-58%

Position-1 CTR on AI Overview queries

Ahrefs, 300,000-keyword study, December 2023 vs December 2025.

48%

Share of Google searches with AI Overviews

March 2026. Up 58% year-over-year.

76% → 38%

Share of AIO citations from top-10 pages

Mid-2025 vs early 2026. Ranking matters less, passage structure matters more.

+35%

Organic clicks gain for brands cited in AIO

Seer Interactive, 53 brands, 5.47M queries analyzed.

The context: why classic SEO no longer cuts it in 2026

For 18 months, site owners have been getting the same two signals: their Google Analytics descending in a staircase pattern and their Google Search Console showing flat impressions but collapsing clicks. The diagnosis is the same everywhere: AI Overviews answers the question directly on the Google page, and the user no longer needs to click.

The real change isn't AI Overviews appearing. It's that this feature went from a gadget tested on 6% of queries in 2024 to a standard component of 48% of searches in March 2026. And the trajectory continues: AI Mode (the full chat experience to the right of the page) hit one billion monthly active users. When AI Mode is active, 93% of searches end without any click to an external site.

Independent data (Ahrefs, Seer Interactive, Digital Applied) confirms what our clients see in their GSC: traditional ranking-and-impression SEO delivers fewer and fewer actual clicks. But the story isn't linear, and that's what makes the 2026 strategy interesting.

What the numbers actually say

The Ahrefs study published in March 2026 is the largest to date: 300,000 keywords analyzed, split into two groups of 150,000 (with and without AI Overview), with GSC aggregate CTR data compared between December 2023 and December 2025.

The results are clear. On keywords without AI Overview, position-1 CTR went from 0.076 to 0.039, a 49% drop. That's already huge and reflects other changes (expanded featured snippets, People Also Ask, etc.). On keywords with AI Overview, position-1 CTR went from 0.073 to 0.016, a 58% drop. The difference (9 points) is the net effect attributable to AI Overviews alone.

The Seer Interactive study (53 brands, 5.47M queries, 2.43B impressions over 14 months) tells a complementary story. Organic CTR on AIO queries dropped 61% since mid-2024. Paid CTR on the same queries dropped 68%. But they also detected a recovery signal: average CTR within AIO itself, after bottoming at 1.3% in December 2025, recovered to 2.4% in February 2026.

Last critical point: sites without AIO on their main queries saw CTR increase, from 2.8% in early 2025 to 3.8% in February 2026. Users who don't get an immediate AI answer click on organic results more readily. The strategic consequence is clear: figuring out which of your queries do NOT trigger AIO becomes as important as understanding which ones do.

Source

On AI Overview keywords, position-1 CTR dropped from 0.073 in December 2023 to 0.016 in December 2025. For every 100 historical clicks, Google now keeps 58.

Ahrefs - AI Overviews CTR study (March 2026)Reference study on the real impact of AI Overviews on organic CTR, GSC methodology across 300,000 keywords.

Source

Organic CTR on AIO queries dropped 61% since mid-2024, paid CTR 68%. Brands cited within AI Overviews earn 35% more organic clicks and 91% more paid clicks compared to uncited brands.

Seer Interactive - organic and paid study (March 2026)Analysis of 53 brands, 5.47M queries, 2.43B impressions between January 2025 and February 2026.

What we actually change for our clients

Here are the four interventions we run on an account hit by an AI Overviews-driven drop. Order matters: each step informs the next.

1. GSC diagnostic crossing impressions vs clicks

Classic symptom of an AIO loss: impressions on your main queries stay flat or grow, but clicks descend. In Google Search Console, compare the last 28 days to the equivalent period a year ago, filter on your most important pages, and identify queries where CTR dropped by more than 30%. For each one, do a manual Google search: AIO present or not? If yes, who's cited? If no, investigate other causes.

2. Restructure content for AI citation

Pages cited in AIO aren't necessarily the best-ranked, but they share structural traits. Digital Applied's 2026 study (which evaluated AI citation factors): accessible URL not blocked by robots.txt (weight 9.5/10), direct answer to the query in the first 200 words of the section (weight 9.2/10), self-contained passages with clear headers (weight 8.6/10), headers framed as questions ("What is X?", "How much does Y cost?") rather than dry keywords.

