We studied the shift in search behavior using public trend data and Writesonic’s latest GEO tool to understand the real impact of AI search on SEO traffic—not just what people assume is happening, but what the data actually confirms.
What we found reveals significant changes in how AI traffic flows, how visibility operates, and what kind of content is surfaced.
Here’s what we learned—and what it means for your SEO strategy going forward.
1. Google AI Overviews are pushing organic results way down the page
Ranking #1 on Google doesn’t mean what it used to.
Google’s AI Overviews now take up an average of 1,764 pixels in height at the top of the search results page. That means even if you rank first, your page can be pushed several scrolls down, especially on mobile.
This means that your content might still rank well, but users won’t see it, or even click it, unless they scroll past a full-screen AI answer (and many people don’t)
AI Overviews now often appear before everything—ads, featured snippets, and even local results. They summarize information from multiple sources right at the top, which means fewer people reach traditional links.
What you can do about it:
To stay visible, your content needs to rank in AI Overviews—not just listed in the organic search results.
You can do this by:
- Writing clear, quotable answers
- Using subheadings and concise sections AI can extract from
- Structuring content so it’s useful in pieces, not just as a full article
You can also use Writesonic’s GEO tool to see where your content is being cited across LLMs and Google AI Overviews—even when that traffic doesn’t show up in traditional analytics tools like GA4 or Search Console.
This helps you understand if you’re still influencing users, even without the click. Plus, you can get insights into which industry related AI summaries do not mention you so you can optimize your content for answer engine optimization (AEO).
2. Organic traffic could drop by 18% to 64%, depending on your industry
AI Overviews are replacing clicks, especially in industries where answers are simple and easy to summarize.
Our research indicates that organic traffic is decreasing by 18% to 64%, depending on your subject matter and the target audience.
Here are the industries that are seeing the biggest drops:
| Industry | % of Queries With AI Overviews |
| Healthcare | 76% |
| Technology | 49% |
| Insurance | 45% |
| Education | 44% |
These industries tend to get hit harder by the emergence of AI search because they often answer broad, informational queries—exactly the kind of content large language models are designed to summarize.
Affiliate and publisher sites are especially vulnerable. For example, listicle blogs like “best tools for freelancers” or “top 10 task apps” can now be turned into quick summaries inside AI-generated search results, giving users what they need without clicking through.
If you want to identify your risk level when it comes to AI visibility, check your brand presence across different AI search engines using Writesonic’s AI traffic analytics or GEO tool.
It shows how your content is ranking on ChatGPT and Google’s AI, even if it’s not resulting in a direct visit. Most importantly, you get to see exactly how your brand is being mentioned by AI, whether the sentiment is positive or negative, and which industry competitors are ahead of you in AI search.
3. AI Overviews cite 3–4 sources on average—but only if they’re SERP-visible
As per the latest data pulled from our generative engine optimization (GEO) tool, on average, Google’s AI Overviews pull from 3 to 4 sources per result. But being cited isn’t just about authority or content quality—it’s also about whether your page shows up in the SERP at all.
According to our analysis and citation data:
- 58% of Overviews cite support URLs only when they already appear in the top 10 results.
- That number drops to just 15% if your page isn’t ranking in the SERP, even if it’s relevant.
This means that AI doesn’t always look deeper into SERPs. If your page isn’t visible on page one, chances are high it’s not being cited at all.
Based on our data, these are the kinds of content that are being cited the most:
| Content Type | Citation Likelihood Ratio |
| Government | 11.75 |
| Ecommerce | 5.10 |
| Support Docs | 3.43 |
| News/Media | 2.56 |
| UGC (Reddit, Quora) | 1.42 |
Platforms like Quora and Reddit lead in volume because they answer niche, specific questions that general content misses. Meanwhile, structured documentation—such as help center pages—gets cited heavily if it’s already visible in the SERP.
What this means for your content:
To improve your chances of being cited by AI:
- Focus on structured content: Support pages, FAQs, and comparison guides.
- Use schema markup like Product, FAQPage, or HowTo to help LLMs understand the page’s purpose.
