Google AI Overviews are AI-generated answer summaries that now appear in roughly one in four UK search results (Ofcom, December 2025). Unlike ChatGPT — where only 8–12% of cited URLs overlap with Google's top 10 — AI Overviews draw 76–86% of their citations from pages already ranking organically (Ahrefs/BrightEdge, 2025). This makes AIO optimisation fundamentally an extension of traditional SEO, but with a critical citation layer on top. For UK businesses specifically, the CMA's proposed conduct requirements under the DMCCA could reshape how Google deploys AI features in this market — a regulatory dimension no other country has.
The stakes are significant. Position-1 organic results lose 58% of clicks when an AI Overview appears above them (Ahrefs, February 2026). Yet brands cited within AIOs retain 35% more clicks than non-cited competitors (Seer Interactive, November 2025). The difference between being cited and being bypassed is the difference between gaining from the AI shift and being displaced by it.
This guide covers how AI Overviews work, what drives citation, the UK regulatory context, and practical steps to protect and grow your visibility. It is part of Whitehat's Answer Engine Optimisation learning path. For ChatGPT-specific optimisation, see our companion guide to optimising for ChatGPT.
Google AI Overviews use Retrieval Augmented Generation (RAG) with a "query fan-out" technique. Google breaks your search query into multiple sub-queries, searches each independently against its organic index, then synthesises the results into a single AI-generated summary with inline citations. This explains why citations sometimes come from pages not ranking for the original query — they rank for a sub-component of it. Each AIO typically cites 4–5 URLs (Writesonic, 1M+ AIOs analysed), though complex queries can surface significantly more.
AIO triggers have shifted dramatically since the May 2025 global rollout. Informational queries accounted for 91.3% of triggers in January 2025 but fell to 57.1% by October 2025 (Semrush, 10M+ keyword study). Meanwhile, commercial queries grew from 8.15% to 18.57%, and transactional queries surged from 1.98% to 13.94%. Even navigational branded searches now face AIO interception, growing from 0.74% to 10.33% — a 1,296% increase. Question-based queries trigger AIOs 84% more often, with long-tail queries of 8+ words showing the highest rates.
Coverage has been volatile. Semrush tracked AIO presence rising from 6.49% of searches in January 2025 to a peak of 24.61% in July, then pulling back to 15.69% in November. Conductor's 2026 benchmarks report 25.11%. The reliable independent range is 15–25% globally, approximately 30% for the UK. Industry penetration varies: science (25.96%) and technology (17.92%) see the highest rates, while real estate and shopping remain minimal — Google protects high-CPC verticals for advertising revenue.
Google AI Overviews and ChatGPT share only 21.26% domain overlap (SE Ranking). The practical implications are substantial: what gets you cited by one platform may have no effect on the other. The table below summarises the key differences based on the latest research.
| Factor | Google AI Overviews | ChatGPT |
|---|---|---|
| Organic overlap | 76–86% from Google top 10 | Only 8–12% from Google top 10 |
| Top social source | YouTube (39.2%), Reddit (20.3%) | Wikipedia (7.8%), Reddit (3–4%) |
| Schema influence | FAQ schema = 60% more likely to be cited | FAQ schema shows no positive effect |
| Page speed impact | Weak correlation (r = −0.12 to −0.18) | Strong: FCP <0.4s = 3× more citations |
| Primary index | Google's own organic index | Bing + OpenAI's own index |
| Opt-out mechanism | None currently (CMA may force one in H1 2026) | Block OAI-SearchBot in robots.txt |
The strategic takeaway: if you already rank well in Google organic results, you are already most of the way to AIO visibility. The optimisation gap is narrower than for ChatGPT, but the actions required are different. Schema markup, YouTube presence, and multimodal content matter far more for AIOs, while page speed and Bing indexing matter more for ChatGPT. Whitehat's AEO service addresses both platforms within a single integrated strategy.
Semantic completeness is the strongest predictor of AIO selection, with one study reporting an r=0.87 correlation. Content scoring highly on completeness is 4.2 times more likely to appear. Practically, this means each key section of your page should work as a self-contained information unit — answerable without surrounding context. AIO-cited articles cover 62% more facts than non-cited ones (Surfer SEO, November 2025).