In practice we retool pages as follows: each H2 becomes a question users actually type, the first paragraph under the H2 answers directly in 2-3 sentences, then you develop further. We also add specific numbers, dates, prices, comparative data. AI models heavily favor content with citable data.

3. Work on third-party mentions (reinforced E-E-A-T)

A fact often missed: no brand gets cited in AIO purely on its own content. Pages that get cited are almost systematically those whose subject is also treated or mentioned on credible third-party sites (press, sector directories, podcasts, authority blog mentions). 96% of AIO citations come from sources with strong E-E-A-T signals. It's digital PR and guest content work that moves the needle, not just proprietary content.

4. Acquisition channel diversification

When organic SEO loses 30 to 50% of its yield structurally, the channel arithmetic changes. On client accounts that depended 60-70% on SEO in 2023, we rebalanced to a mix: 30-40% SEO + GEO, 25-30% paid (Google Ads, Meta), 15-20% local (Google Business, reviews), 10-15% social organic (LinkedIn especially for B2B), and 5-10% referral inside AI tools (ChatGPT, Perplexity, Claude). The exact mix depends on industry, but single-channel dependence becomes an operational risk.

Priority by business type

Si

Content site / blog / media with SEO-dominant traffic

Alors

Urgent diversification toward direct distribution (newsletter, app, community).

You're the most exposed profile. HubSpot lost 70-80% and that's a site with massive domain authority. Content sites depending on AdSense or affiliate are in existential danger. Build a direct channel with your audience that you control 100%.

Si

Local services site (clinic, firm, restaurant, trades)

Alors

Pivot to local SEO + Google Business + reviews, and targeted paid by zone.

AIO touches local-intent queries less ("accountant Sherbrooke", "pediatric clinic Laval"). But informational questions around your trade ("what is an ADHD test?") are eaten by AIO. Focus your SEO on local commercial queries and invest paid for the rest.

Si

Product e-commerce

Alors

Merchant Center feed optimization + Performance Max + enriched product pages.

AI Shopping Ads and Direct Offers in AI Mode require a dense product feed. Product pages with structured FAQs, warranty data, precise comparisons. You enter the AI channel through shopping rather than blog content.

Si

B2B services (SaaS, agency, consulting)

Alors

Investment in depth content (case studies, papers, primary data) + organic LinkedIn + third-party mentions.

B2B keeps significant search volume on complex queries that users will click to validate. But 'what is X?' pages are dead in SEO. Invest in depth (original research, primary data, quantified case studies) and distribution on LinkedIn where your buyer consumes.

90-day plan to recover ground

  1. Weeks 1-2: diagnostic

    Compare your GSC on a 12-month rolling basis, identify the 50 keywords where CTR dropped by more than 30%. For each, do a manual Google search: AIO present? Who's cited? Log in a spreadsheet. Also identify keywords where AIO does NOT appear: these are your priority opportunities.

  2. Weeks 3-4: existing content audit

    On the 20 pages that lost the most, audit structure: are H2s questions? Do the first 200 words of each section answer directly? Are there numbers, data, citations? Do passages stand alone? If not, AI can't extract them.

  3. Weeks 5-8: restructure and publish

    Restructure the 20 priority pages with H2-question, answer-first, self-contained passages, addition of specific data and citations to primary sources. Also publish 5-10 new pieces targeting the no-AIO keywords identified in step 1.

  4. Weeks 9-12: third-party amplification

    Pitch your reworked content to 10 authority sites in your sector for mentions, citations, or guest posts. Start a LinkedIn cadence (3 posts/week) summarizing your new content. If you sell a product or service, launch or increase a paid campaign to plug the traffic hole in parallel.

  5. Week 13: measure and loop

    Compare your GSC for the past 4 weeks vs the previous 4 weeks. For the 20 restructured pages, check CTR. Go to your main AIO-bearing queries: are you cited now? If yes, measure click impact. If not, identify what distinguishes cited pages from yours and iterate.

