Welcome to our live and regularly updated database of GEO & AI SEO research – where we track, collate and share all of the latest original data led studies and insights from GEO experts into the evolving and fast moving field of GEO/AI SEO (AIO/LLMO/AEO).
If you have research you’d like to submit to be added below, please share it with us here. Note that we don’t share opinion pieces, blog posts or how-tos – we only share original, data led research.
Updated: GEO & AI Search Research, Studies & Insights (Chronological Order)
Last Updated: 22nd July 2025
Does Ranking Higher on Google Mean You’ll Get Cited in AI Overviews?
Author: Ahrefs
Publish Date: 21/07/2025
Ranking higher on Google increases your chances of being cited in Google’s AI Overviews, but it’s not guaranteed. Ahrefs found a strong correlation – about 50% of #1 ranking pages are cited, but many AI cited sources don’t rank in the top 10 at all. Other factors like content freshness, specificity, and alignment with AI prompts also influence citations. SEO fundamentals still matter, but optimising for AI visibility requires a broader strategy. Since Google Search Console doesn’t show AI citations, tools like Ahrefs’ Brand Radar are essential. Success now depends on balancing traditional SEO with AI focused content and monitoring tools.
View: Click to open
AI Overviews Cite AI-Generated Content More Than Human Writing
Author: Ahrefs
Publish Date: 14/07/2025
Ahrefs analysed 38,425 URLs cited in Google’s AI Overviews and found that 48.1% contained detectable AI-generated content, while only 26.8% were primarily human written. A further 25.1% were mixed or undetermined. This suggests Google’s AI Overviews are significantly more likely to cite AI generated pages than human authored ones.
Additional findings:
- Pages written with tools like ChatGPT, Claude, and Gemini were overrepresented in citations.
- Content created by human writers using tools like Grammarly or Jasper had lower citation rates.
- AI generated content was most frequently cited in categories like tech, health, and how-to queries.
- Sites with high domain authority were still more likely to be cited overall, regardless of content origin.
Ahrefs concludes that content origin (AI vs human) now plays a role in AI visibility, but not in the way many expect. AI written content, when aligned with searcher intent and structurally clear, is thriving in Google’s AI outputs.
View: Click to open
AI Overview Analysis & Study of 118M Searches: July 2025
Author: Conductor
Publish Date: 10/07/2025
Conductor’s July 2025 AI Overview study analysed 118 million real search keywords to track how Google’s AI-generated overviews (AIOs) are transforming search results. The report found that 18% of all tracked keywords now trigger an AI Overview – a 29% increase since May and a 112% jump since April. Desktop devices account for 61% of AIOs, with mobile presence stabilizing at 37.5%. Industries most impacted include IT Services (38% of keywords trigger AIOs), Healthcare Equipment & Supplies (36%), Life Sciences Tools & Services (36%), Education Services (35%), and Biotechnology (34%). The largest growth was seen in Healthcare Equipment & Supplies (+24 points since April). Most AIO-triggered searches are informational or conversational. The U.S. leads globally, but international AIO presence is expanding rapidly. The study emphasises the need for brands to optimize content for AI-driven search, as AIOs increasingly dominate visibility across sectors.
View: Click to open
700+ Google Search Console Results Showing Better Rankings But Lower CTRs
Author: Lily Ray / Amsive
Publish Date: 10/07/2025
Lily Ray, Vice President, SEO Strategy & Research at Amsive, shares data showing her analysis of a number of Google Seach Console accounts (over 700) from the last 3 months (May – July 2025), year over year. In the screenshot she shares, she shows a large number of sites which have seen an increase in rankings on Google SERPs, but a decrease in CTR as a result of AI search and overviews.
View: Click to open
Does Being Mentioned on Highly Linked Pages Influence AI Mentions?
