AI Overviews, Zero-Click Search, and Why Your Traffic Looks Broken

Reading time: 18 minutes

You’re staring at Search Console.

Impressions are up. Maybe way up.

Clicks are down. Or flat. Or weirdly volatile.

Your rankings haven’t tanked. You didn’t get a penalty. You didn’t break anything obvious.

But the traffic that used to come from Google is just… not there anymore.

You’re not crazy. The rules changed.

Between mid-2024 and late 2025, the fundamental economics of search broke. What worked for 20 years stopped working. The reason has a name: AI Overviews.

This guide explains what happened, what the data actually shows, and what small businesses and agencies need to do about it.

---

TLDR: The Great Decoupling

• AI Overviews launched universally in 2024 and now appear on roughly 13% of all Google searches, up from around 6.5% in mid-2024.
• Multiple independent studies show organic CTR dropping by roughly 60% when an AI Overview appears, and paid CTR dropping even more.
• Zero-click search is now normal, not a bug. Around 60% of searches end without a click to any external website.
• Informational queries take the biggest hit. Transactional and navigational queries are more resilient.
• Local “near me” queries still overwhelmingly show Map Packs and rarely trigger long AI summaries.
• The new success metric is Share of Voice in AI answers, not just keyword rank.
• Surmado Scan, Signal, and Solutions help you measure and respond to this shift without guessing.

---

What Changed: From Referral Engine to Answer Engine

For two decades, Google was a switchboard.

You searched. Google gave you ten blue links. You clicked. The publisher got traffic. Google got your attention and ad revenue. Everyone had a deal.

That model is over.

Now, Google is a destination. The search engine results page (SERP) is where the answer lives.

The New SERP: AI Overviews at the Top

When you search for something informational, Google often shows an AI Overview at the top of the page.

The AI Overview is a multi-paragraph summary. It synthesizes information from multiple sources. It answers your question directly.

On mobile, it fills the entire first viewport. You have to scroll past it to see the first traditional organic result.

This creates what researchers call a “scroll tax.” Users must actively choose to bypass a satisfactory answer to find a link.

Most don’t.

Zero-Click Search: The New Normal

A zero-click search is a session that begins and ends on the SERP.

User asks a question. AI summary appears. User gets what they need. User leaves.

No click to any external website. No traffic for the publisher. No visit to your business site.

This is not a bug. This is the feature.

Multiple studies now show that around 60% of all Google searches end without a click. For low-stakes informational queries, the percentage is even higher.

If the AI Overview tells you the symptoms of the flu or how to fix a leaky faucet, you don’t need to click anything. Your curiosity is satisfied.

That traffic is gone.

---

The Great Decoupling: Impressions Up, Clicks Down

The pattern is consistent across thousands of sites.

Search Console shows impressions climbing. Sometimes doubling. Your site is “showing up” more.

But clicks are flat or falling.

What’s happening: Google is matching your site to more queries because of AI Overviews and carousel features. Your content is being indexed and extracted. But users aren’t clicking because the AI summary already gave them the answer.

This is the “Great Decoupling” of visibility from traffic.

What the 2024-2025 Studies Actually Show

In September 2025, one of the largest ongoing studies of AI Overview impact published updated findings.

The data tracked millions of queries across thousands of keywords.

Key findings:

• Organic CTR for queries with AI Overviews: dropped from around 1.76% in June 2024 to roughly 0.61% in September 2025. That’s approximately a 61% decline.
• Paid CTR for those same queries: dropped from around 19.7% to roughly 6.34%. That’s approximately a 68% decline.
• AI Overview prevalence more than doubled from around 6.5% of queries to over 13%.

Another large analytics firm analyzed nine industries and found that while total search impressions increased by over 49%, click-throughs declined by nearly 30% since May 2024.

Translation: Google is showing more results. Fewer people are clicking anything.

The Difference Between Query Types

Not all queries are affected equally.

Informational queries take the hardest hit. These are “how to,” “what is,” “symptoms of,” and “best practices for” searches.

AI Overviews excel at these. They can summarize step-by-step guides, list symptoms, and explain concepts without the user needing to visit a site.

Transactional queries are more resilient. “Buy running shoes,” “book hotel in Austin,” and “price of iPhone 15” still generate clicks because the user needs to complete a purchase.

Navigational queries are largely unaffected. If someone searches for “Facebook login” or “Surmado.com,” they’re going to click through no matter what.

For small businesses, the implication is clear: top-of-funnel informational content is losing its traffic. Bottom-of-funnel transactional content is holding steady.

Publishers vs. SMBs: Who Gets Hit Hardest?

