Google AI Overviews (AIOs) now appear at the top of results for the majority of high-volume search queries in the UK and US. They summarise the answer before anyone clicks through, and they cite a small number of sources directly.
If your brand is one of those sources, you earn the click. If you're not, you've lost the query, even if you rank in position one below it.
When Google AI Overviews appear, overall organic CTR drops 61% and paid CTR drops 68%. Brands cited inside the AIO earn 35% more clicks than those ranked below it. (Seer Interactive / Ahrefs, 2025-26)
Here's what determines whether you get cited, and the exact steps to improve your chances.
What we've seen on Google AIO across UK brands
Across the 75+ UK brands we have scored in 2026, AIO presence correlates more strongly with off-site citation density (Wikipedia, regulatory pages, trade press, review sites) than with Google authority. In our April 2026 pensions batch, NEST appeared in AI Overviews on every prompt we tested. NEST also has the strongest citation graph in the category (state-backed, Wikipedia article live, Pensions Regulator references, broad trade press). Brands with strong owned-content but thinner off-site presence had Google authority but weaker AIO presence. The signal that gets you cited in an AI Overview is not where your page ranks. It is what other trusted sources say about you.
What Are Google AI Overviews and How Do They Work?
RAG architecture. Google AI Overviews combine a large language model with Google's search index using a process called Retrieval-Augmented Generation (RAG). The model retrieves relevant passages from the web and composes a summary. This means the algorithm rewards fact-rich, structured content rather than keyword-stuffed pages.
They behave differently from other AI platforms. We analysed thousands of queries across six AI platforms at SearchIntel. AI Overviews are the shortest responses by 3x, a median of just 83 words compared to 422 for ChatGPT, 241 for Claude and 679 for Gemini. They're compressed summaries, not conversations.
| Dimension | Google AI Overviews | ChatGPT | Claude | Gemini |
|---|---|---|---|---|
| Median length | 83 words | Mid-length | Brief | Longest |
| Brand mention rate | Lowest of the platforms tested | High | Moderate | Highest |
| Willingness to recommend | 0.2% | 8.9% | 7.3% | 7.3% |
| Tone | Encyclopaedic | Assistant | Concise assistant | Verbose |
AI Overviews are the hardest platform to get cited on. AI Overview responses mention brands much less often than the LLM platforms tested, and brands tend to appear later in the response. Getting cited here requires high authority.
They appear on 79% of queries, but not always with substance. In our dataset, AI Overviews were present on 79% of queries tested. However, extractable text was only available in 39% of cases, meaning AI Overviews are common, but not always substantial enough to include brand citations.
How Google AI Overviews Choose Sources
Google AI Overviews are built on top of the standard Google index. They don't crawl the web separately, they draw from the same content Google already knows about, then apply an additional layer of selection.
In March 2026, Ahrefs analysed 863,000 search results and 4 million AI Overview URLs and found that only around 38% of AI Overview citations come from pages in Google's top 10 organic results, down from 76% in July 2025. The rest are drawn almost evenly from positions 11-100 and from pages beyond position 100. A page-one ranking still helps, but it is no longer a prerequisite for being cited.
That makes passage selection the deciding factor. AIOs choose the specific passages that most directly answer the query, wherever they rank. A page that ranks for a broad keyword but doesn't contain a clear, extractable answer gets skipped. A lower-ranked page, or one outside the top 10 entirely, with a precise, well-formatted answer often gets chosen instead.
The additional selection factors:
- Schema markup, Pages with structured data are 3x more likely to earn AIO citations. FAQ schema, HowTo schema and Article schema all help Google extract answers accurately.
- Direct answer format, AIOs open with a direct fact or answer 99.8% of the time. Content that leads with the answer (not a preamble) is far more likely to be extracted.
- E-E-A-T signals, Author credentials, publication date, outbound links to authoritative sources, and references to original research all signal trustworthiness.
- Content freshness, Recently published or updated content gets preference for time-sensitive queries.
- Page load speed and crawlability, AIO crawlers don't render JavaScript. Clean HTML, fast load times and proper sitemap submission are prerequisites.
The 6-Step Process to Get Cited
Step 1: Identify which queries trigger AIOs for your category
Not all queries show AIOs. Search your key terms in Chrome (UK or US) and note which ones trigger AI Overviews at the top. These are your priority targets. Focus your optimisation effort on queries that already show AIOs, you can't earn a citation on a query type that doesn't have one.
Step 2: Check if you already rank top 10 for those queries
If you're not in the top 10, you need standard SEO work first, this is the gating factor. If you are in the top 10 but not being cited, the problem is content format, not ranking. Check Google Search Console for your current positions.
Step 3: Reformat existing pages to lead with direct answers
For every query you want to win, your page needs a section that directly answers the question in 2-4 sentences, immediately after the heading. No preamble. No "Great question, let's explore..." The answer first, then the detail. This is the format AIOs extract from.