The research points to specific structural patterns that increase citation likelihood. Front-load your main answer within the first 150 words — 44.2% of all AI citations come from the first 30% of text (Growth Memo, February 2026). Optimal passage length for AIO extraction is 127–167 words, with 62% of featured content falling in this range. Content with 15 or more connected entities per 1,000 words shows 4.8 times higher selection probability. Statistics boost citation by 22%, and direct quotes from authoritative sources by 37%.
YouTube has overtaken Reddit as the top social citation source for AIOs. Goodie AI's analysis of 6.1 million citations (August–December 2025) found YouTube's share of social citations rose from 18.9% to 39.2% while Reddit dropped from 44.2% to 20.3% (Adweek, January 2026). YouTube wins because structured transcripts give AI systems clean parseable text, chapter markers enable section-specific extraction, and video demonstrates verifiable expertise. For B2B companies, this means creating video content with full transcripts and VideoObject schema is now a meaningful AIO optimisation tactic.
Pages combining text, images, video, and structured data see 156% higher AIO selection rates. Full multimodal content with schema reportedly drives 317% more citations. Brands are 6.5 times more likely to be cited via third-party sources than their own domains (AirOps, October 2025), which means earned media, Wikipedia accuracy, and review platform profiles continue to matter enormously — just as they do for ChatGPT optimisation.
Multiple rigorous studies confirm that AI Overviews significantly reduce organic CTR. Ahrefs' February 2026 study of 300,000 keywords found position-1 results lose 58% of clicks when an AIO appears, position 2 loses 50.8%, and position 3 loses 46.4%. The absolute CTR for position 1 on AIO queries fell from 7.3% in December 2023 to just 1.6% in December 2025. Pew Research's field study of 68,879 real searches found only 1% of users clicked a link within the AIO itself, and 26% ended their session after seeing one — versus 16% without an AIO present.
The UK publisher impact has been documented in CMA submissions. DMG Media (Daily Mail) reported desktop CTR dropping from 25.2% to 2.8% when AIOs appear, though their overall traffic impact remains "very low single-digit" because AIOs do not yet trigger for many commercially important queries. The PPA documented 10–25% year-on-year CTR declines across its approximately 250 UK member organisations despite stable rankings. Bauer Media confirmed automotive specification content was "hit hard."
However, there is a protective effect from citation. Seer Interactive's study of 3,119 queries across 42 organisations found that cited brands achieve 0.70% organic CTR versus 0.52% for non-cited brands. Cited brands also earn 91% more paid clicks (7.89% vs 4.14% CTR). The caveat: cited brands still saw organic CTR decline 49.4% year-on-year, while non-cited declined 65.2%. Citation provides protection against decline, not absolute growth.
The UK picture appears less severe than the US. Chartbeat data via Reuters Institute (January 2026) shows Google search traffic to publishers declined 17% in Europe versus 38% in the US. Bauer Media's Stuart Forrest reported AIO frequency has "plateaued" at roughly 12% of non-brand terms on mobile in the UK versus 19% in the US. This window may not last — but UK businesses currently have time to prepare.
The UK now operates the most aggressive regulatory framework for AI search globally. On 10 October 2025, Google became the first firm designated with Strategic Market Status under the Digital Markets, Competition and Consumers Act (DMCCA), covering both general search and search advertising. The CMA's 156-page decision established that Google holds over 90% UK search market share, handles approximately 3 billion monthly UK searches, and that 203,847 UK businesses spent £10.2 billion annually on Google Ads. AIOs and AI Mode are explicitly within scope.
On 28 January 2026, the CMA proposed four conduct requirements. The most significant for content publishers is the publisher opt-out: Google must provide "meaningful and effective" controls allowing publishers to opt out of AI Overviews and AI Mode content usage without downranking in traditional search, plus transparency reporting and engagement metrics. Other requirements include fair ranking with 30-day notice before major algorithm changes, enhanced user choice screens on Android and Chrome, and standardised Google Analytics data export via API.