Author: Ahrefs
Publish Date: 08/07/2025
Patrick Stox explores whether mentions on well-linked pages (high referring‑domain count) correlate with AI assistant visibility. Analysing ~76.7 million Google AI Overviews, 957 k ChatGPT prompts, and 953 k Perplexity prompts for June 2025, he calculated Spearman correlations between brand mentions and visibility. Results: Google AI Overviews showed a strong correlation (ρ = 0.70), Perplexity a moderate one (0.40), while ChatGPT had a very weak link (0.12). The study suggests that being cited on popular, highly credible sites boosts visibility in Google’s AI feature, whereas other AI systems are less influenced. The author cautions that correlation doesn’t equal causation and indicates that larger-scale studies will follow.
View: Click to open
Google Seems More Biased Towards Big Brands Than ChatGPT and Perplexity
Author: Ahrefs
Publish Date: 07/07/2025
Continuing the theme of brand visibility, this study examines whether the volume of branded web mentions predicts AI visibility. Again using data from Brand Radar across Google AI Overviews, Perplexity, and ChatGPT, it found a strong correlation between mentions and Google’s AI visibility (ρ = 0.65), but much weaker signals for Perplexity (0.30) and ChatGPT (0.15). The implication is that Google’s AI Overviews favour established brands, likely to combat misinformation, while the other systems show less bias. It reinforces Google’s longstanding preference for brand trustworthiness as a signal in its AI output.
View: Click to open
AI Mode vs Google Search: The Referral Gap
Author: Garrett Sussman / iPullRank
Publish Date: 04/07/2025
Garrett Sussman shares early insights on Google’s AI Mode using Similarweb data from 100,000 searches (May 20 – June 19). The findings reveal a major drop in clickthrough behaviour: only 5% of AI Mode searches lead to an external site click, compared to 25% in traditional Google Search. However, the number of clicks per session is nearly identical (6.0 for Google, 5.9 for AI Mode), suggesting users who do click may still engage meaningfully. Sussman cautions that friction (extra steps in AI Mode UI) and novelty (users still learning how to use the feature) may be skewing behaviour. He emphasises that this is early-stage data and not yet indicative of long-term patterns, but warns that if AI Mode becomes the default within 6–12 months, the industry must start preparing now.
View: Click to open
Does Being Mentioned on High Traffic Pages Influence AI Mentions?
Author: Ahrefs
Publish Date: 03/07/2025
In this study, “web visibility” is defined as the total organic traffic to pages mentioning a brand. Analysing the same AI data set as prior articles, researchers assessed correlations between web visibility and AI mentions. Findings revealed a moderate correlation for Google AI Overviews (ρ = 0.55), weak correlation for Perplexity (ρ = 0.35), and very weak correlation for ChatGPT (ρ = 0.20). This suggests that brands featured on high-traffic pages increase the likelihood of being cited by Google’s AI summarisation tools, underscoring the value of content visibility and distribution for brand recognition in AI search landscapes.
View: Click to open
AI Traffic Has Increased 9.7x in the Past Year
Author: Ahrefs
Publish Date: 26/06/2025
Ahrefs updates its March 2024 study with data from 81,947 sites, revealing that average AI-driven search traffic has surged roughly 10‑fold while traditional search traffic dropped by ~21%. AI referrals now account for about 0.25% of total site traffic on average – still small, but rapidly growing. Interestingly, AI is now Ahrefs’ highest‑converting channel, delivering over 10% conversion rate. While metrics lump AI and regular search together in analytics tools, the study confirms a major shift: AI-overview features and modes are increasingly influencing traffic distribution. The drop in traditional clicks is tied to zero‑click AI summaries, and there’s a call to revisit analytics to better distinguish referral sources and track AI traffic more accurately.
View: Click to open
AI Search Currently Drives Less Than 1% of Traffic To Most Sites
Author: G-Squared Interactive
Publish Date: 25/06/2025
Glenn Gabe analyses AI search traffic using Similarweb clickstream data, comparing it to traditional Google organic traffic. He finds that AI tools like Perplexity and ChatGPT (with SearchGPT) are still driving far less traffic than Google, but their growth is notable – especially for high-ranking, authoritative content. In some cases, Perplexity drives thousands of monthly visits. Key takeaways: Perplexity is more likely to drive direct clicks than ChatGPT (which often references but doesn’t link), and visibility in AI answers doesn’t always translate to traffic. Gabe stresses the importance of branded search terms, featured content, and domain authority to increase AI visibility. While traffic volumes are currently small, trends suggest growing AI influence – and marketers should begin optimising now.
