ChatGPT Thinking mode changes which brands get cited

A recent study by Semrush and Kevin Indig reveals that the "thinking mode" used by ChatGPT significantly alters the brands it cites in its responses. Analyzing 100 prompts and 20 buyer journeys, the research found only a 25.6% overlap in sourced brands when comparing minimal reasoning responses to those requiring high reasoning. This suggests that the complexity of the query directly influences which sources the AI prioritizes and presents to users.

This finding is crucial for SEO professionals and marketers as it highlights the dynamic nature of AI-driven content generation and its impact on brand visibility. Brands need to consider how their content performs not just for simple queries, but also for more complex, multi-step user journeys where ChatGPT might employ deeper reasoning. Optimizing for diverse query types, and understanding how AI models interpret and source information, will be increasingly important for maintaining brand presence in AI-powered search environments.