The most frequently mentioned provider appeared in 8 of 80 completed observations. That is 10% of this measured prompt set—not evidence of market leadership, and not evidence that one provider dominates AI recommendations generally.
A fragmented measured set
Research 001 asked commercial provider-discovery questions across ChatGPT, Perplexity and Yandex Alice, covering International, US, UK and Russian contexts where an engine was accessible. We preserved 80 completed responses and classified provider mentions, explicit recommendations and visible citations separately.
iPullRank had the highest validated mention count: 8 observations, with 4 explicit recommendations and coverage across 6 prompt categories, 2 engines and 2 markets. The next three providers each appeared 6 times. The distribution was therefore shallow rather than concentrated.
Why “8 of 80” needs careful interpretation
The number describes only the frozen prompts, engines, markets, interfaces and collection window used in this study. AI answers vary with model version, search availability, language, location and session context. A provider absent from the sample may still be highly capable; a provider mentioned frequently is not automatically better.
The result does suggest a practical measurement lesson: brand visibility should be monitored across several prompts and engines. A single answer cannot represent category visibility.
Mentions are not recommendations, and citations are not providers
Research 001 treated these as independent signals. A company could be named without being recommended. A media domain could be cited while a different provider was recommended. Preserving those distinctions prevented citation labels and source publishers from inflating provider metrics.
What teams can take from the study
- Use a frozen prompt set before comparing periods.
- Report denominators and coverage, not only counts.
- Separate mention, explicit recommendation and citation.
- Preserve zero results and technical failures.
- Avoid presenting observed frequency as a quality ranking.
Read the full Research 001 and the companion analysis, ChatGPT vs Perplexity: Do They Recommend the Same GEO Providers?.