AI citation patterns show inconsistency across search engines and industries

Research into AI citation patterns reveals that generative search engines like ChatGPT, Gemini, and Perplexity do not follow a single, stable leaderboard for source retrieval. Citation behavior varies significantly based on the specific AI system, the topic being queried, and the accessibility of the source material. Studies indicate that AI systems favor public pages with stable URLs and clear structures, often struggling to retrieve information behind gated interfaces or complex files.

In the cryptocurrency sector, the citeOS Exchange Index highlights a disconnect between AI recommendations and market reality. Despite Binance maintaining a daily spot volume approximately five times larger than Coinbase, Kraken emerged as the most frequently recommended exchange across ChatGPT, Perplexity, and Gemini. Coinbase led in mentions by Claude, while Binance placed third in AI recommendations. The study notes that rankings are inconsistent across different engines, as an exchange may be highly recommended by one chatbot while being absent from the shortlists of others.