- Ensure crawlability: Pages must be visible, fast-loading, and not hidden behind JS frameworks.
- Check SERP presence before optimizing for AI: Use traditional SEO to secure visibility first, then optimize for citation.
In short, if you want to boost your AI search rankings, traditional SEO is still just as important as it helps you rank in SERPs. And if you’re ranking higher on Google, the more likely you are to be mentioned by AI.
4. 58% of Google searches now result in zero clicks
A striking shift in search behavior has emerged, with recent data revealing that 58% of Google searches now result in zero clicks.
AI Overviews has accelerated zero-click searches by providing comprehensive answers directly in search results. As a result, these AI-generated snapshots deliver immediate information satisfaction without requiring users to navigate away from Google.
Why this matters:
For websites that rely on increasing organic traffic to generate leads and revenue, this zero-click trend represents a serious threat. As search increasingly becomes an endpoint rather than a gateway, businesses face declining visitor numbers regardless of their ranking position.
The implications vary by sector, yet the overall pattern is clear: users increasingly consume information directly within search results rather than clicking through to websites.
Unfortunately, this translates to fewer opportunities for conversion, email signups, and other essential business goals that require site visits.
This means your business website has a real shot at getting cited, only if your content helps users make a decision.
For example, if you’ve published a product comparison or pricing breakdown and structured it clearly, LLM tools are more likely to reference it when summarizing “best tools for X” or “how to choose between A and B.”
AI doesn’t just pull from blogs anymore—it looks for content that answers a question directly and can be chunked into parts.
How to optimize for LLMs:
- Write content that targets specific use cases or buyer questions.
- Include direct answers with clean formatting: headings, bullet points, and short paragraphs.
- Add expert input or original insights—ChatGPT is more likely to cite your page if it says something different from what’s already common.
- Include images, charts, and media when relevant (multimodal models are more likely to surface rich content).
💡Learn more about: Will AI Replace Search Engines? What Marketers Must Know
5. Publishers and ecommerce sites are seeing the steepest drops
While traffic loss from AI Overviews is affecting almost everyone, publishers and ecommerce websites are getting hit the hardest, and in different ways.
Publishers rely on informational content, listicles, and affiliate posts—exactly the type of content AI is built to summarize.
Queries like:
- “Best CRMs for recruiters”
- “Top coffee machines under $200”
These now trigger AI Overviews that answer the question upfront. The AI often pulls from multiple sources (sometimes including the publisher) but rarely gives full credit or a reason for the user to click through.
If your model depends on impressions or affiliate revenue, this shift directly impacts earnings, even when rankings haven’t changed.
Our data also shows that e-commerce sites are also being cited inside AI Overviews—but that doesn’t always translate to traffic.
When ecommerce URLs are present in the SERP:
- AI Overviews cite ~5.5 sources on average (vs. 3.9 when ecommerce pages aren’t visible).
- 20% of citations in these cases are e-commerce links.
So, you’re more likely to be cited if your product pages appear in the SERP (instead of informational content)—but users may not click through if the AI has already answered their question or shown a price comparison.
How both industries can adapt:
Publishers:
- Shift toward content AI can’t easily summarize—original research, expert quotes, opinion pieces.
- Focus on long-tail keywords and specific questions AI struggles with.
- Strengthen E-E-A-T signals to improve citation chances.
Ecommerce:
- Use structured data like schema markups and LLM.txt files to improve chances of being cited (especially Product, Offer, and Review schema).
- Add rich content (video demos, comparisons, FAQs) that LLMs can reference but not replicate.
- Consider getting mentions and citations on other high-authority third-party sites, not just your own product pages to boost your digital footprint and visibility for AI.
6. Authority drives citations: Government and support pages lead by far
If there’s one thing AI Overviews prioritize, it’s authority. Across over 660,000 analyzed search results from our AI search analytics tool, the pages most likely to get cited weren’t always the best-ranked—they were the most trustworthy.
