What AI overviews means for SEO in 2025

12 min readJuly 9, 2026By Spawned Team

AI Overviews now appear in ~47% of Google searches. Here's what that actually means for your traffic, rankings, and SEO strategy going forward.

Marketing professional reviewing SEO and AI search performance charts at a desk near a window at dusk

TL;DR: Google's AI Overviews now appear in roughly 47% of searches and send measurably less organic click traffic to the pages they summarize. SEO isn't dead, but the game has shifted: getting cited inside an AI Overview matters more than holding the #1 blue link. Brands that structure content for AI extraction and earn topical authority will win. Those who don't will watch impressions rise while clicks fall.

What are Google AI Overviews and how did we get here?

Google AI Overviews (formerly Search Generative Experience, or SGE) are the AI-generated answer blocks that appear above traditional organic results for a large share of queries. Google launched them to the full U.S. user base in May 2024 and rolled them out internationally through late 2024 and 2025 [1].

The mechanics are straightforward. Google's Gemini model reads a broad pool of documents, writes an answer, and shows it with attribution links to the side of the block. Users get their answer without scrolling, and the cited sources get a small carousel link instead of a traditional blue-link position.

This isn't Google's first answer-box feature. Featured snippets, Knowledge Panels, and People Also Ask boxes all pulled intent away from organic clicks before AI Overviews existed. But AI Overviews are different in two ways. First, they're generative, so they can answer compound questions that no single page ever ranked for. Second, they appear at a scale that dwarfs everything before them. AI-powered search features were always the direction of travel. AI Overviews are the destination.

Know what triggers one. Early research by SE Ranking found that as of early 2025, informational queries trigger AI Overviews far more than transactional ones, and long-tail queries see them at higher rates than short-head terms [2]. That matters for strategy. If your traffic skews informational, you're already in the deep end.

How does AI affect SEO traffic, and what does the click-loss data actually say?

Everyone wants a clean number, and the data is messier than the headlines suggest. The most-cited study, from Semrush's data team in early 2025, found that pages appearing in an AI Overview but not in the top organic positions received roughly 34% fewer clicks than comparable pages on queries with no AI Overview present [3]. Pages that appeared both in the AI Overview citations and in the top-3 organic positions held their traffic better, which makes sense.

Gartner had earlier forecast that organic search volume would fall 25% by 2026 as AI interfaces spread [4]. Nobody has good data on whether that trajectory is tracking to plan. The closest independent evidence says the decline is real but uneven across industries.

The pattern that holds across most data sets: click loss is worst where Google's model can confidently answer from existing trusted sources. Medical, financial, how-to, and definition queries get hit hardest. Local, product comparison, and navigational queries are less affected, at least for now.

Here's the counterintuitive part. Google's own launch blog claimed that links inside AI Overviews get more clicks than the same links in traditional featured snippets [1]. That claim hasn't been independently verified at scale. But it does suggest citation placement inside the Overview block carries real value, even as the query-level click pool shrinks.

| Query type | AI Overview appearance rate (SE Ranking, 2025) | Estimated organic CTR impact | |---|---|---| | Informational | ~64% | High negative | | Navigational | ~8% | Low | | Transactional | ~18% | Moderate | | Local | ~12% | Low to moderate |

Source: SE Ranking AI Overview study, 2025 [2]

How will AI affect SEO rankings, and does position still matter?

Position still matters, but ranking and traffic have decoupled. A page can sit at #1 and lose 30% of its former clicks if an AI Overview answers the query above it. A page ranked #6 that earns an AI Overview citation can capture more qualified attention than its organic position ever delivered.

Citation sourcing is where things get interesting. Studies tracking which pages Google cites inside AI Overviews consistently find that the top 10 organic results supply most citation sources, with pages in positions 1-3 appearing in the citation carousels at the highest rate [3]. Traditional ranking is still the price of admission to the citation pool. You can't skip ranking and go straight to being quoted.

What changed is the objective. The old game was rank at the top and capture 30% CTR. The new game is rank well enough to enter the citation pool, then make your content structured enough that Gemini prefers to quote you over a competitor. Related goals, not identical ones.

