The Number Everyone Wants — and Why It's Complicated
Let's start with what people are actually asking: what share of searches now go to AI instead of websites?
The honest answer is that there is no single clean number, because "going to AI" means at least three different things depending on who you ask:
- Zero-click searches — queries that end on the search results page without any outbound click (includes featured snippets, knowledge panels, AI Overviews, and plain old abandoned searches).
- AI Overview impressions — queries where Google displays an AI-generated summary above the blue links, whether or not the user clicks anything.
- Queries routed to alternative AI engines — searches that never reach Google at all, going instead to ChatGPT, Perplexity, Claude, or Gemini as a standalone tool.
These three buckets overlap but are not identical. Here is what the available data actually shows for each.
Zero-Click Rate: 58–65% as of Mid-2026
SparkToro and Datos published the most-cited zero-click studies. Their 2024 data showed approximately 58.5% of US Google searches ending without a click. Updated estimates for mid-2026 put that figure at 60–65% for desktop and slightly lower for mobile, where users are more likely to tap a result.
To be clear: zero-click search predates AI Overviews. Google has been answering queries directly — via Knowledge Graph cards, featured snippets, and local packs — since at least 2016. What AI Overviews changed is which queries now resolve on-page and how much real estate the AI answer occupies before the first blue link.
Before AI Overviews launched broadly (late 2023 into 2024), the zero-click rate was hovering around 50–53%. The jump to 60–65% over 18 months is substantial and directionally attributable to AI-generated summaries, though Google has not published its own breakdown.
AI Overviews Coverage: Roughly 30–40% of Queries
Third-party tracking tools — including SE Ranking, BrightEdge, and Semrush — have tracked AI Overview appearance rates since the feature's broad rollout. As of Q3 2026, the consensus estimate is that AI Overviews appear on 30–40% of all Google searches, with significant variation by query type:
- Informational / how-to queries: AI Overview appears on 55–70% of results pages
- Comparison queries ("X vs Y"): 45–60%
- Transactional queries ("buy," "price," "near me"): 10–20%
- Local queries ("[service] in [city]"): 8–15%
- Navigational queries (brand + site name): under 5%
This distribution matters enormously for how you think about your own traffic exposure. If your site primarily captures informational traffic — blog posts, guides, FAQs — you are in the highest-risk category. If you primarily capture local-intent or transactional traffic, the AI disruption is real but significantly smaller.
The Alternative AI Engine Problem Is Separate
Google's AI Overviews get the most press because Google still commands roughly 88% of global search market share. But the rise of ChatGPT, Perplexity, and similar tools as primary search interfaces represents a parallel shift that is harder to measure.
OpenAI has reported over 300 million weekly active users for ChatGPT as of early 2026. Perplexity claims approximately 100 million monthly active users. Rough estimates of query volume — extrapolated from session data and user surveys — suggest these platforms collectively handle 14–18 billion queries per month globally, compared to Google's roughly 8.5 billion queries per day.
So alternative AI engines are still a fraction of Google's volume. But the mix matters: users who go to Perplexity or ChatGPT for research are disproportionately high-intent, higher-income, and more likely to be in a buying or decision-making mode. Losing visibility in those surfaces is not just a traffic problem — it is a reach problem with a specific demographic.
What Actually Happens to Click-Through Rates
The fear is that AI Overviews kill all clicks. The reality is more nuanced.
Studies from SE Ranking and Search Engine Land (Q1–Q2 2026) found that when an AI Overview appears, organic CTR for positions 1–3 drops by 15–25% compared to the same query without an AI Overview. That is significant but not catastrophic for sites that rank in the top three.
For positions 4–10, the CTR drop is steeper — 30–45% — because users who would have scrolled past the AI Overview to find a result further down the page are less likely to do so now.
The counterintuitive finding: sites that are cited inside the AI Overview as a source see a CTR lift of 20–40% on those citations, even though the total click volume for the query is lower. The user who clicks through from an AI Overview citation is already partially sold — they want the full context, not just the summary. Conversion rates on that traffic are measurably higher.
