AI models show distinct information sourcing patterns in brand and banking queries

Recent research reveals significant differences in how AI models, specifically ChatGPT and Gemini, source information when answering consumer queries. An analysis by StoryHub indicates that ChatGPT tends to prioritize official corporate websites, with approximately 60% of its citations coming from product or service pages. In contrast, Gemini relies heavily on comparison and ranking sites, which account for about 70% of its citations.

A separate study by Bakuri and Dr. AISEO focused on the banking sector, finding that while both models primarily use official financial institution websites, their secondary sources differ vastly. Gemini consulted media and third-party articles 38% of the time, whereas ChatGPT did so only 9% of the time. ChatGPT showed a higher tendency to reference public institutions and industry associations.

Both studies highlight the volatility of AI responses. StoryHub found that repeating the same question can result in a 30% turnover in the brands mentioned. Similarly, the banking study noted that the specific pages accessed by AI models can change between identical queries, suggesting that businesses cannot rely on a single interaction to determine their visibility in AI-generated answers.