AI SEO as a discipline is forming around exactly this tension. Tactics that improve traditional ranking (E-E-A-T signals, backlinks, page speed) still apply. But you now also have to think about how your content gets extracted and summarized, which is a different job. More on that below.

One honest caveat. The ranking-to-citation relationship varies by industry, and nobody has published a broad cross-vertical study. The pattern above is the best available generalization, not a law of physics.

AI Overview trigger rate by query type

| | | |---|---| | Informational | 64% | | Transactional | 18% | | Local | 12% | | Navigational | 8% |

Source: SE Ranking, AI Overviews study, 2025

What is AI search market share, and how fast is this changing?

Google still processes roughly 8.5 billion searches per day, which makes it the dominant surface for AI Overview exposure by volume [5]. But the broader AI search landscape now includes ChatGPT (which opened its native web search to all users in late 2024), Perplexity AI, Microsoft Copilot (inside Bing), and Meta AI across WhatsApp, Instagram, and Facebook.

Perplexity reported 100 million weekly active users in January 2025 [6]. ChatGPT's search feature reached over 1 billion searches per week in early 2025, by OpenAI's own reporting [7]. Those numbers look large until you set them next to Google's scale, but the growth rates matter more than the current base. Monthly active users for AI-native search tools roughly tripled between mid-2024 and mid-2025 by most estimates.

For most brands, Google AI Overviews are the immediate priority because that's where your existing organic traffic lives. But if you sell in B2B tech, finance, or any category where early adopters lead buying decisions, the share of your audience using ChatGPT or Perplexity to research vendors is probably higher than your analytics show. AI assistants don't reliably pass UTM parameters.

This is why generative engine optimization has become its own field alongside traditional SEO. The signals that get you cited in ChatGPT or Claude overlap heavily with what works for Google AI Overviews, but they aren't identical. Treating them as separate visibility channels is becoming a practical necessity.

Which types of content are most hurt by AI Overviews?

Simple informational content is the clearest casualty. If your page exists mainly to answer a definition, explain a concept, or serve a numbered list any LLM can reproduce, AI Overviews will absorb that query and your traffic will fall. This isn't a prediction. It's been observed in site-level data across dozens of publisher case studies since mid-2024.

The content categories most at risk:

Glossary and definition pages
Basic how-to content that doesn't require hands-on expertise
News summaries (as opposed to original reporting)
Generic comparison pages with no original data or first-hand testing
FAQ pages that answer questions AI models were trained on

Content categories that hold up better:

Original research with proprietary data
First-person experience accounts (product reviews from actual use, field reports)
Local and hyperlocal content that requires physical proximity
Highly technical content with specificity that general models struggle to reproduce accurately
Content that changes fast enough that model training data is already stale

Google's own documentation on helpful content [8] has long emphasized experience, expertise, authoritativeness, and trustworthiness. AI Overviews speed up the reckoning for content that faked those signals. Pages built to capture SEO traffic rather than genuinely help someone are the first to get abstracted away.

How do you get cited in AI Overviews, and what does optimization actually look like?

Getting cited in AI Overviews is partly traditional SEO (you need to rank and hold topical authority) and partly structural work that makes your content easy for a language model to extract with confidence.

The structural side breaks into a few practical moves.

Answer questions directly at the top of sections. AI models grab the first substantive answer to a question, then look for supporting detail. If your page buries the answer in paragraph four, you lose to a page that leads with it. A 40 to 60 word direct answer at the start of each section, before any caveats or background, sharply improves extraction quality.

Use explicit question-and-answer structure. H2s phrased as real questions people ask, with the answer in the first paragraph, mirror exactly how retrieval-augmented generation (RAG) systems pull content. This is not accidental.

Add schema markup. FAQ schema, HowTo schema, and Article schema give Google structured signals about what your page contains. They've existed for years, but their weight for AI extraction is higher now because structured data reduces ambiguity for the model.

Build topical authority over keyword coverage. AI models prefer sources they've seen cited consistently across many related queries. A single optimized page rarely earns sustained citations. A cluster of related, deeply linked content covering a whole topic space tells the model your site is a reliable primary source.

Cite your own original data. A concrete number, study result, or proprietary finding that lives nowhere else is one of the strongest citation magnets available. If you ran a survey, published a dataset, or tracked something over time, make that data findable, citable, and shareable.

