation

• Pricing: From $1,800/month for managed execution
• Limitation: Done-for-you service, not a self-serve dashboard. Teams needing real-time prompt monitoring with direct UI access will find Profound or AthenaHQ better fits.

FAQ

Why is my Google Search Console showing more impressions but fewer clicks?

Does being cited in an AI Overview actually increase clicks, or just brand visibility?

• 35% higher organic CTR (0.70% vs 0.52% for non-cited at same rank)
• 91% higher paid CTR

Citation inside the AI summary is a measurable click driver, not just a brand signal — even within a zero-click environment.

Which queries are most affected by AI Overview CTR drops?

• Non-branded queries: -19.98% avg CTR when AI Overview appears
• Worsens to -37% when AI Overview + Featured Snippet appear together
• Branded queries: +18.68% CTR (high-intent buyers use AI summary to validate before converting)

What is llms.txt and do I need it?

Not technically mandatory, but a meaningful differentiator:

• AI crawlers (GPTBot, ClaudeBot, PerplexityBot) struggle with JavaScript-rendered pages
• llms.txt removes that friction
• Adoption is only ~10% of domains (per Ahrefs) — implementing now is a real differentiation signal

How long until I see measurable improvements in AI citation rates?

Timelines for teams deploying content + infrastructure simultaneously:

• Initial visibility lifts: 2–8 weeks
• Meaningful pipeline impact (qualified inbound + AI-attributed demos): 60–90 days
• Compounding effect: kicks in months 3+ as feedback loop accumulates signal

Sources

1. Search Engine Land: Google AI Overviews drive drop in organic, paid CTR
2. Ahrefs: AI Overviews Reduce Clicks — Updated Study (Feb 2026)
3. Seer Interactive: AIO Impact on Google CTR — September 2025 Update
4. Dataslayer / Seer Interactive Study: Google AI Overviews — The End of Traditional CTR
5. Amsive: Google AI Overviews — New Research Reveals CTR Drop
6. Search Engine Land: Google AI Overviews hurt click-through rates (Amsive)
7. Search Engine Land: Google AI Overviews hurting clicks — Pew Research study
8. Gartner: Search Engine Volume Will Drop 25% by 2026
9. Ahrefs: AI Overviews Reduce Clicks — Original Study (April 2025)
10. Semrush: Zero-Click Searches and AI Overviews
11. Yotpo: What is llms.txt?
12. Analyt Solutions: Schema Markup and LLMs.txt
13. Ahrefs: Answer Engine Optimization
14. Search Engine Land: AI Overview citations, clicks, what to do
15. Search Engine Journal: Studies Suggest How to Rank on Google's AI Overviews
16. The HOTH: Generative Engine Optimization Guide
17. Recomaze: AI SEO Mistakes That Kill AI Search Visibility
18. SE Ranking: Review Platforms in AI Overviews

Related Reading

• Why Organic Traffic Is Declining in 2026
• How Much B2B Organic Traffic Are AI Overviews Taking?
• AEO vs. SEO vs. GEO: Which Strategy to Prioritize in 2026