GEO, LLMO, AEO, AI SEO—whatever you want to call it—is a fast-growing field. And the truth is, no one has fully cracked the code yet.
Marketers, SEOs, and content teams are doing lots of experiments and studies on the topic, and it can be hard to keep up. That’s why I’ve pulled everything together in this article, so you can track the latest findings, news, and trends on AI-driven search.
I will update this guide every month with fresh insights to help you stay ahead of the curve. Feel free to bookmark this page so you can always come back later.
Key Takeaways
- AI isn’t replacing Google; it’s expanding search. The average Google Search usage went up to 12.6 sessions per week after people began using ChatGPT.
- Brands are 6.5x more likely to be cited through third-party sources than their own domains.
- Clicks in AI Mode are reserved for transactions. 75% of AI Mode sessions end without external visits.
- Wikipedia and Reddit are among the most frequently cited domains in AI Overviews, AI Mode, and ChatGPT.
- Organic CTR has dropped 61% for queries where an AI Overview is present. But when your brand is cited in the AI Overview, organic CTR is 35% higher.
- Websites with more organic traffic tend to get more mentions in AI Overviews and Perplexity. But that’s not the case with AI Mode and ChatGPT.
- Content depth, readability, and freshness matter more than traditional SEO metrics like traffic and backlinks when it comes to securing AI mentions and citations.
- Structured content (headings, lists, FAQ, etc) is the most effective format in AI search.
- 44.2% of all LLM citations come from the first 30% of text (the intro).
- Bottom funnel content like case studies and pricing pages get the highest AI referral traffic, while top funnel content (what is and how-to guides) saw massive drops in the past two years.
- Major news publishers like BBC, The New York Times, and The Guardian are taking a stand against AI, blocking AI crawlers from using their content for training data.
- AI recommendations are highly inconsistent: there’s a <1 in 100 chance that ChatGPT or Google’s AI, if asked 100X, will give you the same list of brands in any two responses.
AI Search Engine Usage Statistics
AI chatbots like ChatGPT, Perplexity, and Gemini are becoming popular search channel alternatives, especially for ultra-long and super-specific queries.
But interestingly, according to recent data, people haven’t abandoned traditional search engines just yet.
Here are AI search adoption statistics for 2025:
- Google is still the most popular search engine with over 5 billion users globally, processing an estimated 16.4 billion searches per day. (Demandsage, August 2025)
- AI adoption rate has increased massively: from 14% in February to 29.2% in August. (HigherVisibility, August 2025)
- ChatGPT is the most dominant force in the LLM space, driving 800 million weekly users and processing 2.5 billion prompts each day. (Exploding Topics, August 2025)
- Nearly 40% of Americans use at least one AI chatbot once per month or more. (SparkToro, August 2025)
- 20% of Americans are heavy users of LLMs. They use ChatGPT, Perplexity, Claude, Copilot, Gemini, and Deepseek more than 10 times a month. (SparkToro, August 2025)
- Despite a high AI adoption rate, 95% of Americans continue using traditional search engines like Google, Bing, Yahoo, and DuckDuckGo. (SparkToro, August 2025)
- The average Google Search usage went up to 12.6 sessions per week after people began using ChatGPT, compared to 10.5 weekly sessions before ChatGPT adoption. (Semrush, August 2025)
- ChatGPT usage for general information searches has tripled from 4.1% to 12.5% in just 6 months (February to August 2025). (HigherVisibility, August 2025)
- The most common use cases of ChatGPT are Practical Guidance (29%), Seeking Information (24%) and Writing (24%). These account for 77% of all ChatGPT conversations. (OpenAI and Harvard, September 2025)
- AI Mode usage has increased roughly 4x since its launch, growing from 0.25% in May to over 1% in July. (Semrush, September 2025)
- AI Mode hits 75 million users, said Google’s Nick Fox. (Search Engine Journal, December 2025)
- Copilot (25.2x growth) and Claude (12.8x growth) are the fastest-growing AI platforms, signaling AI discovery is moving into embedded workplace tools. (Previsible, December 2025)
- YMYL industries show the biggest AI adoption: Legal (11.9x), finance (2.9x), and health (2.9x). (Previsible, December 2025)
Brand Mentions & Citations in AI Answers
Search visibility isn’t just about ranking in traditional SERPs anymore. It’s about being consistently recommended and cited by AI-generated answers.
But, in order to win the war for AI inclusion, you have to know the battlefield first.
