
LLM Citations: How Each AI Engine Actually Cites Sources (Data Study)
Across six AI engines and six verticals, citation depth varies ~20×: Google AI Mode averages 15–22 sources per answer, ChatGPT/Gemini/Copilot land around 4–8, and Perplexity ranges from ~4 down to literally zero depending on the topic. Reddit sits top-2 in every vertical. AI Overview decides whether to answer at all: it triggers on 98% of shopping queries and 3% of dining.