View: Click to open
AI Search Intent Study: What 50M+ ChatGPT Prompts Reveal
Author: Profound
Publish Date: 25/06/2025
Analysing over 50 million ChatGPT prompts, Profound found a sharp drop in informational intents—from around 52% to just 32%. In contrast, transactional and navigation intents have grown, signalling that users now expect ChatGPT to help with tasks—not just deliver information. This “intent shift” challenges traditional SEO strategies, which have emphasised informative content. Profound suggests content creators now need to focus on practical utility—tools, templates, checklists—that fit this new behaviour pattern.
View: Click to open
AI Visitors Visit Fewer Pages and Bounce More Often Than Traditional Search Visitors
Author: Ahrefs
Publish Date: 24/06/2025
Drawing from the same 81,947‑site dataset, this analysis compares AI-source visits (via ChatGPT, Perplexity, etc.) with traditional search traffic. Findings show AI users visit fewer pages (4 vs 5.2 for search) and engage less per session (session duration divided by pages visited = 2.27 for AI vs 2.79 for search). They also bounce more frequently, indicating shallower browsing. The study suggests these visitors are likely seeking targeted answers rather than exploring a site broadly—highlighting a shift in user behaviour that site owners should acknowledge when assessing traffic quality from conversational AI referrals.
View: Click to open
The New Normal
Author: Kevin Indig
Publish Date: 17/06/2025
Kevin Indig explores how AI is reshaping search metrics and strategies, with interesting predictive data on when ChatGPT might overtake Google Search based on modelling various growth rates. He advocates a 360° approach: adjusting to AI-driven query behaviour, evolving measurement frameworks, and closely monitoring emerging KPIs. The memo emphasises preparing for AI Mode’s eventual mainstream rollout and the need to redefine success metrics – away from clicks toward user satisfaction and task completion.
View: Click to open
Does AI Search Traffic Convert Better Than Traditional Search? For Ahrefs, Yes: 0.5% of Visitors Drove 12.1% of Signups
Author: Ahrefs
Publish Date: 16/06/2025
Ahrefs reports that AI search visitors convert 23× better than traditional search visitors: 12.1% of Ahrefs signups stem from AI traffic, despite it only representing about 0.5% of visits. AI-sourced users go through 50% more pages and have lower bounce rates, but spend less overall time onsite. These signals point to higher purchase or signup intent—AI users seem to come more ready to act. The authors caution, though, this trend may not scale linearly as AI becomes more common, and analytics attribution remains imprecise.
View: Click to open
80% of Our AI Search Traffic Goes to Our Homepage, Product Pages, and Free Tools
Author: Ahrefs
Publish Date: 16/06/2025
This analysis of Ahrefs Web Analytics (30‑day timeframe) reveals 80% of AI-sourced visits land on high-intent pages: free tools (36.5%), product pages (23.1%), and the homepage (20.4%). This contrasts with domain advice focused on informational content—AI assistants disproportionately direct users toward conversion-focused or branded pages. A small percentage (~3.6%) even lead to non-existent “hallucinated” pages, mainly from ChatGPT. While “best-of” content and guides still attract AI traffic, brands should optimise high-intent pages for AI visibility. The study also encourages monitoring misdirected AI traffic and setting up redirects for hallucinated URLs.
View: Click to open
86% of Top Mentioned Sources Are Not Shared Across ChatGPT, Perplexity, and AI Overviews
Author: Ahrefs
Publish Date: 12/06/2025
Ahrefs Brand Radar analysed ~76.7M Google AI Overviews, 957k ChatGPT prompts, and 953k Perplexity prompts for June 2025. It found striking divergence in citation sources: only 7 out of the top 50 domains were common to all three platforms—just 14%. Preferences differ by assistant: Google AI leans heavily on authoritative and user-generated sites (Wikipedia, YouTube, Reddit), ChatGPT cites publishers and news outlets, and Perplexity draws from regional and niche sources. This highlights platform-specific algorithmic biases and emphasizes that SEO optimisation should tailor for multiple AI ecosystems, not just Google.
