As more consumers look for information from generative artificial intelligence (AI) before purchasing products, on the 6th the first domestic results quantifying how often specific brands are exposed in AI were released. The results showed that an industry ranking was not the only factor determining exposure frequency in AI. Even with the same question, the brands most frequently surfaced differed by AI.

AI brand analysis company ‘AIBIX Lab’ unveiled its AI competitiveness index ‘AIBIX’ for 715 brands across 30 categories in the food and beverage and franchise sectors. They repeatedly posed identical consumer questions to four major generative AIsChatGPT·Gemini·Perplexity·Claudeand quantified, on a 100-point scale, how often particular brands appeared in AI answers, among other measures. This is a different concept from overall consumer preference. The survey was conducted from the 20th of last month to the 4th of this month.

The analysis found that market ranking did not directly translate into exposure frequency within AI. For example, by number of stores among franchise burger brands, the order is Mom’s Touch·Lotteria·Burger King·McDonald’s, but for AI exposure it appeared as Burger King·McDonald’s·Lotteria·Mom’s Touch.

In 20 of the 30 categories, the brands most frequently exposed differed by AI. Among franchise coffee chains, Mega MGC Coffee, which ranks first by store count, ranked first on Gemini, but third on Claude, fifth on Perplexity, and seventh on ChatGPT. For cup noodles, Shin Ramyun Cup topped ChatGPT and Claude, while Yukgaejang Sabalmyeon ranked first on Gemini and Perplexity. Meanwhile, 127 of the 715 brands measured (17.8%) were never mentioned by any of the four AIs.

Recently, companies have been grappling with how to respond to this ‘brand exposure within AI’. Although AI-recommended purchases by consumers are steadily increasing, companies cannot clearly know AI’s sources or criteria for recommendations. As identical questions yield different answers depending on the user, timing, and model, companies are using these variations as clues to backtrack the principles behind AI exposure. They are working to optimize their websites so AI crawlers can read them well, and striving to post specifics such as product ingredients·content·prices instead of vague ad copy. Efforts to improve online information so that a company is cited or recommended more frequently and accurately in AI answers are called GEO (Generative Engine Optimization·generative AI search optimization). An industry official predicted, “Going forward, managing content structures to be well cited in AI search results will become a new area of communications work.”

However, in this survey, it was found to be very rare for each brand’s official materials to be used as grounds for AI answers. Kim Jun-Hyeon, CEO of AIBIX Lab, said, “Because of structures that make AI access difficult, the website content that companies painstakingly create is effectively treated by AI as information that doesn’t exist,” adding, “If official information is lacking, AI may rely on unofficial content such as blogs or user reviews, and key details like a product’s ingredients·content·price may not be accurately reflected in answers.”