Large publishers and generalist content farms are seeing the most brutal declines.

Studies show traffic drops ranging from 25% to 80% depending on the vertical. Sites that specialized in definitions, how-to guides, and aggregated lists are being replaced by AI summaries.

For small and medium-sized businesses, the pain is different.

You’re not losing 80% of your traffic overnight. But you are losing the discovery layer.

Queries like “best HVAC company in Dallas” or “most reliable movers near me” used to send you top-of-funnel traffic. Now, those queries increasingly trigger AI Overviews that synthesize reviews and business details into a summary.

The user gets recommendations without ever clicking your site.

Your website still matters for conversion. But the AI summary is now handling discovery.

---

Local Reality: “Near Me” Fortress vs “Best Of” Battlefield

If you’re a local business, the AI era presents a split reality.

Some queries are still safe. Others are being absorbed into the AI machine.

The “Near Me” Fortress

Good news: proximity-heavy queries are highly resilient.

Searches like “coffee shop near me,” “plumber near me,” and “dentist open now” still overwhelmingly show Map Packs and Google Business Profile results.

Early studies analyzing millions of keywords found that queries containing “near me” triggered Map Packs nearly 100% of the time and rarely triggered long AI Overviews.

Why this matters: classic Local SEO fundamentals still work.

• Complete your Google Business Profile 100%.
• Keep your name, address, and phone number consistent across directories.
• Upload recent photos.
• Respond to reviews.
• Add your hours, services, and attributes.

Proximity is still the dominant ranking factor for immediate-need searches.

The “Best Of” Battlefield

The trouble starts with discovery queries.

“Best Italian restaurant for a date.” “Plumber good with emergencies.” “Movers who handle pianos.”

These queries are more likely to trigger AI Overviews.

The AI reads reviews. It extracts descriptors. It pulls out phrases like “good for kids,” “fast response time,” “cozy date night spot,” and “transparent pricing.”

It assembles a narrative summary. The user gets three recommendations with reasons. No clicks required.

For local businesses, this creates a new challenge: you need to be cited inside the AI summary, not just ranked in the organic results below it.

Clarity Signals: The New Local SEO

To appear in AI summaries, you need what researchers are calling “Clarity Signals.”

A Clarity Signal is a fact that is unambiguous, verifiable, and consistent across multiple sources.

How AI systems verify information:

AI models work by triangulation. If Yelp, Apple Maps, and your website all say your hours are “Mon-Fri 8am-6pm,” the AI assigns high confidence to that fact.

If those sources disagree, the AI either picks the most authoritative source or excludes you entirely to avoid hallucinations.

What this means in practice:

1. Cross-directory consistency. Your NAP (name, address, phone), hours, services, and descriptions must be identical across Google Business Profile, Yelp, Facebook, Apple Maps, and your website.

2. Detailed review content. AI systems read the text of reviews, not just star ratings. They look for tokens like “same-day service,” “veteran-owned,” “transparent pricing,” and “emergency availability.” A steady stream of recent, descriptive reviews provides the raw material for AI summaries.

3. Structured data as a language. Schema markup (LocalBusiness, Service, FAQPage) is the native format AI systems prefer. If you add clean schema to your website, you’re handing the AI a structured resume of your business.

Google Business Profile as the New Homepage

For many local businesses, the Google Business Profile is now where the entire customer journey happens.

Discovery happens there. Consideration happens there (photos, reviews, Q&A). Conversion happens there (call button, directions, booking).

Your website becomes a verification artifact. The AI and the GBP handle the funnel.

This is not a replacement for a website. But it does mean that an incomplete or generic GBP is a death sentence in the AI era.

---

From SEO to GEO: Generative Engine Optimization in Practice

Traditional SEO optimized for ranking in a list.

Generative Engine Optimization (GEO) optimizes for being cited in an answer.

The goal is no longer just “be position one.” The goal is “be the source the AI is quoting.”

Citation as the New Ranking

One large study found that brands cited in AI Overviews earn roughly 35% more organic clicks than brands that are not cited, even when both appear on the same SERP.

Why: being cited signals authority. It tells the user that this brand is the definitive source for that information.

Being ranked #1 below the AI Overview is not enough. You need to be inside the AI Overview.

The Barbell Content Strategy

To survive in the AI era, you need to avoid the “kill zone.”

The kill zone is the middle ground. Generic definitions. Basic how-tos. Aggregated lists of tips. This is content that AI Overviews can synthesize perfectly.

Investing in kill-zone content in 2025 yields diminishing returns.

Instead, adopt a barbell strategy:

Side A: Proprietary, First-Hand, Experience-Based Content

AI models cannot generate new data. They can only synthesize existing data.