View: Click to open
AI Scraping Is On The Rise. TollBit State of the Bots – Q1 2025
Author: Tollbit
Publish Date: 11/06/2025
TollBit’s updated Q1 2025 report (following on from its Q4 2024 report) shows AI assistant traffic rose 39.8% quarter-over-quarter, now making up 4.7% of total traffic across 2,752 publisher sites. ChatGPT led with 55.3% of identifiable assistant visits, followed by Perplexity (22.7%) and Claude (6.9%). News and health sites continue to see the highest AI-driven engagement. The report also highlights a sharp increase in “shadow AI traffic” – bots with hidden or no user-agent strings – accounting for 62% of all assistant traffic, up from 49% in Q4 2024. This growth signals increasing AI content scraping without attribution or monetisation. TollBit again urges publishers to recognise AI as a traffic source requiring visibility, governance, and monetisation strategies.
View: Click to open
The 10 Most Mentioned Domains for ChatGPT, Perplexity, and AI Overviews Across 78.6M Searches
Author: Ahrefs
Publish Date: 11/06/2025
Ahrefs used its Brand Radar dataset (~76.7 M Google AI Oversees, 957 k ChatGPT, 953 k Perplexity prompts) to identify the top 10 domains most frequently cited by AI assistants. Wikipedia leads overall—16.3% in ChatGPT, 12.5% in Perplexity, and 8.4% in Google AI Oversees—while YouTube ranks high in Perplexity (16.1%) and Oversees (9.5%) but is absent in ChatGPT. Google favours user-generated content (Reddit, Quora) in Oversees (7.4% and 3.6%), whereas ChatGPT emphasises news outlets like Reuters, AP, and AS.com. Impression/potential reach analysis, weighted by search volume, shows institutional and medical sources like Mayo Clinic also hold considerable visibility in Oversees. The findings underline how each AI assistant follows a distinct content citation bias.
View: Click to open
We Studied the Impact of AI Search on SEO Traffic. Here’s What We Learned.
Author: Semrush
Publish Date: 09/06/2025
Semrush examined over 500 SEO and digital marketing query topics to project how AI search will affect traffic and revenue. Their model indicates that by early 2028, AI-sourced visits could surpass traditional search visits for such topics—potentially sooner if Google’s AI Mode becomes the default. This trend suggests AI search’s rapid ascension—with significant implications for industry traffic patterns and optimisation strategies.
View: Click to open
Semrush AI Overviews Study: What 2025 SEO Data Tells Us About Google’s Search Shift
Author: Semrush
Publish Date: 05/05/2025
Semrush analysed over 10 million keywords (January–March 2025) to assess the growing prevalence of AI Overviews in SERPs. The share of queries triggering AI Overviews doubled from 6.49% in January to 13.14% in March. These features predominantly appear on informational queries (88.1%), and navigational triggers also doubled. Sector-wise, topics like science (+22.3%), health (+20.3%), people & society (+18.8%), and law & government (+15.2%) saw the highest increases. Surprisingly, zero-click rates for the same keywords decreased slightly after AI Overviews were introduced—suggesting users may still click through after reading a summary. The key takeaway: AI Overviews are reshaping search—marketers must optimise for them to maintain visibility.
View: Click to open
AI Overviews Reduce Clicks by 34.5%
Author: Ahrefs
Publish Date: 17/04/2025
Analysing 300K informational keywords, this Ahrefs study compared CTRs from March 2024 (pre-AI Overviews) and March 2025. Position‑one CTR dropped from 5.6% to 3.1% for those without AI Overviews, while AI Overview-triggering keywords saw CTR fall even more dramatically—from 7.3% to 2.6%. This amounts to an estimated 34.5% CTR reduction attributed directly to AI Overviews. The mechanism resembles Featured Snippets, providing answers directly in the SERP and decreasing traditional link clicks. Despite Google’s assertion that links within Overviews get more traction, Ahrefs notes that current tools cannot differentiate these click types. The study warns that as AI summaries become more routine, passive “zero-click” searches will likely increase, further impacting organic traffic.
