Ask ChatGPT, Gemini, and Perplexity the same question, and you'll often get three different sets of citations. That's not random.
Different retrieval, different results
Each engine crawls, indexes, and ranks sources with its own pipeline. Overlap exists, but it's smaller than most marketers assume.
What this means for your content
Optimizing for one engine's citation pattern doesn't guarantee the others follow. Structure content for clarity and factual density first, platform-specific quirks second.