Winners in the AI Search Era: How Beauty Brands Build Authority in Generative Recommendations

When consumers start asking ChatGPT, Perplexity, or Claude about "affordable foundations for oily skin," the digital marketing battlefield for beauty brands is undergoing a structural shift. The latest AI search research report released by Foundation Agency provides the most systematic evidence to date: in the world of generative AI recommendations, content owned by a brand is no longer the protagonist; instead, "third-party validation" from media, retailers, and communities is the key to determining visibility.

The study covered four major AI platforms, conducted three rounds of testing on 20 prompts simulating real consumer questions, and ultimately collected 1,335 brand mentions and 690 cited sources. The data clearly reveals a decentralized recommendation ecosystem: no single brand dominates all conversations, but those that have built deep recognition in specific tracks—affordable cosmetics, clean skincare, luxury beauty, or hero products—receive disproportionate exposure.

For example, under the "budget beauty" prompt, e.l.f., Maybelline, NYX, and L'Oréal Paris frequently appeared; while ILIA, Kosas, Merit, Saie, and Tower 28 dominated conversations about clean beauty. In the eye product category, Charlotte Tilbury, Makeup by Mario, Urban Decay, and Natasha Denona became top picks; lip care was dominated by Aquaphor, La Roche-Posay, Laneige, and Vaseline. This "category specialization" model is entirely different from traditional search SEO logic—in the past, brands pursued rankings for generic keywords, but now AI tends to recommend brands that are strongly associated with specific problems in consumers' minds.

The different citation preferences of various AI platforms further compound marketing complexity. ChatGPT heavily relies on traditional beauty and lifestyle media such as Vogue, Allure, and Cosmopolitan; Gemini mixes editorial content, official product information, and authoritative directories; Perplexity places more importance on retailer lists and community-driven platforms; while Claude shows the greatest diversity, often recommending brands that other models have never mentioned. This fragmentation means that brands cannot cover all channels with a single strategy—each platform constitutes an independent visibility game.

The most startling finding of the study might be the distribution of cited sources.The most startling finding of the study might be the distribution of cited sources. Among all 690 sources analyzed, content from earned media accounted for 44%, while brand-owned content made up only 7%. Even more notably, Reddit surpassed Vogue and Allure to become the single most cited source in the entire study. This validates a growing consensus: AI models, when generating responses, trust real discussions and multi-party verification within communities more than one-way brand communication.

For beauty marketers, this implies a fundamental strategic reconfiguration. The traditional "SEO + content marketing" framework is being replaced by a new triangle of "category authority + media relations + community engagement." Brands need to: define a category label recognizable by AI, invest in deep collaborations with high-credibility media, and cultivate authentic conversations on platforms like Reddit and community forums—rather than simply expanding their own blogs or product pages.

Of course, this trend also raises a deeper question: When AI recommendations increasingly rely on "what others say" rather than "what the brand says," can brands indirectly control the narrative by building genuine industry expertise and consumer trust? Those that learn to relinquish some control in the AI era may ultimately be the winners.