In Magna Research, we measure how brands appear in AI answers from ChatGPT, Gemini, Perplexity and Google AI Overviews, and we publish the results. We build each study across different countries, languages and industries, around the questions buyers actually ask. In every study we state the measurement dates, queries, platforms and limitations openly.
AI engines can recommend different brands for the same question; being visible on one platform does not mean you are visible on the others. That is why we measure visibility not on a single platform but by comparing across platforms, queries and time periods.
We examined which companies AI systems recommend to buyers looking for furniture manufacturers in the German market, across five German-language commercial queries. Average overlap between platforms stayed at 14.8%.
Read the StudyWe examined which Turkish apparel manufacturers AI systems recommend when German fashion brands look for production partners, across six German-language queries. Average overlap between the lists given for the same question was only 2.1%.
Read the StudyWe examined how visible AI systems make Turkish automotive suppliers when recommending suppliers to German OEMs and Tier-1 companies, across six German-language queries. Overlap was 40.9% in Europe-wide queries but fell to 3.5% in Türkiye-focused queries.
Read the StudyWe examined which e-commerce and retail brands six AI platforms recommend across 10 different shopping needs, and how the results change by category. Overlap was 52.9% in electronics and 8.5% in premium fashion.
Read the StudyWe examined which travel, booking and comparison platforms AI systems recommend to people planning holidays and accommodation, across 10 travel needs. 80.5% of the names appeared on only one platform.
Read the StudyWe examined which brands AI systems recommend across 10 different car-buying needs in Türkiye, and how these lists change by platform and purchase intent. The lists from ChatGPT, Gemini and Perplexity overlapped by 47.9%.
Read the StudyOur research is not tied to a single platform or a single measurement method. Depending on the research question, we look separately at brand visibility, source use, search performance, industry competition or differences between countries and languages.
We examine how brands appear in answers from ChatGPT, Gemini, Google AI Overviews, Perplexity and similar AI systems: mentions, citations, source visibility, place in recommendation lists, position and comparison with competitors. AI visibility is an observed state; it is not the same thing as a site's technical readiness or the traffic it receives from AI platforms.
We analyze which web sources AI answers use for given queries, which domains recur and to what extent brands appear as sources. Our aim is not to make firm cause-and-effect claims about the platforms' closed algorithms, but to compare the source patterns that can be observed.
We examine the visibility of brands, categories, industries or query groups on Google. Where needed, we compare SEO data and AI visibility data in the same study, but we do not present organic search success as an AI visibility result.
We compare brands within an industry using a common method to examine visibility, competition and source patterns. Because every industry has its own range of questions and competitive structure, we design each study around its industry.
We compare how the same industry or group of questions performs under different country, language and market conditions. For example, we use this approach to measure how an exporter appears in the questions buyers in a target market ask in that country's language.
We start every study by defining the question we want to answer; then we set the industry, market, platforms, queries, measurement dates and the metrics we will use. We publish the findings together with the scope they were produced in and their limitations.
Research question → Scope → Data and observation → Classification → Analysis → Limitations → Findings
We explain all of our measurement rules on our GEO Methodology page.
In our AI visibility research, we take measurements with Magna AI Analyzer, a tool we developed ourselves. For the queries we define, it records the visibility of brands and their competitors, mentions, citations, place in recommendation lists, position and change over time, by platform and query. Where needed, we assess these measurements together with manual checks, Search Console and web analytics data. You can find details of our tools on our AI Visibility Tools page.
If you want to see the same measurement for your own brand and competitors, choose our AI Visibility Analysis; if you want us to improve how your brand appears in AI answers based on the findings, choose our GEO Services. You can see all of our GEO, AEO and AI visibility services together on our AI Visibility: GEO and AEO Services page.
We publish studies that measure, industry by industry, how brands appear in AI answers from ChatGPT, Gemini, Perplexity and Google AI Overviews, along with citation and source analyses and data-driven comparisons.
No. AI visibility is the focus of the research we publish, but SEO, search visibility, industry competition and cross-market comparisons are also among our research areas.
Platform coverage varies by study. In each study we state which of ChatGPT, Gemini, Perplexity, Google AI Overviews and other AI answer systems were examined.
Yes. In our AI visibility research, we take measurements with Magna AI Analyzer, a tool we developed ourselves. We explain any other data sources we use in the method section of each study.
Not on its own. A single mention or the result of a single query does not show overall AI visibility. We assess visibility within a defined set of queries, platforms, periods and competitors.
No. A brand or site being shown as a source in an AI answer does not mean it was recommended to the user in that answer. Mentions, citations, recommendations and position are separate signals, and we measure them separately in our research.
Only when the measurement conditions are sufficiently the same. When the query set, platform, country, language or metric changes, comparing results directly can be misleading; that is why we record the measurement conditions in every study.
No. The research presents the data we observed within a given scope and our analysis of it; it does not guarantee future Google rankings, inclusion in AI answers, traffic or sales.