- The average AI Mode answer contains 12.6 links. (SE Ranking, August 2025)
- In comparison, AI Overviews link to 13.3 sources on average. (SE Ranking, August 2025)
- Despite being Google products, AI Overviews and AI Mode show limited response overlap: just 10.7% of URLs and 16% of domains. (SE Ranking, August 2025)
- AI Mode has extremely high volatility. When performing three tests for the same query, AI Mode had overlapping results with itself just 9.2% of the time. (SE Ranking, August 2025)
- 90.8% of AI Mode links are placed in a separate block, while only 8.9% are embedded directly within the answer. (SE Ranking, August 2025)
- AI assistants hallucinate links nearly 3x more often than Google Search. ChatGPT is the biggest offender with 2.38% of all cited URLs returning a 404 error code. (Ahrefs, September 2025)
- When users enter a specific product (e.g MacBook Air), AI Mode will show the brand’s product page. If the prompt is more general (e.g laptop), it prefers to show marketplaces instead. (Growth Memo, October 2025)
- Brands are 6.5x more likely to be cited through third-party sources than their own domains. (Airops, October 2025)
- ChatGPT now show fewer brand mentions in its responses after the recent October algo update, dropping from around six–seven to three–four per answer. (Profound, November 2025)
- AI Overview content changes 70% of the time for the same query. And when it generates a new answer, 45.5% of citations get replaced with new ones. (Ahrefs, November)
- YouTube mentions and branded web mentions are the top factors that correlate with AI brand visibility in ChatGPT, AI Mode, and AI Overviews. (Ahrefs, December 2025)
- AI Overviews and AI Mode cite different sources: only 13.7% of citations overlap between the two Google AI features. (Ahrefs, December 2025)
- AI recommendations are highly inconsistent: there’s a <1 in 100 chance that ChatGPT or Google’s AI, if asked 100X, will give you the same list of brands in any two responses. (SparkToro, January 2026)
- 44.2% of all LLM citations come from the first 30% of text (the intro). 31.1% of citations come from the 30-70% of a text (the middle). 24.7% of citations come from the last third of an article (the conclusion). (Growth Memo, February 2026)
ChatGPT Statistics
Find out how exactly ChatGPT’s query fan-out works and how the chatbot decides which brands to surface and which sources to cite.
- Out of 8,500+ prompts analyzed in ChatGPT, around 31% trigger a web search. (Nectiv, October 2025)
- Prompts with a local intent trigger a web search in 59% of all instances. (Nectiv, October 2025)
- General “Commerce” is also likely to trigger the search function (41% of the time). Credit Cards and Fashion were the least likely to have ChatGPT perform a search with only 18% and 19% of instances respectively. (Nectiv, October 2025)
- On average, ChatGPT performs 2 searches per query. Each search is typically between 5-6 words long. (Nectiv, October 2025)
- Jobs & Career and Software have an average of nearly 3 fan-out searches, the highest of any sectors. (Nectiv, October 2025)
- The most common terms that trigger the search function in ChatGPT are ‘reviews‘, ‘2025‘, ‘free‘, ‘features‘, and ‘comparison‘. (Nectiv, October 2025)
- Commercial intent prompts are much more likely to trigger web search in ChatGPT (53.5%) compared to informational queries (18.7%). (Josh Blyskal, January 2026)
ChatGPT ranking factors:
- LLMs.txt doesn’t matter but domain authority does. Sites with over 32K referring domains are 3.5x more likely to be cited by ChatGPT than those with up to 200 referring domains. (SE Ranking, November 2025)
- Domains with millions of brand mentions on Quora and Reddit have roughly 4x higher chances of being cited than those with minimal activity. (SE Ranking, November 2025)
- Domains with profiles on platforms like Trustpilot, G2, Capterra, Sitejabber, and Yelp have 3x higher chances to be chosen by ChatGPT as a source, compared to sites without such presence. (SE Ranking, November 2025)
- Pages with FCP under 0.4 seconds average 6.7 citations, while slower pages (over 1.13 seconds) drop to just 2.1. This means that fast-loading pages are 3 times more likely to be cited by ChatGPT compared to slower ones. (SE Ranking, November 2025)
- Content depth, structure, and freshness also matter a lot if you want to secure ChatGPT citations. (SE Ranking, November 2025)
- ChatGPT is more likely to cite content that uses definite language (not vague), contains a question mark, has a high entity density, has a balanced mixed of facts and opinions, and uses simple writing structures. (Growth Memo, February 2026)
User Behaviors in AI Mode
AI has a massive impact on the way people find information and interact with websites.
