The Real Numbers Behind AI Search Eating Blue-Link Traffic

The Question Everyone Is Asking, Finally With Numbers

For the past two years, marketers have been talking about the death of the blue link as though it were a future event. It isn't. The shift has already happened across a meaningful share of search volume, and the data — fragmented as it is — now tells a clear enough story to act on.

Here is what we know as of mid-2026, where the numbers come from, and what they actually mean if you run a business that depends on organic search for new customers.

Zero-Click Search: The Baseline Before AI Overviews

Zero-click search — queries that end without the user visiting any external website — predates generative AI. Featured snippets, knowledge panels, People Also Ask boxes, and local map packs have been absorbing clicks since at least 2016.

By 2022, Semrush and SparkToro research estimated that roughly 50% of Google searches ended without a click. That figure was already alarming for publishers. Then Google launched AI Overviews in the US in May 2024, and the trajectory steepened.

By Q2 2026, multiple independent measurement sources — including Datos clickstream panels, BrightEdge's query-type analysis, and Semrush's Traffic Analytics — converge on a range of 58–65% of all Google searches ending without an external click. The variance comes from methodology: clickstream panels capture desktop behavior better than mobile, and mobile now accounts for roughly 63% of all searches.

Where AI Overviews Actually Show Up

Not every search gets an AI Overview. Google has been selective about deployment, and the pattern is now legible:

High AI Overview frequency (>50% of queries in category):

• Health and medical symptoms
• Financial definitions and how-to
• DIY home repair
• Recipe and cooking technique
• General 'what is' / 'how does' queries

Moderate AI Overview frequency (15–35%):

• Software comparisons and reviews
• Travel planning
• Local area information (not 'near me' transactional)
• Educational content

Low AI Overview frequency (<10%):

• 'Near me' and local transactional queries
• Brand-name navigational searches
• Product purchase queries with commercial intent
• News and current events

BrightEdge's mid-2026 data puts AI Overviews appearing on approximately 25–35% of all US desktop queries. That number is lower on mobile (closer to 18–22%) because screen real estate compresses the format.

The practical takeaway: if your business generates leads from informational content — blog posts explaining how something works, comparison guides, FAQ pages — those pages are the ones getting displaced. If your business generates leads from 'best [service] near [city]' or '[product] buy' queries, you are largely still in the game.

The Click-Through Rate Collapse on Informational Queries

Here is where it gets concrete for owner-operators. Advanced Web Ranking's CTR study for Q1–Q2 2026 shows that when an AI Overview is present:

• Position 1 organic CTR drops from ~13% to 4–6%
• Positions 2–5 become nearly invisible, with combined CTR under 3%
• The AI Overview itself links to 3–5 sources, but those citation links receive only 1–3% CTR each

For a site that was getting 10,000 monthly visits from a position-1 informational ranking, that can translate to a drop to 3,000–5,000 visits — not from losing the ranking, but from the query resolving before the user scrolls to the blue link.

This is why many site owners in 2025–2026 reported ranking stable or even improving in position while watching traffic fall. The ranking didn't move. The query did.

What Is Still Sending Clicks

The data is not uniformly bad. Three query categories remain click-heavy:

1. Transactional queries with purchase intent. Searches like 'buy running shoes size 11' or 'book haircut downtown Austin' still resolve in clicks because the user needs to complete an action on an external site. Google has limited incentive to eliminate this traffic — it's the foundation of the ads ecosystem.

2. Local 'near me' and map-pack queries. According to Google's own search quality disclosures and third-party local SEO tracking from BrightLocal, local intent queries still drive map-pack clicks more than 70% of the time. For a salon, restaurant, auto shop, or any business where the customer needs to physically show up or call, local search remains the highest-value channel and is relatively insulated from AI Overview displacement.

3. Navigational and branded queries. When someone searches your brand name, they are going to your site. AI doesn't intercept this.

The implication is significant: if you are a local business, the search landscape has not collapsed for you as badly as the aggregate numbers suggest. The sites that have been devastated are content publishers, affiliate sites, and businesses that built their entire acquisition model on ranking for informational keywords.

The Rise of Non-Google AI Answer Engines

Google is not the only surface to track. Perplexity, ChatGPT Search, and Microsoft Copilot now collectively handle an estimated 3–5% of total search volume in the US — small in absolute terms but growing at roughly 80% year-over-year as of mid-2026.

More importantly, these platforms skew toward the exact user profile most valuable to small businesses: higher-income, tech-forward consumers who are comparison-shopping or researching a significant purchase. A 2025 study from Datos found that Perplexity users have median household incomes 30% above the US average.

For owner-operators, appearing as a cited source in Perplexity or ChatGPT Search answers is a different optimization problem than traditional SEO. These engines weight structured, authoritative, recently-updated content and pull heavily from sources with clean schema markup and strong E-E-A-T signals. The post on how Perplexity's indexing has changed covers the mechanics in detail.

The Honest Traffic Math

Let's put numbers to a real scenario. Suppose a local service business (a plumbing company, a dental practice, a gym) has historically gotten:

• 40% of organic traffic from informational blog content ('how to fix a leaky faucet', 'what causes tooth sensitivity')
• 40% from local/transactional queries ('plumber near me', 'dentist open Saturday')
• 20% from branded searches

Applying the AI Overview CTR reduction to the informational segment, that 40% bucket likely shrinks to 15–20% of its former volume. The local and branded segments remain largely intact. Net result: total organic traffic down 20–30%, but lead-generating traffic down far less — because the informational content was mostly top-of-funnel awareness, not bottom-of-funnel conversion.

This is why the panicked headlines about 'SEO is dead' overstate the case for local and transactional businesses, while underplaying the genuine crisis for content-first publishers.

What This Means for How You Spend Your Content Time

The strategic shift is straightforward once you accept the data:

Stop: Writing informational blog content designed to rank for generic how-to queries. AI will answer those for free, and even a position-1 ranking now delivers a fraction of the clicks it used to.

Start: Writing content that establishes your business as a citable authority — the kind of specific, experience-backed content that AI engines quote rather than paraphrase. First-person case studies, proprietary data, local-specific guides, and opinion pieces from a named expert are harder for AI to synthesize and more likely to earn citation links inside AI answers.

Double down: On local search signals — Google Business Profile completeness, review velocity, NAP consistency, and local schema. These signals feed the map pack and local AI answers, and local intent queries are still sending clicks. See the review velocity benchmark data for specifics on what the numbers look like.

Measure differently: Stop using total organic sessions as your primary KPI. Segment by query type — branded, local/transactional, informational — and track each separately. The aggregate will look worse than the reality of what's actually driving leads.

The Number That Matters Most

All the percentages above are useful context, but the single number that matters for any individual business is simpler: what share of your current organic traffic comes from queries that still send clicks?

If you have Google Search Console access, filter your top 50 queries by clicks. Classify each as informational, transactional, local, or branded. The queries that are informational and generic are the ones at risk. Everything else is more durable than the aggregate statistics suggest.

The search landscape in 2026 is not uniformly hostile — it is selectively hostile, and the selection is legible if you look at the data by query type rather than in aggregate.