Content based on original research, proprietary surveys, first-hand experiments, and real stories from your team is highly valuable.

Studies show that adding clear statistics boosts GEO performance by around 30%. Adding expert quotes boosts it by around 41%.

Examples:

• “We tested 12 HVAC brands over 6 months. Here’s what broke.”
• “Our moving crew tracked 200 jobs. Here’s the #1 reason pianos get damaged.”
• “We surveyed 500 customers. 78% said transparent pricing mattered more than price.”

This content gives AI systems fresh information to cite.

Side B: Hyper-Transactional Content

Content focused on the immediate moment of conversion.

Pricing calculators. Stock availability. Appointment booking flows. Service area maps with real-time availability.

AI summaries can describe your services, but they can’t book an appointment or show real-time pricing.

Users still need to click for these actions.

The Kill Zone:

“What is a mortgage?” “How to choose an HVAC company.” “10 tips for moving without stress.”

This is content AI Overviews eat for breakfast. Stop writing it.

Technical Practices for LLMs

Optimizing for an AI system requires different technical standards than optimizing for a traditional crawler.

1. Answer-First Formatting

Structure content with the answer immediately following the question.

Example:

Q: Do you offer same-day HVAC repair in Dallas? A: Yes. We offer 24/7 emergency HVAC repair with same-day service, including weekends and holidays. Call 214-555-1234.

This is easy for AI systems to extract and cite.

2. Semantic HTML and Schema

Use FAQPage, HowTo, LocalBusiness, Service, and Product schema on all relevant pages.

AI systems can read HTML, but schema removes ambiguity. It tells the AI exactly what each piece of data represents.

3. Source Preference and Consistency

Research from academic institutions suggests that AI models exhibit “source preference bias.”

Once a domain is established as a reliable source for a specific topic, the model enters a positive feedback loop. It cites that domain more frequently for related queries.

Implication: establishing early authority in the AI era is critical for long-term visibility.

If you’re the first clean, consistent, well-structured source an AI encounters for a topic, you have an advantage.

---

How Surmado Helps You Survive the Great Decoupling

The shift to AI-first search requires new tools and new metrics.

Surmado offers three products designed specifically for this era.

Surmado Signal: Measuring AI Share of Voice

Traditional rank trackers tell you where you rank on Google.

Surmado Signal tells you how often and how well AI platforms talk about your business.

What Signal tests:

• 7 AI platforms: ChatGPT, Claude, Gemini, Perplexity, Meta AI, Grok, DeepSeek.
• 5 buyer personas asking 50+ questions about your industry.
• Presence Rate: how often you’re mentioned (0-100%).
• Authority Score: how confidently AI systems recommend you (0-100%).
• Ghost Influence: how often competitors are mentioned instead of you.

Why this matters:

You can’t fix what you can’t measure. Signal shows you which queries and personas you’ve already lost to zero-click search.

It establishes your baseline for GEO campaigns.

Cost: $50 (2 credits). One async report delivered in 15-30 minutes.

Surmado Scan: Fixing the Technical and Content Plumbing

AI systems can’t cite you if they can’t read you.

Scan audits your site for the technical and structural issues that block AI platforms from understanding your business.

What Scan checks:

• Schema markup (LocalBusiness, FAQ, Service, Product).
• Core Web Vitals (LCP, CLS, INP).
• Crawlability and indexability.
• Mobile performance.
• Accessibility and security.

What you get:

A prioritized action plan. 5-10 fixes ranked by impact.

Scan helps you implement the Clarity Signals and schema disciplines that GEO requires.

Cost: $25 or $50 (1-2 credits) depending on site size.

Surmado Solutions: Strategy in a Weird New Market

The AI era requires strategic decisions, not just tactical fixes.

Which content should you cut? Which should you double down on? How do you sequence experiments over 90 days?

Solutions runs a six-AI adversarial debate analyzing your business, competitive landscape, and market positioning.

What Solutions delivers:

• Prioritized recommendations with ROI analysis.
• Multi-quarter roadmap connecting SEO and GEO tactics.
• Real Options Valuation for high-uncertainty decisions.
• Adversarial critique to stress-test your assumptions.

Cost: $50 (2 credits).

Async, APIs, and Automation

Every Surmado report is async and webhook-friendly.

Devs and agencies can plug Surmado into their own stack to monitor GEO and AI visibility over time.

You can run Scan, Signal, and Solutions via API. You can pipe results into your dashboards, CRMs, or BI tools.

All reports return structured JSON. All long-running tasks support webhooks for completion notifications.