Here are some interesting findings from recent studies:
- Around 93% of AI Mode searches end without a click. This is more than twice the rate of AI Overviews, where 43% result in zero clicks. (Semrush, September 2025)
- Users spend double the time in AI Mode compared to AI Overviews (49 seconds vs 21 seconds in average). (Growth Memo, October 2025)
- The median time spent on different tasks in AI Mode: 77 seconds for comparing brands or products, 71 seconds for learning information, and 52 seconds for choosing or purchasing products. (Growth Memo, October 2025)
- In 75% of AI Mode sessions, users never left the pane. In other words, most AI Mode sessions end without external visits. (Growth Memo, October 2025)
- Clicks in AI Mode are mostly reserved for transactions. (Growth Memo, October 2025)
Top Cited Domains in AI Answers
Understanding which domains are most frequently cited in AI answers is crucial for marketers and publishers alike.
This data tells you where you should build your brand presence and which content formats AI systems are likely to pull from.
- Brand popularity (measured by search volume) has a high correlation with mentions in AI chatbots, especially ChatGPT. (Growth Memo, March 2025)
- Wikipedia is the most cited source in ChatGPT (7.8%), followed by Reddit (1.8%), Forbes (1.1%), and G2 (1.1%). (Profound, June 2025)
- Google is more likely to pull data from Wikipedia, YouTube, Reddit, and Quora for AI Overview responses. (Ahrefs, June 2025)
- The top mentioned domains in Perplexity answers are YouTube, Wikipedia, Apple, and Google. (Ahrefs, June 2025)
- AI Overviews prefer websites with strong brand signals (like brand mentions and branded searches) more than ChatGPT and Perplexity. (Ahrefs, July 2025)
- Besides Wikipedia and Reddit, review and affiliate sites like Forbes, Nerdwallet, Investopedia, Value Penguin, and others are heavily cited in LLM responses. (Amsive, August 2025)
- The top cited domains in ChatGPT in the US are Reddit, Wikipedia, Amazon, Forbes, and Business Insider. (Ahrefs, September 2025)
- Wikipedia, YouTube, Google’s blog, and Reddit are the most cited domains in Google’s AI Mode. (Ahrefs, September 2025)
- G2 is the most cited software review platform on ChatGPT, Perplexity, and Google’s AI Overviews. (Radix)
Google AI Overview Statistics
We’re entering the era of zero click search.
AI Overviews are pushing down organic search results and stealing clicks from publishers.
But what keywords actually trigger an AI Overview? How significant the effect is on traffic? And what sectors are impacted the most?
Let’s look at the numbers.
- 60% of searches in traditional search engines end without a click due to AI summaries. (Bain, February 2025)
- Click-through rate (CTR) drops from 15% to 8% when an AI Overview is present (Pew Research Center, July 2025)
- Only 1% of searches lead to users clicking a link within an AI Overview. (Pew Research Center, July 2025)
- Users tend to end their search session more often when an AI answer is shown – 26% compared to 16% for results pages without AI Overviews. (Pew Research Center, July 2025)
- AI Overviews show up 25% of the time. When a video is shown, it’s almost always coming from YouTube. (Conductor, November 2025)
- Organic CTR for queries where an AI Overview is present has dropped 61% YoY (June 2024 – September 2025). (Seer Interactive, November 2025)
- But when your brand is cited in the AI Overview, organic CTR is 35% higher. (Seer Interactive, November 2025)
- AI Overviews appear in 99.9% of informational keywords. 46% of which are long-tail queries (7 words or more) and 57.9% are question queries. (Ahrefs, November 2025)
- Sectors with the highest AI Overview share are Science (43.6%), Health (43.0%), Pets & Animals (36.8%), and People & Society (35.3%). (Ahrefs, November 2025)
- While sectors with the lowest share of AI Overviews include Shopping (3.2%), Real Estate (5.8%), Sports (14.8%), and News (15.1%). (Ahrefs, November 2025)
- Only 7.9% of local searches trigger an AI Overview. (Ahrefs, November 2025)
- AI Overviews now reduce clicks by 58%. (Ahrefs, February 2026)
Overlap Between Organic and AI-Driven Search
Does better SERP performance translate into higher visibility in AI-generated answers? The answer will depend on the platform.
What we so far from multiple studies, AI Overviews have the strongest correlation with traditional search rankings.
In contrast, LLMs like ChatGPT and AI Mode tend to draw from a wider range of sources, often citing lower-ranking or even non-ranking pages if they provide contextually relevant information.