Critère
Primary success metricPosition 1-3 ranking + impression volumeCitation in AIO + clicks from AIO + visibility in ChatGPT/Perplexity
Content structureLong-form dense, keywords in titleSelf-contained passages, H2 as question, answer-first
Backlink importanceVery high (ranking signal)Moderate (authority signal), third-party mentions matter more
Content freshnessImportant but not criticalCritical: cited content averages 25.7% fresher
Data and numbers in contentDesirableNear-mandatory to be cited
Schema markupImportant for rich resultsNot required per Google, but helps content comprehension
Time to first results3-12 months depending on competition4-12 weeks to appear in AIO

SMB owner FAQ

My SEO traffic dropped 40% over 6 months. Is it recoverable?

Partially, and it depends what type of pages you lost. If they were mostly 'what is X?' pages answering simple questions, that traffic isn't recoverable because AIO answers directly and users won't click. If they were product, services, or comparison bottom-funnel pages, yes, you can usually recover 60-80% with restructuring + third-party mentions work + adding citations. Rule: the more commercial the intent, the more recoverable.

Should I block GoogleBot to prevent my content from being cited in AIO?

No, and it would be counterproductive. Blocking GoogleBot removes your pages from Google's index entirely, so also from traditional search and all Google-driven traffic sources. The 'nosnippet' directive does prevent your content from being extracted into AIO, but it also removes you from featured snippets and reduces organic CTR. The economic math strongly favors letting Google index. Brands cited in AIO earn on average 35% more organic clicks than uncited brands on the same queries.

ChatGPT and Perplexity, is that really measurable traffic for an SMB?

Small but growing fast. On client accounts we monitor, referral traffic from ChatGPT, Perplexity, and Claude currently represents 2-8% of total organic traffic, versus nearly zero 18 months ago. For a B2B services site, AI traffic quality is often higher than classic SEO because the user arrives after a contextualized recommendation. Worth the work to structure your content to appear there, even if volume stays secondary to traditional SEO for now.

How long before a restructured page appears in AIO?

Most cited data points to 4-12 weeks after publication or restructuring. Faster than traditional SEO (3-12 months to move in average competition) because AI models are retrained or re-indexed more frequently. But never instant. If you restructure this week, look for first signals in the GSC AI Search report in 6-8 weeks.

Schema markup, is it really important for AIO?

Google confirmed in May 2026 that there is NO specific markup required for AIO and no 'AI' schema exists. But standard schema (Organization, FAQPage, HowTo, Article) remains useful because it helps Google understand the structure and context of your page, which is exactly what AI Overviews needs to decide whether to extract a passage. Keep your schema, don't invent a new type, don't pay anyone to 'optimize for AIO' with special schema that doesn't exist.

Source

Editorial content represents 53.46% of AI citations, news 14.09%, social 8.71%, press releases 0.04%. Pages with URL accessibility (9.5/10), self-contained structure (8.6/10), and answer-near-top (8.8/10) have the highest citation probabilities.

Digital Applied - AI citation ranking factors study (2026)Analysis of factors correlated with page selection as a source in AI Overviews, ChatGPT, Perplexity and other AI engines.

Verdict

AI Overviews isn't a wave that will pass. It's a structural change in how Google delivers the value of search, and it's accelerating. The numbers confirm what our clients see in their Google Analytics: classic ranking-and-impression SEO yields fewer and fewer actual clicks.

The good news: pages cited in AIO recover significantly more clicks than uncited pages, and queries without AIO recover lost CTR. The 2026 strategy isn't to abandon SEO, it's to restructure it for two types of win: get cited in AIO on your main queries, and dominate organic on bottom-funnel commercial queries where AIO doesn't appear (or barely).

If your organic traffic has dropped significantly since 2024 and you want to know precisely what's recoverable in your case, write us. We do the GSC impressions/clicks cross-diagnostic free for sites over 5,000 monthly organic sessions.

Free diagnostic: what you can recover from your lost SEO traffic

Book a call