Tools like Brandrank.ai visibility insights analysis and platforms built for AI search visibility metrics and KPIs can show you where you appear (or don't) across AI engines. That's the starting point for any optimization effort. You can't improve what you're not measuring.

Does Google AI Overviews hurt small sites and favor big brands?

Honest answer: probably yes, on average, but not universally. The citation patterns in 2024-2025 research show that high-authority domains appear in AI Overview citations at rates well above their share of the organic top 10 [3]. If Gemini has ingested thousands of positive signals about a domain over many years, it's more likely to trust and cite that domain.

Niche expertise is a real counterweight. A site that owns a specific topic so thoroughly that it's the definitive source, even with modest domain authority by broad-web standards, can earn steady AI citation in that space. The pattern that hurts small sites most is general-purpose content at medium quality, which is exactly what models can reproduce without crediting anyone.

The sites most at risk sit in the middle. Not authoritative enough to be a default citation, not specialized enough to be the only credible source. Thin affiliate sites, generic informational blogs, and content farms are being structurally marginalized. Some of that was already happening through algorithm updates. AI Overviews just moved up the timeline.

Small brands that do original work, speak from real expertise, and cover one domain with genuine depth have better prospects than their domain authority scores would predict. The investment is in substance, not SEO tricks.

How do you measure SEO performance now that AI Overviews exist?

Google Search Console still has no dedicated AI Overview filter as of mid-2025 [9]. That's a real gap. You can see impression and click data, but you can't natively attribute traffic loss to an AI Overview appearing on a specific query.

Here's what works today.

Segment your keyword tracking by query type. Informational queries that trigger AI Overviews will show impression growth alongside CTR decline. That divergence is the signal. If impressions for a query are rising while CTR drops below 2%, an AI Overview is almost certainly present.

Use third-party SERP monitoring tools that track AI Overview presence by keyword. Several SEO platforms added this in late 2024 and 2025. SE Ranking, Semrush, and BrightEdge all have some form of AI Overview detection built into their rank trackers [2][3].

Track citation presence separately from organic rank. For your most important queries, you want to know two things: does an AI Overview appear, and if so, are you cited in it? These are different metrics from position 1-3, and they need different tooling. AI visibility tools built for this are now available and worth evaluating if you manage significant organic traffic.

Monitor zero-click rate by landing page category. If a group of pages generates searches but not visits, that's where AI Overviews are absorbing the intent.

Spawned's AI visibility audit is one option for brands that want a structured snapshot of where they stand across both Google AI Overviews and the wider AI search ecosystem, before deciding where to spend effort.

The underlying principle: measure visibility in AI surfaces separately from traditional organic performance. Blend them into one metric and you hide what's actually happening to your traffic.

What SEO strategies actually work now, given AI Overviews?

The strategies that work are mostly the ones that always should have worked, with the execution bar raised hard.

Build content around questions, not keywords. The move from keyword optimization to question answering has been a trend for years, but AI Overviews make it a hard requirement. A page optimized for the keyword "project management software" gets steamrolled by an AI Overview. A page that authoritatively answers "what project management software works best for a remote team of 10" has a shot at being cited, because it answers a specific question with specific depth.

Invest in original data and research. Surveys, proprietary datasets, original analysis, and primary reporting are the content types AI models can't reproduce and have to attribute. One well-designed annual survey that produces citable statistics is worth more for AI visibility than fifty generic blog posts.

Build author authority. Google's E-E-A-T framework explicitly weighs author credentials and track record [8]. Bylines tied to real professional profiles, demonstrable expertise in the subject, and consistent publishing under a named identity all feed into whether your content gets trusted as a citation source.

Focus on transactional and comparison content. These query types see lower AI Overview rates. Users searching to buy, compare, or evaluate something still reach traditional organic results, at least for now. If your content mix leans heavily informational, rebalancing toward middle and bottom-of-funnel content cuts your AI Overview exposure.

Get mentioned on other authoritative sites. AI models are trained on the web. If your brand or content is cited, linked, mentioned, and discussed by authoritative sources in your space, you build the off-page signals that make the model treat you as a reliable citation. Digital PR, expert contributions to industry publications, and genuine thought leadership all matter here.