- Websites with more organic traffic tend to get more mentions in AI Overviews and Perplexity. (Ahrefs, June 2025)
- However, there’s a weak correlation between high organic traffic with ChatGPT inclusion. (Ahrefs, June 2025)
- 76.1% of URLs cited in AI Overviews also rank in the top 10 of Google search results. (Ahrefs, July 2025)
- ChatGPT Search primarily cites lower-ranking pages (position 21+) about 90% of the time. (Semrush, July 2025)
- Only 12% of URLs cited by ChatGPT, Perplexity, and Copilot rank in Google’s top 10 search results. (Ahrefs, August 2025)
- 80% of LLM citations don’t even rank in Google’s top 100 for the original query. (Ahrefs, August 2025)
- AI Mode also has limited overlap with traditional Google results. Only 14% of URLs cited by AI Mode rank in the top 10. (SE Ranking, August 2025)
- If your website ranks first on SERP results, there’s a 33.07% chance that it will also appear in AI Overviews. (Writesonic, August 2025)
- Just 10% of ChatGPT’s short-tail query results overlap with Google SERPs. (Ahrefs, September 2025)
- 28.3% of ChatGPT’s most cited pages have zero organic visibility. (Ahrefs, October 2025)
AI Referral Traffic and Conversions
Traffic from AI is still relatively small (only 0.1% of the total web traffic), but it’s growing fast.
Early research also showed that AI referral visitors convert better than traditional search visitors. Although more recent studies say otherwise.
These stats highlight the current size and quality of AI traffic.
- AI traffic has grown by 9.7% since 2024. (Ahrefs, June 2025)
- Gen AI traffic is growing 165x faster than organic search traffic. (WebFX, June 2025)
- ChatGPT is the biggest AI referrer, now sending more referral traffic than Reddit and LinkedIn. (Ahrefs, June 2025)
- AI traffic drove 12.1% more signups for Ahrefs, despite making only 0.5% of all visitors. (Ahrefs, June 2025)
- LLM traffic has higher conversion rates than organic traffic: ChatGPT (15.9%), Perplexity (10.5%), Claude (5%), and Gemini (3%). Google’s organic conversion rate is 1.76%. (Seer Interactive, June 2025)
- Users from LLMs like ChatGPT and Perplexity browse fewer pages and spend less time on your website. (Ahrefs, June 2025)
- 50% of links in ChatGPT 4o responses point to business or service websites. (Semrush, July 2025)
- LLM visitors convert 4.4x better than organic search visitors. (Semrush, July 2025)
- AI-referred sessions saw a 527% year-over-year increase. (Search Engine Land, August 2025)
- Consultancy-driven sectors like legal, finance, health, and insurance seem to drive higher AI visitors compared to other industries like SaaS and eCommerce. (Search Engine Land, August 2025)
- ChatGPT’s referral traffic to e-commerce websites generates lower conversion rates and revenue per session than Google’s organic and paid search visitors. (Maximilian Kaiser and Christian Schulze, October 2025)
- Despite strong growth earlier this year, AI referral traffic has dropped -42.6% since July. (Kevin Indig, November 2025)
- On average, traditional organic traffic accounted for 25% of all website traffic, while AI referral traffic accounted for just 1.08%. (Conductor, November 2025)
- 87.4% of all AI referral traffic comes from ChatGPT. (Conductor, November 2025)
Content Optimization for AI Search
Optimizing content for search engines is a bit different than optimizing for LLMs.
Generative AI engines no longer just match exact keywords – they interpret intent and context. These systems look for comprehensive, semantically rich answers that align with what the user is really asking.
- (Suggested reading: How to Create Extractable Snippets for AI Tools)
Some models also demonstrate a clear bias towards content freshness more than the others.
These stats will give you a better picture of what it takes to create AI-friendly content that gets picked up by AI Overviews, ChatGPT, and other LLMs.
