Studies and papers
The most cited research on AI search visibility includes the 2023 GEO paper (visibility gains in a simulated engine), C-SEO Bench (2025; most rewrite tactics ineffective for citation rank), a 2023 Stanford audit of citation accuracy, Pew Research Center’s 2025 analysis of clicks with AI summaries, and Ahrefs’ 2025 correlation study of brand signals. Each is summarized below with its date and method.
Which academic studies matter most?
| Study | Date | Method | Key finding |
|---|---|---|---|
| GEO: Generative Engine Optimization (Aggarwal et al.) | Nov 2023; KDD 2024 | 10K-query benchmark, simulated engine (gpt-3.5-turbo) | Up to 40% visibility gain; citations, quotes, statistics best[1] |
| C-SEO Bench (Puerto et al.) | Jun 2025; NeurIPS 2025 | 1.9K queries, 6 domains, 4 models, citation rank | 3 of 54 cases significant; retrieval order dominates[2] |
| Evaluating Verifiability in Generative Search Engines (Liu et al.) | Apr 2023 | Human audit of 4 engines, 1,450 queries each | 51.5% of sentences fully supported; 74.5% citation precision[3] |
See the GEO paper and replications.
What does behavioral data show about AI summaries?
Pew Research Center analyzed browsing by 900 US adults in March 2025 (68,879 unique Google searches):[4]
- 18% of searches produced an AI summary.
- With a summary, users clicked a traditional result in 8% of visits; without one, 15%.
- Users clicked a link inside the summary in 1% of visits.
- Sessions ended on 26% of pages with a summary versus 16% without.
- Wikipedia, YouTube and Reddit were the most cited sources, together 15% of AI summary sources.
- The median summary was 67 words; 88% cited three or more sources.
What does industry research show about brand signals?
Ahrefs (May 26, 2025) studied 75,000 brands using Spearman correlations against AI Overview brand mentions measured in Brand Radar: branded web mentions 0.664, branded anchors 0.527, branded search volume 0.392, Domain Rating 0.326, referring domains 0.295, backlinks 0.218. Roughly 26% of brands had zero mentions.[5] This is correlational vendor research, not a controlled experiment.
How should you read AI visibility statistics?
- Check the date: engines change quickly, and this page labels each figure.
- Check the engine: findings for one engine do not transfer automatically.
- Check the metric: word share, citation rank, mention rate and clicks are different things.
- Check who measured: vendor studies may be sound but have commercial context.
Frequently asked questions
What is the most important GEO study?
The 2023 GEO paper defined the field; C-SEO Bench (2025) is the main independent retest and is essential context for its claims.
Do AI Overviews reduce clicks?
Pew’s March 2025 data showed lower click rates on traditional results when an AI summary appeared (8% vs 15% of visits). Google says clicks from pages with AI Overviews are higher quality.
Are there studies on ChatGPT citations?
The 2023 Stanford verifiability audit predates ChatGPT search and did not test ChatGPT because it lacked inline citations at the time.
See also
References
Sources accessed October 14, 2026 unless another date is given. See all sources.
- “GEO: Generative Engine Optimization”. arXiv (Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, Deshpande). v1 submitted November 16, 2023; published at KDD ’24.
- “C-SEO Bench: Does Conversational SEO Work?”. arXiv (Puerto, Gubri, Green, Oh, Yun). June 2025; NeurIPS 2025 Datasets & Benchmarks.
- “Evaluating Verifiability in Generative Search Engines”. arXiv (Liu, Zhang, Liang; Stanford). April 2023; Findings of EMNLP 2023.
- “Google users are less likely to click on links when an AI summary appears in the results”. Pew Research Center. July 22, 2025; March 2025 browsing data.
- “An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)”. Ahrefs. May 26, 2025.