Don't abandon SEO. This one is important. Some of the coverage around AI Overviews implies traditional SEO is obsolete. It isn't. Ranking in the top 10 is still the prerequisite for entering the AI Overview citation pool. You can't optimize for AI citation without optimizing for organic ranking. The work is additive, not a substitute. AI SEO tools that layer AI visibility metrics on top of traditional rank tracking are the practical way to manage both.

What does this mean for Google's business model, and will AI Overviews stay?

Google's revenue depends on advertising, specifically on users clicking paid search ads. AI Overviews that reduce click-through to organic results don't directly hurt ad revenue if users still click ads. Google has been careful to keep paid results visually distinct from and above AI Overviews in most layouts.

The deeper risk to Google is query volume migration. If a growing share of search intent moves to ChatGPT, Perplexity, or other AI-native tools, that's where Google's business model faces real pressure. The May 2024 launch was partly defensive. Google needed to show its search product could give AI-generated answers so users wouldn't go elsewhere [1].

AI Overviews are not going away. Google has too much at stake competitively to pull the feature, and it keeps expanding coverage, adding UI treatments like follow-up questions, and folding more Gemini capabilities into search. The direction is clearly toward AI answers as the default for informational queries.

For Google AI search specifically, the implication is that the SERP of 2027 will look more like a chat interface with optional source expansion than a list of ten blue links. That trajectory is visible now. Planning for it isn't premature.

How should brands think about AI visibility as a separate metric from SEO?

The cleanest mental model: SEO measures whether you appear in the results. AI visibility measures whether you're recommended by the model. Related, not the same.

A brand can have perfect SEO, ranking #1 for its most important terms, and still be effectively invisible in AI-generated answers if the model doesn't trust or recognize it as authoritative. The reverse also happens, though less often: a brand with modest traditional ranking becomes the default citation source in its niche because it produces distinctive, well-structured content.

The metrics you care about for AI visibility: citation rate in AI Overviews (how often you appear in the source carousel), mention rate in AI assistant responses (how often ChatGPT, Perplexity, or Claude name your brand when someone asks a category-level question), and sentiment in those mentions (positive, neutral, or negative framing). None of these show up in Google Search Console today.

Building AI visibility takes the same foundational work as SEO but with extra emphasis on off-page citation signals, structured content extraction, and brand mentions across the web that train models to link your name to your category. It's a longer-term investment with compounding returns, which is why starting now matters.

Brands that treat AI visibility as a real measurement discipline in 2025 will hold a structural advantage by 2027, when AI-generated answers are the primary surface for most informational queries. The window to build that position ahead of competitors is real, and it's open now.

Sources

Frequently Asked Questions

How will AI affect SEO in the next few years?

The trajectory is clear: AI-generated answers will absorb more informational query traffic, cutting clicks to organic results for those query types. Traditional ranking signals still matter because they decide which pages enter the AI citation pool. SEO is evolving to include AI visibility optimization alongside classic rank tracking. Brands that build topical authority and original content now will be better positioned as AI surfaces expand across Google, Bing, and AI-native tools.

Does getting cited in an AI Overview actually drive traffic?

Yes, but less than a top organic position used to deliver. Google's own launch documentation claimed citation links in AI Overviews outperform equivalent featured-snippet links, though that hasn't been independently verified at scale. The value of an AI Overview citation is part direct traffic, part brand exposure. Users who see your brand cited as a trusted source form a positive impression even if they don't click through right away.

Which queries are least affected by AI Overviews?

Transactional queries (ready to buy), local queries (nearby results, maps, hours), navigational queries (searching for a specific brand or site), and comparison queries for high-consideration purchases all see lower AI Overview rates. Informational and how-to queries are most affected. If your site's traffic is heavily transactional or local, AI Overview impact is lower, though this may shift as Google expands AI answer coverage.

What is E-E-A-T and does it still matter for AI Overviews?

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. It's Google's framework for evaluating content quality. It still matters, arguably more than before, because AI models use authoritativeness signals to decide which sources to cite. Named authors with demonstrable credentials, citations from respected publications, and a clean track record of accuracy all feed E-E-A-T and AI citation likelihood.

