On 17 July 2026, Nick Fox, Google’s Senior Vice President of Knowledge and Information, published a widely discussed claim. AI features in Google Search, he wrote, are “now sending billions of clicks to websites every week.” The SEO community pushed back the same day. Independent researchers had already measured something different — Google AI Overviews website traffic data showing a 58% drop in click-through rates for pages sitting in the top search position.

Both sides are telling the truth. That, precisely, is why the Google AI Overviews website traffic debate is so consequential — and why most of the coverage has been useless for Malaysian business owners.

This article explains why the data appears contradictory, what it actually means for a business in Malaysia or Singapore, and what to do differently starting this month. If your website traffic has been sliding despite stable keyword rankings, you are not imagining it. However, the explanation — and the fix — is more specific than most coverage suggests.

What Google Actually Said — and What It Chose Not to Share

The “billions of clicks” claim, unpacked

Nick Fox’s post was specific in one direction and vague in another. The specific claim: AI features in Google Search alone now send billions of clicks to websites every week. The vague part: no denominator, no per-site breakdown, no comparison to pre-AI-Overview click volumes, and no data by industry or query type.

Two supporting features received mentions in the post. Preferred Sources allows users to designate websites they trust — links from those sites then appear with a visible badge inside AI Overviews and AI Mode. Google reports that these labeled links roughly doubled click-through rates, and that users have selected more than 345,000 unique sources. Subscription Linking, meanwhile, highlights subscription content for paying readers inside Google surfaces.

Both features exist and both matter. However, neither came with supporting click-level data for the “billions” claim. Industry observers immediately asked for the breakdown. Google has not provided it.

Why the number is real — and irrelevant to your business

Google processes approximately 8.5 billion searches every day globally. “Billions of clicks weekly” across that volume is not a high threshold. Moreover, aggregate volume and individual page performance are different measurements — and the two can move in opposite directions simultaneously.

Your business does not receive billions of anything. It receives a click-through rate on a specific set of queries in a specific market. For a legal firm in Johor Bahru, a dental clinic in Subang Jaya, or an accounting practice in Singapore, the global aggregate tells you almost nothing. The click-through rate on the queries where your pages actually rank is the only number that determines whether your SEO investment is working.

In other words, the size of the pool matters far less than your share of it.

The Data Google Doesn’t Publish

What the independent research actually shows

Three studies, all independent of Google, point in the same direction. Together, they form the clearest picture available of how Google AI Overviews website traffic patterns have shifted in 2026.

Ahrefs, February 2026 — an analysis covering 300,000 keywords using actual Search Console data found that pages ranking in position one showed a 58% lower average click-through rate when an AI Overview appeared above them. In Ahrefs’s earlier 2025 study, that figure was 34.5%. The effect worsened as AI Overviews expanded to more queries.

Agarwal & Sen, Carnegie Mellon / Indian School of Business, June 2026 — a randomised field experiment with 1,065 participants found that AI Overviews reduced organic clicks to external websites by 39.8%, while zero-click searches increased by 34.5%. Sponsored results were unaffected. A randomised experiment controls for confounding variables that aggregate data cannot.

Pew Research — users clicked a traditional result 8% of the time when an AI summary appeared, compared with 15% without one. Links cited inside the AI summary itself attracted clicks on just 1% of visits.

Google has challenged some methodologies. That said, three independent research efforts reaching consistent conclusions are difficult to set aside.

The 68% zero-click reality

SimilarWeb data places zero-click searches — queries that end without a visit to any external website — at 68% of all Google searches in 2026. A decade ago, that figure was 45%. The steepest acceleration arrived in the past two years, driven largely by AI Overviews expanding across more than 20% of all queries.

Google AI Mode surpassed one billion monthly users at Google I/O 2026. Total search volume is growing, because people use Google more when they can ask longer, conversational questions. More searches, however, does not automatically produce more website visits. A growing share of those queries now resolves inside the results page itself — and never sends a visitor to any external site.