Here’s what our data shows:
| Source Type | Citation Likelihood Ratio |
| Government | 11.75 |
| Support Docs | 3.43 |
| Ecommerce | 5.10 |
| News Sites | 2.56 |
| UGC (e.g., Reddit) | 1.42 |
Government and support URLs dominate citations, even if they’re not ranking in the top 3 positions. That’s because AI systems lean heavily on trusted domains with a history of accuracy, clear structure, and low risk.
This shift shows that Google’s core ranking systems (and by extension, AI Overviews) are heavily tied to Experience, Expertise, Authority, and Trustworthiness (the E-E-A-T framework).
In fact, studies show:
- Content with stronger authority signals improved AI visibility by 89%
- Trust-specific updates led to a 134% citation boost
In other words, LLMs aren’t just looking for relevance. They’re scanning for reputation.
How to make your content more citation-worthy:
- Add author bios and credentials: Especially for YMYL (Your Money Your Life) topics, make it clear who’s behind the content and why they’re credible.
- Use statistics and external citations from well-known, reputable sources.
- Secure backlinks from authoritative domains to validate your trustworthiness.
- Update content regularly: Stale pages are less likely to be cited, even if they were once ranked.
Also, consider building your own support library or help center content. Pages with structured answers to common questions tend to get cited frequently, especially if they already appear somewhere in the SERP.
7. Multimedia and original content are harder for AI to replace
Multimedia-rich content provides a powerful defense against the growing prevalence of AI-generated search results. Google’s multi-modal AI models understand more than just text—they process images, videos, and audio—yet AI still struggles to fully replicate or replace these rich media formats.
Why original content stands out:
AI Overviews are great at pulling widely available information. But when your content includes first-party data, expert opinions, or multimedia elements, it becomes a lot harder to reduce into a single paragraph.
That gives users a reason to click through—and gives AI a reason to cite you as the source, not just a supporting link. As a result, content with authoritative elements improves rankings by 89% in generative AI search environments. In short, content backed by unique data becomes irreplaceable in the information ecosystem.
For example:
- Proprietary research and surveys often get directly quoted because no other site has the same data.
- Expert interviews or detailed case studies add perspectives AI can’t easily replicate.
- Videos and images (especially ones not found elsewhere) increase the odds of being featured in image or video carousels—places where AI Overviews don’t yet dominate.
Video and visual content deserve special attention as a defense mechanism against automated AI search results.
To maximize video effectiveness:
- Target specific search queries with keyword-rich titles and detailed descriptions
- Include timestamps that help both users and AI understand your content structure
- Integrate videos within existing web pages rather than relying on standalone video content
Custom images that appear in organic results below AI snapshots can compel users to visit your site. Your multimedia can also appear in Google Image or Video search, which currently doesn’t feature AI Overviews.
The key is creating value that AI cannot replicate—whether through original research, expert insights, or multimedia experiences that require users to visit your site for the full context.
💡Also read about: The Impact of AI Engines on SEO
8. AI Overviews appear in 47% of all search queries
AI search results are no longer a niche concept. They’re showing up in nearly half of all Google searches—and growing. Across over 660,000 SERPs analyzed, AI Overviews appeared in 47% of Google queries. And that number is higher in certain industries and query types.
AI Overviews are far more likely to appear in informational searches, especially in sensitive or regulated categories. Here’s a breakdown:
| Industry | AI Overview Coverage |
| Healthcare | 76% of queries |
| Technology | 49% |
| Insurance | 45% |
| Education | 44% |
AI summaries are also more likely to trigger for:
- Question-style queries (e.g., “how to choose the right CRM”)
- Long-tail keywords (e.g., “best software for small recruitment agencies”)
- Mid-tail, category-level topics (e.g., “best productivity apps”)
This means, the more general or informational your content is, the more likely it’s being summarized—possibly without credit.
But AI Overviews still struggle with:
- Highly nuanced or context-heavy queries
- Topics that require subjective judgment
- Pages that offer step-by-step visuals or rich interaction
Which means, these topics and niches are your opportunity.
If you rely on SEO for discovery, don’t just optimize to rank—optimize to be useful in a way AI can’t replicate.
9. Citation chances drop 4x when you’re not on page one
Our data analysis of 660,876 search results reveals a striking disparity: support URLs appear in AI Overviews four times more frequently when they rank in the top 10 positions.