View: Click to open
Google AI Overviews: New CTR Study Reveals How to Navigate Negative SERP Impact
Author: Amsive
Publish Date: 16/04/2025
This study analysed 700K keywords across five industries and found that the introduction of Google’s AI-generated Overviews has significantly disrupted clickthrough rates. On average, CTR dropped by 15.5%, with non-branded (–19.98%) and lower-ranked keywords (–27.04%) hit hardest. Overlapping Featured Snippets combined with AI Overviews led to an even steeper CTR decline of 37%. Interestingly, branded queries that did trigger an AI Overview saw a CTR increase of 18.7%, suggesting brand credibility can offset visibility loss. The research recommends adapting SEO strategy: aim for top positions but also focus on securing Featured Snippets, optimise for high-intent non-branded queries, and double down on branded content. The takeaway is to recalibrate your SEO playbook for an AI-dominated SERP environment.
View: Click to open
Does Brand Awareness Impact LLM Visibility?
Author: Seer Interactive
Publish Date: 16/04/2025
Seer Interactive analysed correlations between brand mention volume (MSV) and LLM visibility. Overall correlation was modest (ρ ≈ 0.18), second only to Domain Rank (ρ ≈ 0.25). In high-trust verticals like finance, brand awareness appears to meaningfully improve LLM mentions. The conclusion: awareness contributes to LLM visibility, but only in tandem with strong domain authority, backlinks, and credible content. For industries reliant on trust, awareness-building campaigns—PR, expert engagement, publisher mentions—are recommended to boost AI visibility.
View: Click to open
Marketing’s New Middleman: AI Agents
Author: Bain
Publish Date: 14/04/2025
Bain & Company argues that AI agents are fast becoming influential intermediaries between brands and buyers, shifting how consumers and businesses make decisions. These agents don’t just provide information – they make or narrow down choices. This has major implications for marketers, particularly in B2B and high-consideration consumer sectors. Key data and insights include: 28% of consumers have already used generative AI to assist in purchasing decisions. Among these, 70% say it improved decision quality and 63% say it saved time. Bain predicts AI agents will soon dominate early discovery and evaluation stages of the buyer journey.
View: Click to open
AI Scraping Is On The Rise. TollBit State of the Bots – Q4 2024
Author: Tollbit
Publish Date: 24/02/2025
TollBit’s Q4 2024 report analyses 295 million visits across 2,420 publisher sites to track how LLMs and bots interact with web content. The headline finding: traffic from AI assistants and bots increased by 17.2% quarter-over-quarter, now accounting for 3.4% of total traffic. ChatGPT, Perplexity, and Claude led the charge, with ChatGPT visits up 20.7%. News publishers and health sites saw the highest share of AI traffic, with Perplexity disproportionately favouring health content. The report also notes a rise in unidentified bot traffic (up 30.4%), suggesting growing use of non-transparent agents. TollBit emphasises that most of this AI traffic is unmonetised—publishers receive no revenue despite their content powering AI outputs. To address this, TollBit promotes its tooling to help publishers identify AI agents, measure content usage, and control access. The report urges media companies to begin treating AI traffic like a commercial channel and to prepare monetisation strategies accordingly.
View: Click to open
87% of SearchGPT Citations Match Bing’s Top Results
Author: Seer Interactive
Publish Date: 06/02/2025
Seer Interactive found that in SearchGPT (ChatGPT with live search), 87% of citations align with Bing’s top 20 organic results, with many from the first page. In contrast, only around 56% match Google’s top results. This suggests that Bing’s SERP landscape heavily influences ChatGPT’s web citations. The study advises SEO to diversify efforts—tracking Bing alongside Google—and consider partnerships with external trusted publishers.
