Key point. This is where Elementera publishes its own research on AI search: what Google AI Overviews, ChatGPT and other answer engines cite, how often their answers change, and whether AI crawlers can read your site. Every figure carries its sample size and a 95% interval, and where we can, we publish the data.
Most numbers about AI search come from a single snapshot, with no sample size and no sense of how much the answer moves from one day to the next. We built our research the other way round. When we study AI answers, we ask each question at least three times and on more than one day; in every study we fix the sample and the rules before we collect, and we publish each figure with the range it could plausibly fall in. The studies below are the result, grouped by the question they answer.
Key takeaways
- Studies, data and tools in one place. Each entry links to the full write-up, and the datasets and free tools sit beside the study they came from.
- Every number has its n. Each figure on these pages carries its sample size and a 95% interval, so you can see how sure the number is.
- Measured more than once. Our research measurements ask each question at least three times and on more than one day, because AI answers change between asks.
- Free to cite. Journalists, researchers and teams are welcome to quote our figures with a link to the study.
Can AI crawlers read your site?
The first step of AI visibility is access: an answer engine can only cite pages its crawler may fetch.
- The AI Crawler Access Report. We read the robots.txt of 600 randomly sampled startup websites. 3 of 569 sites (0.5%; 95% interval 0.2% to 1.5%) block an AI search crawler, while 18 of 569 (3.2%; 95% interval 2.0% to 4.9%) block an AI training crawler, usually by name.
- The dataset. One row per sampled site and one column per crawler for all 18 crawlers we tracked, as a CSV download, with random ids instead of company names.
- Free AI crawler checker. Paste your robots.txt and see, engine by engine, whether the crawlers behind ChatGPT, Claude, Perplexity, Google and Bing may read your pages. It uses the same rules as the report.
What do AI answers cite, and how is that different from ranking?
Ranking and being cited are related, and they are not the same thing. These studies measure the gap.
- Is good AEO just good SEO? What our measurements show. In our study of 150 non-branded Google searches (US, July 2026), about 57% of the websites cited in AI Overviews did not appear in that same search's top 10 organic results (95% interval 54% to 60%). The page sets that against Google's guidance and outside studies, and shows where good SEO stops doing the AEO work.
- Citation rate by organic position. In the same Google data, "The website in the first organic position was cited by the AI Overview 56% of the time (95% interval 51% to 61%). Websites in positions 6 to 10 were cited 34% of the time (31% to 37%)." The interactive chart is in the AEO vs SEO study.
Where does a company drop out of AI answers?
- The AI search visibility funnel. Our framework for finding where a company is lost, from a crawler that cannot fetch the page to a visitor who arrives from an AI answer and leaves, with the check for each stage, built on the measurements above.
How do we measure?
Our method is the same across studies. We fix the sample and the rules before collecting, ask each question about AI answers at least three times and on more than one day, report every figure as a count with its share and a 95% interval, and say when two figures cannot be told apart at the sample size we have. When a study's data can be shared without naming companies or people, we publish it beside the study.
Frequently asked questions
Can I cite your research? Yes. Quote the figure with its sample size and link to the study page, so readers can check the method behind it. Each study page has a suggested citation where we publish the data.
Do you publish the data behind your studies? Where we can do it without naming companies or people, yes. The AI Crawler Access Report's full per-site dataset is a free CSV download, with random ids in place of company names.
How often is the research updated? Each study shows the date its data was collected and the date the page was last updated. When we re-run a study, the new figures and date replace the old ones on the same page.
Who does the research? Elementera's founders, who between them have published peer-reviewed machine learning research, design and run the studies. Each page names its author and the founder who reviewed it.
Work with us
There are two ways to work with us. Teams that want the work done for them can use our managed AEO services: Book a call. Teams that want to see where they stand first can start with Lenz, our free analysis of how ChatGPT and Gemini answer their buyers' questions: Get your free AI visibility report.
How this page was made
Dr. Maziar Gomrokchi wrote this page with the help of an AI writing assistant (Claude, made by Anthropic). Each summary is taken from the study it links to, as we published it. Soroush Mehri reviewed the page, and Dr. Maziar Gomrokchi approved it for publication.