Category 1 measures (publisher controls, fair ranking) are slated for implementation in H1 2026, with a 90-day grace period after final publication. Non-compliance penalties: up to 10% of global annual revenue. A second batch addressing whether Google must compensate publishers has been postponed 12+ months. Separately, the government must publish an economic impact assessment on AI's use of copyright works by 18 March 2026.
Reduced competition for citations means higher visibility for businesses that remain. Your AIO presence could strengthen significantly with fewer sources competing for the 4–5 citation slots.
AIO response quality may degrade in the UK, leading Google to reduce deployment. Businesses relying on AIO visibility could see that channel contract. Diversify across ChatGPT, Perplexity, and organic.
Optimise for both AIOs and traditional organic. Build multi-platform AI visibility. Monitor CMA developments. This way you benefit regardless of which regulatory outcome materialises.
Currently, no granular opt-out exists. Google-Extended in robots.txt controls AI model training for Gemini/Vertex AI but does not prevent AIO citations — Google confirmed this during May 2025 antitrust testimony. The nosnippet meta tag blocks AIO appearance but kills all snippet visibility. Until the CMA requirements take effect, the only options are full participation or full withdrawal. Whitehat's SEO services include ongoing monitoring of these regulatory developments and their practical implications for client strategies.
Google AI Mode launched at Google I/O in May 2025 and reached UK users on 28 July 2025. Powered by a custom Gemini 2.5 model, it offers full conversational search with follow-up questions, multi-step reasoning, and "Deep Search" for expert-level reports. It now has 75 million daily active users globally (Google's Nick Fox, Q3 2025 earnings call).
AI Mode operates very differently from standard AI Overviews. Only 13.7% citation overlap exists between the two for the same query (Ahrefs, December 2025, 730K query pairs). AI Mode responses are 4 times longer, cite approximately 9 unique domains per response versus AIOs' 7.7, and critically — only 14% of AI Mode-cited URLs rank in Google's top 10 (SE Ranking, August 2025). This makes AI Mode more similar to ChatGPT than to standard AIOs. Wikipedia appears in 28.9% of AI Mode citations versus 18.1% in AIOs, and Quora is cited 3.5 times more frequently.
The implication: 93% of AI Mode searches end without a click (Semrush, September 2025). As AI Mode usage grows, UK businesses need strategies that address both the high-organic-overlap world of standard AIOs and the low-overlap, conversational world of AI Mode. This means applying the dual strategy that Whitehat covers across the full AEO learning path — traditional SEO foundations for AIOs, plus answer capsules, entity clarity, and off-site signals for AI Mode and ChatGPT.
Google Search Console offers an AI Mode filter (available since June 2025) under Performance showing impressions, clicks, CTR, and position for AI Mode citations. However, AI Overview clicks are blended into standard organic data with no separate filter — if a page appears in both the AIO and organic results, clicks cannot be distinguished. GSC retains only 16 months of data via API, meaning pre-AIO baseline data is now falling off for early-rollout queries.
GA4 cannot distinguish AIO clicks from standard organic traffic. Industry estimates suggest true AI traffic is 2–3 times what analytics platforms report due to referrer stripping, particularly from mobile apps. For third-party AI sources, create a custom GA4 channel group using regex matching against known AI domains — the same approach recommended in our ChatGPT optimisation guide.
For dedicated AIO tracking, Semrush offers the most complete suite with Position Tracking AIO filters and an AI Visibility Toolkit (from $139.95/month). Ahrefs Brand Radar tracks 100M+ AI prompts across six platforms including Google AIOs (from $199/month). Budget-friendly options include Otterly.AI (from $29/month) and Advanced Web Ranking with dedicated UK AIO support (from ~$79/month). Whitehat recommends tracking AIO citation rate for target keywords, comparing CTR trends for AIO-triggering versus non-AIO queries in GSC, and monitoring branded search volume as a proxy for AIO brand exposure. For a comprehensive measurement approach, see our guide to AEO measurement and KPIs.