View: Click to open
Google Triggers 100% More AI Overviews for Longer Queries, New Report from BrightEdge Finds
Author: Brightedge
Publish Date: 30/01/2025
From September to December 2024, the proportion of long-tail queries (eight or more words) triggering AI Overviews doubled, showing Google’s increasing confidence in answering complex questions with AI. Approximately 25% of those longer queries now generate an AI Overview. BrightEdge highlights Google’s ability to handle nuance at scale, and implies SEO strategies must now cater to longer, more conversational queries.
View: Click to open
Marketing Leaders Want to Meet AI Search Head-On: New Survey Results
Author: Botify
Publish Date: 28/01/2025
Botify’s January 2025 survey found marketing leaders recognise the urgency of integrating AI search strategies. Organisations are prioritising structured data, metadata automation, AI-ready indexing, and cross-platform tracking (Google, Bing, ChatGPT). Key initiatives include SmartIndex (real-time AI-friendly indexing), SmartContent (AI-enriched content generation), and SmartLink (automated internal linking), showcasing a strategic shift to support visibility across both traditional and AI search.
View: Click to open
STUDY: What Drives Brand Mentions in AI Answers?
Author: Seer Interactive
Publish Date: 07/01/2025
Seer analysed 10,000 LLM-generated brand-recommendation prompts (mainly finance and SaaS). Page‑1 Google rankings had the strongest correlation with brand mentions (~0.65), followed by Bing (~0.5–0.6). Surprisingly, backlinks and multimedia content had little impact. After filtering out aggregators and forums, correlation strengthened, reinforcing the importance of rankings and PR/partnership strategies. The study suggests that while SERP prominence matters, brands should also invest in PR and partnerships to boost LLM mention likelihood.
View: Click to open
New Report From .Trends & Statista Reveals How AI Search is Changing the Web
Author: Semrush
Publish Date: 02/12/2024
This report provided early market insights, highlighting that as of July 2024, ChatGPT and Google’s Gemini dominate AI search traffic—capturing around 78% between them, with Perplexity and Bing composing the rest. It also noted approximately 13 million US adults had already adopted generative AI as their primary search tool, with projections reaching 90 million by 2027. It confirms AI search isn’t niche—it’s becoming mainstream, demanding adaptation from marketers and content creators.
View: Click to open
We Studied 200,000 AI Overviews: Here’s What We Learned
Author: Semrush
Publish Date: 30/10/2024
Focusing on the structure of Google AI Overviews, Semrush analysed how many top‑10 organic URLs appear in AI Overviews. They found low overlap: over 80% of mobile AI Overviews include three or fewer top‑10 results and only 46% of desktop and 34% of mobile Overviews included the #1 organic result. Ads rarely overlap with AI‑shown URLs. This suggests AI Overviews use different selection criteria, meaning high organic rank doesn’t ensure an AI citation. Brands have a chance to feature in Overviews even if they aren’t top in classic SEO—communicate expertise clearly to be surfaced by these AI summaries.
View: Click to open
AI Overviews Study: Inside Google’s New Search Reality
Author: Botify / DemandSphere
Publish Date: 01/10/2024
Botify’s Q4 2024 report, based on 120 000 SERPs, shows AI Overviews appearing in up to 47% of searches and occupying 75.7% of mobile viewport space when paired with Featured Snippets. Most AI Overview citations come from top‑12 organic rankings, with strong semantic alignment between page content and AI summaries. The study calls for optimising for ranking and content similarity to increase AI Overview placement.
View: Click to open
New Research From BrightEdge Finds Google’s AI Overviews Are Getting Smarter
Author: Brightedge
Publish Date: 19/09/2024
BrightEdge’s research reveals Google’s AI Overviews are evolving: they now lean heavily on specialised expert sources, comparative shopping content, and visual modules like carousels. This signals more discerning source selection and richer formats. Additionally, Google’s SearchGPT referral growth is outpacing competitors—underlining that businesses need to target both search ranking and source authority to capture AI visibility.
View: Click to open