AIs are inconsistent in brand/product recommendations, raising concerns for marketers relying on AI tracking. Research by Rand Fishkin and a colleague shows AIs rarely provide the same list of recommendations, complicating visibility metrics. Despite running numerous prompts via popular AI tools (ChatGPT, Claude, Google AI), the findings indicate answers vary widely, making “ranking position” tracking unreliable. Nevertheless, visibility percentages across multiple prompts can reflect brand presence in AI responses. Marketers should critically evaluate AI tracking tools, avoiding those without transparent methodologies.