Can small websites compete for AI Overview citations?

Yes, in specific niches. AI Overview citation patterns favor high-domain-authority sites broadly, but niche expertise is a real counterweight. A small site that is the definitive source on a narrow topic, with original research and consistently structured content, can earn regular AI citation in that space. Generic content at medium quality on any topic is much harder to place. Depth and specificity beat breadth and volume for small publishers.

Should I stop creating informational blog content because of AI Overviews?

Not entirely, but raise the bar significantly. Generic informational content that AI can reproduce without citing you is worth less than it used to be. Original data, first-person experience, specific technical depth, or a distinctive point of view all make informational content citation-worthy. Before publishing, ask: is there anything here a language model couldn't generate itself? If the answer is no, that's your quality gap.

How do I check if AI Overviews are appearing for my target keywords?

Search the keywords manually in an incognito window from a U.S. browser. Several rank tracking tools, including SE Ranking and Semrush, now flag AI Overview presence in their keyword data. Google Search Console doesn't yet provide a direct AI Overview filter. Third-party AI visibility platforms can automate this monitoring across large keyword sets and alert you to new AI Overview appearances on important queries.

Does schema markup help with AI Overview citations?

Yes, with caveats. FAQ schema, HowTo schema, and Article schema give Google explicit structured signals about your content's structure and intent. These reduce ambiguity for the model when it decides whether to extract and cite a section. Schema isn't sufficient on its own: the underlying content still needs to meet quality and authority thresholds. But on two comparable pages, the one with proper schema is more likely to be cited.

What is generative engine optimization (GEO) and how is it different from SEO?

Generative engine optimization (GEO) is the practice of optimizing content to be cited and recommended by AI-generated answer engines: Google AI Overviews, ChatGPT, Perplexity, and similar systems. It shares foundational practices with SEO (authority signals, quality content, structured markup) but adds specific techniques for AI extraction, including direct question-answer structure, citation-worthy original data, and off-page mention building across AI training surfaces.

Is zero-click search a new problem created by AI Overviews?

Zero-click search, where users get their answer on the SERP without visiting a site, predates AI Overviews. Featured snippets, Knowledge Panels, and direct answer boxes were already cutting click rates for informational queries. AI Overviews extend this pattern significantly in scope and quality of answers, speeding up a trend that was already underway. SparkToro and others have tracked Google's zero-click rate at over 50% of searches for several years.

How often do AI Overview citations change for the same query?

Frequently. AI Overview citations aren't static rankings: they can change based on query phrasing variations, model updates, and changes to the cited pages themselves. Research tracking AI Overview citations over time shows meaningful week-to-week variation in which specific pages appear in the citation carousel, even when the top organic results stay stable. This is why monitoring citation presence continuously beats a single snapshot.

Do paid ads protect traffic from AI Overview click loss?

Partly. Google keeps paid search results above and visually separate from AI Overviews, so brands running paid search on their key terms keep visibility even when AI Overviews absorb organic clicks. The risk is cost. If organic traffic falls and you compensate with paid, your customer acquisition cost rises. Paid search is a tactical buffer, not a long-term substitute for building AI citation authority.

What's the best way to track brand mentions in AI assistant responses?

Native analytics tools don't capture this. Specialized AI visibility platforms query ChatGPT, Perplexity, Gemini, and similar systems with category-relevant prompts and track how often your brand is mentioned, in what context, and with what sentiment. This is currently the only reliable method. Some brands run manual prompt testing on a regular schedule as a lower-tech alternative. This measurement gap is one of the main reasons AI visibility has emerged as a distinct discipline.

Is it possible to be removed from AI Overview citations if Google has inaccurate information?

Yes. Google provides a feedback mechanism on AI Overview responses and has update pathways for factual corrections. If your brand is misrepresented in an AI-generated summary, submitting feedback through Google's standard mechanisms is the starting point. Updating the source content on your own site to be accurate and unambiguous also reduces the chance the model extracts incorrect information. There's no guaranteed removal or correction timeline.

what ai overviews means for

Ready to try it?

Build your first app in a few minutes.

Start Building