Why Both Stories Can Be True — and What It Means for Malaysian SMEs

The aggregate-vs-rate distinction

Google and the critics are both technically accurate. Aggregate weekly clicks can grow when total query volume rises fast enough, even while the click-through rate on each individual page falls at the same time. Consider a shopping mall that doubles in size. Total foot traffic rises. Nevertheless, each individual tenant may draw a smaller share of visitors than before, because there are simply more places for shoppers to go.

Global publishers with millions of indexed pages benefit from volume growth even at lower per-page click rates. Your business, however, is not a global publisher. A clinic in Johor Bahru, a legal firm in Singapore, or a retailer in Petaling Jaya operates on a finite set of specific queries in a specific market. For those businesses, the rate on each individual page — not the global aggregate — is the meaningful number.

The metrics your business should actually track

Stop measuring SEO success purely by organic sessions. In 2026, three additional metrics deserve equal attention.

The first is AI impression data. Google Search Console launched a dedicated generative AI performance report on 3 June 2026. It shows how often your pages appear in AI Overviews, AI Mode, and other generative features — separately from traditional search impressions. Most Malaysian businesses have not yet opened this report.

The second is branded search volume. When an AI answer references your business, many users search your brand name directly afterwards rather than clicking through immediately. Rising branded search alongside falling informational traffic is a recognisable pattern, not a coincidence.

The third is direct traffic. AI awareness frequently converts to direct visits later — someone encounters your brand in an AI answer, then returns to your site by typing the URL directly. Attribution tools rarely capture this connection, but the volume is real and growing.

Four Signs Google AI Overviews Are Already Affecting Your Traffic

These symptoms appear consistently across Malaysian business accounts. Check your Search Console data against each one before drawing conclusions about your overall SEO performance.

1. Impressions are rising but clicks are flat or falling.
Compare impressions versus clicks over the past twelve months in Search Console. If impressions trend upward while click-through rate compresses, AI Overviews are likely appearing on queries where you rank. The generative AI performance report separates this view clearly. Many businesses mistake rising impressions for a positive signal — they are not, when clicks fail to follow.

2. Top-ranked pages are losing traffic without losing their rankings.
A page holds position one. However, click-through rate has dropped from 12% to 5% over six months. Nothing on the page changed. What changed is that an AI-generated answer now appears above it for many queries in your category. This is the clearest and most common symptom.

3. Informational content has lost more traffic than service pages.
AI Overviews resolve informational queries more often than transactional ones — definitions, how-to guides, general explanations. Therefore, if your traffic decline concentrates on blog posts and educational content rather than service pages and contact forms, AI intercept is the most likely explanation.

4. Branded search volume is rising unexpectedly.
This is a positive signal worth monitoring. When someone encounters your brand cited inside an AI answer, they often search your brand name directly afterwards rather than clicking through in that moment. Rising branded searches alongside falling informational traffic is consistent with AI visibility without immediate click-through — and it is an audience worth building on.

The Strategy Shift: From Ranking to Being Cited

What content gets cited in AI Overviews (and what doesn’t)

Google’s AI surfaces do not simply pull from the highest-ranking page. They pull from content that answers questions directly, supports claims with verifiable evidence, and demonstrates topical authority across a subject — not just on a single article.

The Google AI Overviews system responds strongly to E-E-A-T signals — experience, expertise, authoritativeness, and trustworthiness — which have always mattered for traditional SEO and now matter even more for AI citation. Named authorship, professional credentials, local data points, and original research all improve a page’s probability of being cited.

For Malaysian and Singaporean businesses, this creates a specific advantage. Local data is genuinely scarce in AI training sets. Consequently, a page that cites specific JB property pricing, Singapore SME salary benchmarks, or Malaysian consumer behaviour statistics is more likely to appear in local AI answers than a page recycling global figures. Specificity and local authority are advantages here — not limitations.

Google’s Preferred Sources feature reinforces this further. When a reader designates your website as a trusted source, links to your pages receive a visible badge across AI Overviews and AI Mode. Your existing audience can therefore become a direct pathway into AI-surfaced content.