Since AIO optimisation builds on existing SEO, these actions focus on the citation layer — the specific additions that increase your chances of being cited within AI Overviews.
44.2% of citations come from the first 30% of content. Place your primary answer as the opening paragraph — not after two paragraphs of context. Each H2 section should begin with a 127–167-word passage that works as a standalone answer.
Pages with FAQ schema are 60% more likely to be featured in AIOs. Unlike ChatGPT (where schema shows no effect), Google actively uses structured data to parse and verify content. Include Article schema with dateModified, author credentials, and publisher details.
Cover topics comprehensively — AIO-cited articles contain 62% more facts. Target 15+ connected entities per 1,000 words. Include statistics (22% citation boost) and expert quotes (37% boost). Long-form content of 2,900+ words averages 5.1 citations versus 3.2 for content under 800 words.
YouTube's share of AIO social citations rose from 18.9% to 39.2% in five months. Upload videos with full transcripts, chapter markers, and VideoObject schema. This is the single biggest AIO-specific tactic that does not apply to ChatGPT.
Category 1 measures are expected in H1 2026 with a 90-day grace period. If Google introduces publisher opt-outs, competitors who leave create opportunity for those who stay. Track announcements from the CMA and prepare to adjust your strategy in either direction.
Use GSC's AI Mode filter for Google-specific data. Create a custom GA4 channel group for AI referrals. Consider Semrush AI Visibility Toolkit or Ahrefs Brand Radar for cross-platform citation monitoring. Compare AIO-triggering queries against non-AIO queries to calculate your "AI Tax" on CTR.
Currently, no granular opt-out exists. Blocking Google-Extended only prevents AI model training, not AIO citations. The nosnippet meta tag blocks AIOs but kills all snippet visibility. The CMA's proposed conduct requirements may force Google to offer a genuine opt-out without ranking penalties in H1 2026.
Position-1 results lose 58% of clicks when an AIO appears above them. However, brands cited within AIOs retain 35% more clicks than non-cited competitors. The UK impact is currently less severe than the US, with AIO frequency plateauing at roughly 12% of non-brand mobile terms.
AI Overviews are automatic summaries above search results, drawing 76–86% of citations from top-10 organic pages. AI Mode is a conversational interface with only 13.7% citation overlap with AIOs and only 14% of cited URLs from the top 10 — making it more similar to ChatGPT. AI Mode has 75 million daily active users.
Yes, specifically for Google's AI features. Pages with FAQ schema are 60% more likely to be featured in AIOs. Article, FAQPage, HowTo, and VideoObject schema all help Google parse and verify content. This is a notable difference from ChatGPT, where schema shows no positive effect on citation rates.
The CMA's proposed conduct requirements could force Google to offer UK publishers a genuine opt-out from AIOs without ranking penalties, provide 30-day notice before algorithm changes, and deliver standardised analytics data export. These measures are expected in H1 2026 and could reshape how AIOs work in the UK market.
If you already rank well in Google organic results, AIO optimisation requires less new work — 76–86% of citations come from pages already in the top 10. ChatGPT requires a more distinct strategy since only 8–12% of its citations overlap with Google rankings. Ideally, address both within an integrated AEO strategy.
Whitehat's AEO audit queries Google AI Overviews, ChatGPT, and Perplexity with your top buyer queries and reports exactly which brands are cited — yours and your competitors'. Includes actionable recommendations and regulatory impact assessment.
Book Your Free AEO AuditNo obligation. Takes 30 minutes. Includes competitor comparison.
This article was researched and written by Whitehat SEO's content team using data from Ahrefs (300K keywords, Feb 2026), Semrush (10M+ keywords, Dec 2025), Seer Interactive (3,119 queries/42 organisations, Nov 2025), BrightEdge (May 2025), Pew Research (68,879 searches, Jul 2025), Goodie AI (6.1M citations, Jan 2026), CMA DMCCA documentation (Oct 2025/Jan 2026), Growth Memo (Feb 2026), Surfer SEO (Nov 2025), SE Ranking (Nov 2025), and additional sources cited throughout. All statistics verified against primary sources at time of publication. Last updated: February 2026.