This is the core strategic implication: the goal is no longer just to rank — it is to be cited.
Which Businesses Are Most Exposed
Not all owner-operators are equally affected. Here is a practical breakdown:
High exposure:
- Blogs and content sites that monetize informational traffic (how-to guides, definitions, comparisons)
- E-commerce sites that rely on comparison or review content to drive top-of-funnel traffic
- Service businesses that publish educational content to attract leads
Moderate exposure:
- Local service businesses with strong Google Business Profile presence (AI Overviews appear less often on local queries, and local packs still drive calls and directions)
- E-commerce stores with strong branded search and repeat customers
Lower exposure:
- Businesses that primarily capture navigational traffic (users already know the brand and search for it directly)
- Businesses with strong direct traffic channels (email lists, social audiences, referral networks)
If you run a salon, a plumbing company, or a local retail shop, your primary search surface — local pack results, Google Business Profile, maps — is less disrupted by AI Overviews than a national content publisher's is. That does not mean you are immune, but the urgency is different.
How to Position Your Content to Be Cited, Not Replaced
The shift from ranking to citation requires a different kind of content architecture. AI systems — whether Google's or a standalone LLM — favor content that is:
Specific and attributable. Broad overviews get summarized and discarded. Specific data points, original findings, and first-hand expertise get cited because the AI cannot generate them from training data alone.
Structured for extraction. Content with clear headers, defined terms, numbered steps, and explicit answers to questions is easier for AI systems to parse and quote. This is the core logic behind Answer Engine Optimization (AEO): write content that answers questions in a form AI can lift verbatim and attribute.
Authoritative on a narrow topic. A site that covers one topic deeply — with consistent publishing, clear authorship, and structured schema — is more likely to be treated as a reliable source than a site that covers everything shallowly. Owner-operators who genuinely know their trade have a natural advantage here if they publish what they know in structured form.
Schema-marked. Schema.org markup — particularly FAQPage, HowTo, Article, and DefinedTerm types — signals to both Google and LLMs that your content is structured, authoritative, and extractable. It does not guarantee citation, but it removes friction.
The goal is no longer just to rank — it is to be cited. And being cited starts with writing content that AI systems can extract, attribute, and trust.
The Measurement Problem
One underappreciated consequence of the AI search shift is that standard analytics tools are increasingly blind to AI-referred traffic. When a user reads an AI Overview and clicks a cited source, Google Analytics typically records that session as organic search. When a user reads a Perplexity answer and clicks a source link, that often shows up as direct or referral traffic, depending on UTM handling.
This means your traffic reports may be understating how much AI-mediated traffic you already receive — and overstating the pure algorithmic-ranking contribution. If you have noticed organic traffic declining while conversion rates on remaining organic traffic are rising, AI citation is a plausible explanation worth investigating.
The practical response: tag your content carefully, monitor referral sources from known AI domains (perplexity.ai, chatgpt.com, bing.com/chat), and track branded search volume as a proxy for overall brand awareness — including awareness built through AI citations.
What the Numbers Mean for Your Content Investment
If you are deciding how to allocate content effort in late 2026, here is the honest calculus:
- Informational content still has value, but the return now comes more from AI citation and brand authority than from direct click volume. Publish it with AEO structure or do not publish it at all.
- Transactional and local content (product pages, service pages, location pages) is less disrupted and should still be optimized for traditional ranking signals.
- Original data, case studies, and first-hand expertise are the highest-value content types right now, because they are the hardest for AI to replicate and the most likely to be cited as sources.
- Thin content — posts that restate what is already widely known — is being commoditized faster than ever. AI systems can generate it themselves; they have no reason to cite yours.
The owner-operator who writes a detailed post about how they reduced no-show rates at their salon by 40% using a specific waitlist protocol has something no AI can fabricate. That post will get cited. The owner-operator who publishes a generic "5 tips for reducing no-shows" post is competing with AI-generated content that is indistinguishable from theirs.