Step 4: Add FAQ schema to every content page
FAQ schema markup signals to Google that your page directly answers questions. Add 3-5 Q&As in FAQ schema format to every blog post and pillar page. Use the exact phrasing people search for as the question text. This is one of the highest-ROI technical changes for AIO visibility.
Step 5: Build E-E-A-T signals throughout the page
Include: author name with credentials, publication and update dates, outbound links to studies or data sources, references to original research or proprietary data. AIOs favour sources that demonstrate genuine expertise, not just content that covers the topic superficially.
Step 6: Monitor your AIO citation rate over time
Google Search Console shows AIO impressions and clicks. Track which of your pages earn AIO citations and which don't, then analyse what's different between them. Pages that earn citations show you what's working. Pages that rank but don't earn citations show you where to improve format and schema.
Target Comparison Keywords for Maximum Impact
Our data shows that Google AI Overviews use noticeably more superlative language ("best", "top", "leading") on comparison queries vs discovery queries. Other AI platforms hedge instead, but when someone asks "best X", AI Overviews actually pick sides.
This makes "best X" and "top X" queries the highest-value targets for AIO citation. Create content that directly answers comparison questions with clear, defensible rankings or recommendations backed by data.
Pitfalls to Avoid
Beyond the technical exclusions below, watch out for these content-quality traps:
- Thin content, Pages that only summarise others' work or lack original insights will not be cited. Invest in original research and unique perspectives.
- Over-optimisation, Don't spam keywords or artificially insert facts. AI models penalise unnatural writing.
- Ignoring other platforms, AI Overviews are the hardest platform to get cited on, with brand mention rates well below the LLM platforms. A comprehensive AI search visibility strategy covers all platforms, not just Google.
What Google AI Overviews Won't Cite
Understanding the exclusions helps focus your effort:
- Pages blocked from crawling, Any pages with noindex tags, disallow in robots.txt, or behind login walls are invisible to AIOs.
- JavaScript-rendered content, If your key content only appears after JavaScript executes, AIO crawlers likely can't read it. Move critical content to server-rendered HTML.
- Paywalled content, AIOs don't cite subscription content. If your best research is gated, it won't appear.
- Highly promotional content, AIOs avoid content that reads as advertising. Factual, informative content gets cited; product pitches don't.
- Low-authority domains, AIOs skew heavily towards established domains with strong trust signals. New sites building authority from scratch need time before this channel becomes available.
The Content Types That Earn AIO Citations Most Often
Based on our analysis of which content gets cited across 5,600+ AI queries:
| Content Type | AIO Citation Rate | Why |
|---|---|---|
| FAQ pages with schema | High | AIOs extract Q&A format directly; schema makes structure machine-readable |
| How-to guides | High | Procedural queries ("how to...") trigger AIOs frequently; step-by-step format extracts cleanly |
| Definition / explainer pages | High | "What is X" queries almost always trigger AIOs; clear definitions at the top of the page get cited |
| Data and research pages | Medium-High | Original statistics and data are cited as authoritative sources; proprietary data has no competition |
| Product pages | Low | Too promotional; AIOs prefer informational sources |
| News articles (recent) | Medium | High freshness scores for time-sensitive queries; decays quickly |
Frequently Asked Questions
How do Google AI Overviews choose which sources to cite?
Ranking helps, but it is no longer the deciding factor. As of March 2026, only around 38% of AI Overview citations come from pages in Google's top 10, down from 76% a year earlier (Ahrefs), with the rest drawn from positions 11-100 and beyond. What matters most is passage selection: AIOs favour pages with structured data and schema markup, direct question-answer formatting, E-E-A-T signals, original data, and comprehensive topic coverage. Schema markup gives a 3x citation advantage over comparable unstructured pages.
Does ranking on Google guarantee appearing in AI Overviews?
No. A strong organic ranking improves your chances but is neither required nor sufficient. Google AI Overviews select specific passages from pages that best answer the query, so even a top-ranking page needs to directly address the question with a clear, extractable answer. Pages that rank for broad terms but don't answer specific questions clearly are often skipped.
What is the impact of appearing in Google AI Overviews on website traffic?
Brands cited as sources inside Google AI Overviews earn 35% more organic clicks than brands ranked below the AIO but not cited. When AI Overviews appear, overall organic CTR drops 61%, meaning the traffic increasingly flows to cited sources, not to the wider results page.
How long does it take to start appearing in Google AI Overviews?
If your pages already rank in Google's top 10, adding schema markup and improving answer-format content can show results within 2-4 weeks as Google recrawls. If you need to build rankings from scratch, expect 3-6 months for competitive queries. The fastest wins come from improving structured data on pages that already rank.