Content architecture changes that improve AI citability

Building content clusters around your core topic areas signals to Google’s AI that your site owns a subject, not just a single keyword. Isolated articles on unrelated topics rarely establish the topical authority that earns consistent AI citation. In other words, five deep articles on a connected subject consistently outperform fifteen scattered posts across different topics.

Direct answers in the first 100 words of any page are the highest-impact content change most Malaysian business websites can make today. If a page targets the query “how long does trademark registration take in Malaysia,” that answer belongs in the first two sentences — not buried eight paragraphs down.

Structured markup also matters significantly. FAQ schema, Article schema, and HowTo schema allow Google’s systems to parse answers without ambiguity. For search engine optimisation in the AI era, structured markup is the mechanism by which content becomes machine-readable across AI surfaces. AI citations for backlinks follow from the same logic — content cited by AI answers tends to attract more traditional backlinks as well, because AI citation signals authority to other publishers.

What Malaysian Businesses Should Do This Month

The following four steps require no external tools and no additional budget. Each takes between twenty minutes and one hour. Together, they will give you the clearest picture available of how Google AI Overviews website traffic is affecting your specific pages — and what to change first.

Step 1: Open the Generative AI performance report in Search Console.
Go to Performance → Generative AI. This report has been available since 3 June 2026. It shows which of your pages appear in AI features and on which queries. Most Malaysian businesses have not yet opened it. The data is already there, waiting.

Step 2: Identify your highest-impression, lowest-CTR queries.
Sort your performance report by impressions, descending. Then filter for click-through rates below 3%. These pages have strong visibility — but something is intercepting the click before it reaches your site. For most businesses, AI Overviews are the most likely cause.

Step 3: Rewrite the opening paragraph of your top five pages.
Take the question each page targets. Answer it directly and completely in the first two sentences. Then expand. This single change improves AI citability more reliably than most technical optimisations — and it costs nothing.

Step 4: Add FAQ schema to your core service and blog pages.
Google’s AI consistently extracts FAQ content. If your page includes a frequently asked questions section, wrap it in FAQ schema markup. Google’s Structured Data Markup Helper makes this straightforward without developer support.

These steps will not reverse the structural shift in search behaviour overnight. However, they will place your content in a meaningfully better position to appear inside AI answers — rather than ranking below them without being cited.

How Xwork Builds for AI Search Visibility — Not Just Rankings

Most SEO agencies in Malaysia still measure success by keyword rankings and monthly organic sessions. That was an accurate measure in 2022. In 2026, however, it is an incomplete one — it tells you how often you appear in traditional blue links, while saying nothing about whether Google’s AI answer is citing you, ignoring you, or surfacing a competitor instead.

Xwork’s CODE/RAVEN RANK GENERAL protocol runs two complementary layers simultaneously.

The first is Phantom — Xwork’s dedicated GEO (Generative Engine Optimisation) and AI Search strategy service. Phantom audits existing content for citability, optimises new content so that AI Overviews, Gemini, ChatGPT, Perplexity, and Copilot are more likely to surface it, and monitors whether your brand appears inside AI-generated answers — not just in traditional results. It is a capability that no other digital marketing agency in the JB–Singapore corridor currently offers under one roof.

The second layer is the traditional SEO foundation: keyword strategy, content architecture, technical optimisation, and four blog posts per month — each simultaneously optimised for ranking and for AI citation.

A RANK GENERAL client receives the following each month:

• Phantom AI citation audit and optimisation pass on core pages
• Four blog posts targeting both ranking keywords and AI surface queries
• Monthly analytics report separating traditional and AI search visibility performance
• Brand monitoring across ChatGPT, Gemini, Perplexity, Copilot, and AI Overviews
• Keyword and topical authority architecture review

RANK GENERAL starts at RM 2,500 per month with a six-month minimum. The minimum exists because AI search visibility compounds. The first two months establish the content architecture. Months three and four produce measurable citation signals. By month six, you hold a defensible position in both traditional and AI search for your core topics.

For Malaysian businesses ready to measure both dimensions of their search presence: learn more about RANK GENERAL.