- When it comes to securing AI mentions and citations, content depth (sentence and word counts) and readability matter most, while traditional SEO metrics like traffic and backlinks have little impact. (Growth Memo, March 2025)
- Response time is another strong factor: content that loads faster is more likely to be included by AI. (Growth Memo, March 2025)
- Q&A is the best format for AI search. Structured content (headings and lists) is almost as effective for non-question queries, while dense paragraphs perform worst. (Chris Green, June 2025)
- 85% of AI Overview citations were published in the last two years. 44% are from 2025. (Seer Interactive, June 2025)
- 50% of Perplexity citations are content published in 2025 alone. (Seer Interactive, June 2025)
- ChatGPT is more likely to reference older content than AIO and Perplexity, with 29% of its citations dating back to 2022 or earlier. (Seer Interactive, June 2025)
- AI Overviews are more likely to cite AI-generated content than human-written content. (Ahrefs, July 2025)
- 86.85% of the time, AI Overviews don’t include the exact search query. (Writesonic, August 2025)
- Case studies and pricing pages are the best content types to drive traffic in the age of AI. Top funnel content (what is, how-tos, guides) saw massive drops in past two years. (Siege Media, September 2025)
- Content freshness score is a major ranking factor across seven AI models: GPT-4o, GPT-4, GPT-3.5, LLaMA-3 8B/70B, and Qwen-2.5 7B/72B. (Metehan Yesilyurt, October 2025)
- Content translation can boost your AI visibility. Translated sites gain 327% more visibility in AI Overviews compared to untranslated sites. (Weglot, October 2025)
- Google is pushing content that demonstrates expertise and unique perspective in AI Overviews. Low-quality content that just repeats what’s already out there without adding something new will be down-ranked. (Liz Reid, VP of Google Search, October 2025)
- More G2 reviews lead to higher visibility, but the number isn’t significant: a 10% increase in reviews leads to a 2% increase in AI citations. (Kevin Indig, October 2025)
- Most cited pages in ChatGPT by content type: Wikipedia (29.7%), Homepage/Landing Page (23.8%), and Educational Page (19.4%). (Ahrefs, October 2025)
- AI Overviews love factual statements: the typical AIO-cited article covers 62% more facts than the typical non-cited one. (Surfer SEO, November 2025)
- Core sources (pages that are cited every time AI Overview is generated for a topic) cover 42% of key facts for their topic. (Surfer SEO, November 2025)
- Google’s John Mueller said you don’t need to create bot-only Markdown or JSON clones of existing pages just to be understood by LLMs. Clean HTML works just fine. (John Mueller, November 2025)
- Articles over 2,900 words average 5.1 citations, while those under 800 get 3.2. For smaller domains, content length has roughly 65% more impact on ChatGPT citations than it does for top domains. (SE Ranking, November 2025)
- Content updated in the past three months averages 6 citations versus 3.6 for outdated pages. (SE Ranking, November 2025)
- In terms of structure, pages that use 120–180 words between headings receive 70% more ChatGPT citations than pages with sections under 50 words. (SE Ranking, November 2025)
- Using question-based headings and FAQ sections also boost your chances of being cited by ChatGPT. However, FAQ schema markup isn’t essential. (SE Ranking, November 2025)
- “Best X” listicles are the most cited page types in ChatGPT responses, accounting for 43.8% of all page types. (Ahrefs, December 2025)
AI Bot, AI Crawler & AI Agent Statistics
If you want to appear in AI answers, your content should be accessible to AI bots and crawlers.
And now, with the rise of agentic AI, techical optimization is even more important if you want to protect (and improve) conversions.
- 92% of the time, ChatGPT agents rely on the Bing Search API (instead of the live SERPs) to search for information. (Search Engine Land, October 2025)
- 46% of ChatGPT bot visits begin in reading mode, a plain HTML version of a webpage with no images, CSS, JavaScript, and schema markup. (Search Engine Land, October 2025)
- After landing on a page, 63% of ChatGPT agents leave immediately. (Search Engine Land, October 2025)
- Common reasons for AI agents bouncing include HTTP errors (4XX and 5XX), 301 redirects to unexpected URLs, loading issues (including slow load time), CAPTCHAs, and bot blocking. (Search Engine Land, October 2025)
- 79% of major news publishers (like BBC, The New York Times, and The Guardian) block AI training bots via robots.txt. 71% of them also block AI retrieval bots. (BuzzStream, December 2025)
- An analysis of 66.7 billion web crawlers found that LLM training bots are losing access, as more and more websites block them. In contrast, AI search bots continue to expand their reach on the web. (Hostinger, January 2026)
- LLM bots (like ChatGPT and Perplexity) crawl websites much more frequently than traditional search engines (like Google and Bing). (Shannon Vize, January 2026)
SEO is Evolving, But It’s Far from Dead
AI is the hottest thing at the moment, and no one can deny its huge impact on the way people search for information online.
But studies have shown that customers haven’t abandoned traditional search engines. AI isn’t replacing Google – it’s expanding search.
And since LLMs still need Google to access the world’s information, SEO remains essential. The playbook may be changing, but the game is far from over.
Want to boost your visibility in both traditional and AI-driven search? Partner with an AI SEO agency today!