And only 15% of AI Overviews cite support URLs when they don’t appear in top search results, compared to 60% when they do.
So even though AI tools technically have access to the full index, they still behave like users: they trust what’s already surfaced by the algorithm.
| SERP Position | % of Citations |
| Position 1 | 18.5% |
| Position 2 | 15.6% |
| Position 3 | 13.5% |
| Position 9 | 4.7% |
If you’re not on page one, your chances of getting cited fall by nearly 75%—even if your content is relevant.
What this means for your SEO strategy:
This isn’t a case of SEO vs. AI. It’s combining SEO with AI search. Traditional SEO ranking factors still matter because:
- You need to rank in SERPs to be eligible for AI citations (not always, but it definitely helps!).
- Structured content and technical SEO can increase both your visibility and extractability.
- Schema markup remains critical for product, review, FAQ, and support content.
If you’ve deprioritized SEO ranking in favor of AI optimization alone, it’s time to revisit that balance. The key is to integrate GEO with SEO.
10. High domain authority wins citations, and long-tail queries trigger AI the most
We looked at the top 200 most-cited domains in AI Overviews and grouped them by Domain Rating (DR). The difference was huge.
- Sites with a DR of 88–100 received more than 6,000 average citations
- Sites in the 63–88 range got far fewer
- Domains under 63 were barely cited at all
AI search tools clearly favor trusted sources—government websites, help docs, and major publishers top the list. Even if a low-DR site ranks on page one, it’s often skipped when AI decides what to cite.
In short: If your domain lacks authority, you’re probably invisible to AI—even if you’re visible to users. This means traditional SEO efforts that include boosting DR through link-building are still effective, even for AI visibility.
Our data also revealed that AI search models prefer specific, question-based searches.
AI Overviews are most likely to appear for:
| Keyword Type | % of AI Overview Triggers |
| Long-tail queries | 33% |
| Question keywords | 22.6% |
| Short-tail | 9.9% |
| Single-word terms | 4.4% |
That means vague, broad terms like “best running shoes” rarely trigger Overviews. But specific, intent-heavy searches like, “best budget-friendly running shoes for teens” or “how to automate candidate follow-ups” are far more likely to surface AI responses—and that’s where your brand has a shot at being cited.
This means you need both authority and search intent alignment to appear in AI search results.
Here’s how to get there:
- Build your DR through high-quality backlinks and partnerships
- Target long-tail, question-based keywords that AI is already summarizing
- Make your answers clear, quotable, and easy to extract
- Target niche and specific search queries and topics that align with your industry
Ranking isn’t enough anymore. You have to be the kind of site—and the kind of content—AI wants to quote.
SEO isn’t dead, but it’s no longer enough on its own
AI is reshaping how people search—but that doesn’t mean SEO is irrelevant. It means SEO needs backup.
The fundamentals of SEO still matter:
- You still need indexable, optimized content that ranks
- You still need to earn authority and backlinks
- You still need to meet search intent and structure your content well
But that’s just one half of the equation now.
The other half is AI visibility—being cited in Overviews, responses, and zero-click summaries. That’s where strategies like AEO (AI Engine Optimization) come in. AEO focuses on how LLMs read, understand, and select content. It’s not just about being found—it’s about being featured.
To compete today, you need both:
- SEO to get into the SERP
- AEO to get into the AI answers
Writesonic helps with both through:
- Intuitive SEO tools and AI agents help you automate SEO workflows like keyword research and creating optimized, structured content that performs in traditional search.
- GEO (Generative Engine Optimization) system shows where your content appears in AI platforms like Google AI Overviews, ChatGPT, and Perplexity.
- Visibility into what your competitors are doing in terms of SEO and AI visibility, so you can track how your brand shows up as per industry standards.
AI isn’t replacing SEO. It’s forcing it to evolve.
If you’re not adapting your strategy for how search actually works in 2025, you’re not just losing rankings—you’re losing relevance.
👉 Try Writesonic’s SEO + GEO tools and stay ahead of the curve.