---

90-Day Playbook for Small Businesses and Agencies

Here’s a concrete, numbered plan you can follow.

Phase 1: Days 1-7 (Get Oriented)

1. Run Surmado Scan and Signal. Establish your baseline for technical health and AI visibility before changing anything.

2. Audit Search Console. Look for the pattern: impressions up, clicks down. Identify queries where AI Overviews are likely appearing.

3. Review your Google Business Profile. Check that name, category, hours, photos, description, and services are complete and accurate.

4. Check cross-directory consistency. Compare your NAP data across Google, Yelp, Facebook, Apple Maps, and your website. Fix any discrepancies.

5. Read 10 recent reviews. Look for descriptive language mentioning your differentiators. Note any patterns in what customers say.

Phase 2: Days 8-30 (Fix the Obvious Leaks)

6. Clean up NAP inconsistencies. Update every directory listing to match your canonical NAP exactly.

7. Add or fix schema markup. Implement LocalBusiness, FAQPage, and Service schema on key pages. Use Google’s Rich Results Test to validate.

8. Add one plain-language FAQ section per core service. Write FAQs the way a real customer or AI would ask. Answer directly in the first sentence.

9. Request detailed reviews from recent customers. Guide them to mention specific features in their reviews, not just “great service.”

10. Fix Core Web Vitals issues flagged by Scan. Prioritize fixes that improve mobile performance and accessibility.

Phase 3: Days 31-90 (Build GEO Assets)

11. Publish one proprietary or experience-based piece. This is your Side A barbell content. Original data, actual experiments, real outcomes.

12. Build or improve one hyper-transactional asset. Pricing calculator. Service area map. Real-time availability checker. Booking flow.

13. Re-run Surmado Signal and Scan. Measure changes in AI visibility, Presence Rate, Authority Score, and critical technical issues.

14. Document results and refine. What moved? What didn’t? What new Ghost Influence patterns appeared? Adjust and iterate.

---

For Developers and Agencies: Metrics and Automation

If you’re an agency or dev team, you need to track AI visibility as a complement to keyword rankings.

AI Visibility Metrics

Share of Voice: How often your brand appears in AI-generated answers compared to competitors.

Presence Rate: Percentage of persona-based queries where you’re mentioned (0-100%).

Authority Score: Confidence level AI systems assign to your brand (0-100%).

Ghost Influence: Frequency of your features being attributed to competitors.

API and Webhook Integration

Use Surmado’s API to:

• Run Scan, Signal, and Solutions programmatically.
• Track how AI recommendations shift by persona over time.
• Monitor competitor movements in AI visibility.
• Pipe results into your own dashboards, CRMs, or reporting tools.

Example workflow:

1. Client onboarding: run Scan and Signal via API.
2. Store baseline metrics in your CRM.
3. Run Signal monthly via webhook.
4. Track Presence Rate and Authority Score over time.
5. Generate monthly reports showing AI visibility trends alongside Google Analytics.

This lets you pitch “AI visibility reporting” as a service without building your own multi-AI testing infrastructure.

---

The Bottom Line

The eighteen months from mid-2024 through late 2025 will be remembered as the era when the search economy broke.

Impressions and visibility decoupled from traffic and revenue.

AI Overviews absorbed the informational queries that used to drive top-of-funnel awareness.

Zero-click search became the dominant user behavior.

For small businesses and agencies, the path forward is clear:

1. Accept that traffic volume is no longer the north star. Share of Voice in AI answers is.

2. Shift investment from kill-zone content to proprietary data and hyper-transactional assets.

3. Obsess over Clarity Signals: cross-directory consistency, detailed reviews, clean schema, and authoritative citations.

4. Measure AI visibility as aggressively as you measure keyword rankings.

5. Use tools like Surmado Scan, Signal, and Solutions to see what AI systems actually say about you, not what you hope they say.

The businesses that master Generative Engine Optimization in 2025 will dominate their markets.

The ones that cling to the old referral model will watch their competitors get recommended by AI while they wonder why their traffic looks broken.

Your impressions are up. Your clicks are down. The world changed.

Now you know what to do about it.

---

Related Reading:

• What is Generative Engine Optimization (GEO)?
• AI Visibility: How ChatGPT, Claude, and Gemini Discover Your Business
• How AI Platforms Discover Your Business
• Google Business Profile Optimization
• Schema Markup for GEO

Take Action:

• Test Your AI Visibility with Signal ($50, 2 credits)
• Audit Your Technical Foundation with Scan ($25-50, 1-2 credits)
• Get Strategic Guidance with Solutions ($50, 2 credits)