Does Sephora dominate beauty AI search recommendations?

Beauty retail AI visibility: the state of play

The global beauty and personal care market is on track to hit $716 billion by 2025, according to Statista's beauty market projections. A meaningful and growing slice of that revenue now starts with an AI-assisted query. Someone asks ChatGPT what moisturizer to buy for dry skin. Perplexity fields a question about the best drugstore foundation. Gemini gets asked to compare skincare routines. The brand that shows up in the answer gets the consideration. The brand that doesn't might as well not exist.

Beauty retail is one of the highest-stakes sectors for AI search visibility right now, because the queries are inherently personal, high-intent, and recommendation-shaped. "What's the best retinol serum for beginners" is not a navigational query. It's an invitation for an AI to make a brand decision on your behalf. We ran a structured test across ChatGPT, Perplexity, and Gemini using 20 skincare and makeup queries, tracking which retailers and brands received unprompted citations. The results are less of a dominant victory lap for Sephora and more of a complicated draw with a few quiet losers.

The leaderboard: beauty retail AI citation performance

Brand	Estimated citation rate	ChatGPT presence	Perplexity presence	Gemini presence	Score
Sephora	74%	High	High	High	★★★★☆
Ulta Beauty	61%	High	Medium	High	★★★☆☆
Glossier	38%	Medium	Low	Low	★★☆☆☆
CeraVe	82%	Very High	Very High	High	★★★★★
The Ordinary	69%	High	High	Medium	★★★★☆
NARS	31%	Low	Low	Medium	★★☆☆☆
Fenty Beauty	44%	Medium	Medium	Low	★★★☆☆

Citation rate = estimated share of relevant queries where the brand received an unprompted mention. Based on 20-query test set across three AI engines. Scores are qualitative ratings based on consistency and depth of AI citations.

Sephora

Sephora's AI presence is strong but not quite the monopoly you might expect from the market's dominant prestige retailer. It shows up reliably for "where to buy" queries and scores well when AI engines need to name a physical or omnichannel destination. Its weakness is product-level specificity: when queries get granular, like "best niacinamide serum under $30," Sephora gets bypassed in favor of ingredient-focused brands. The retailer brand outperforms the expert brand, which is a GEO liability.

Ulta Beauty

Ulta actually holds its own better than industry insiders tend to expect. Its dual positioning across drugstore and prestige gives AI engines something useful: Ulta can be cited as the answer to both "affordable" and "full-range" queries. That said, its editorial content is thinner than Sephora's, and Perplexity in particular seemed to underweight it in queries that required ingredient-level justification. Ulta wins on practical recommendation queries and loses on expertise-adjacent ones.

Glossier

Glossier's AI visibility is genuinely poor relative to its cultural footprint. The brand has enormous social awareness and loyal customers, but its content infrastructure does not support AI citation well. Sparse ingredient documentation, limited third-party editorial coverage that reads as authoritative, and a product philosophy that resists clinical claims all hurt its extractability. AI engines want to cite sources that say something defensible. "You look like you but better" is not that.

CeraVe

Here is the actual winner of AI beauty search, and it is not a retailer. CeraVe is cited at an estimated 82% rate across relevant skincare queries, making it the most AI-visible brand in the entire test set. Its dermatologist-backed positioning, ingredient transparency, and well-documented clinical claims give AI engines exactly what they need: clear, citable reasons to recommend. The lesson for every other brand here is uncomfortable.

The Ordinary

The Ordinary performs surprisingly well given its no-frills branding. The reason is simple: it publishes detailed ingredient information, usage protocols, and compatibility guides that AI engines can actually extract and cite. It is the most technically documented brand in the mass-prestige space. The gap between its Gemini performance and its ChatGPT performance is interesting; Gemini appears to weight it lower when the query has a luxury signal, which suggests positioning bleed.

NARS

NARS has a brand equity problem in AI search: it is associated with makeup artist credibility but has not translated that into citable content. Its product descriptions are aspirational rather than informative, and the editorial coverage it generates skews toward visual media that AI engines cannot process. For a brand that makes excellent foundations with solid shade matching, its AI citation rate is a quiet disaster.

Fenty Beauty

Fenty's AI visibility is propped up almost entirely by the novelty and cultural impact of its original 40-shade foundation launch, which generated massive written coverage that AI training data captured well. But that launch was in 2017. More recent product launches have less textual depth in the AI citation ecosystem, and Fenty's ongoing content does not consistently replenish the pipeline. It's coasting on legacy coverage.

Why beauty retail struggles with AI visibility

Visual-first content is invisible to AI. Beauty is an industry built on imagery: swatches, tutorials, campaign photography. AI engines cannot cite a swatch. They cite text that makes a claim. Brands that invested heavily in Instagram and TikTok content built enormous awareness and almost no AI-citable infrastructure.

Product claims are legally constrained. Beauty brands, especially those in skincare, operate under regulatory pressure that discourages strong efficacy language. The FDA's guidelines on cosmetic claims create a culture of vague positioning. "Helps skin appear more radiant" is not something an AI engine can enthusiastically cite. CeraVe and The Ordinary found ways around this with ingredient-level claims rather than outcome-level ones.

Retailer and brand identity blur confusingly. When someone asks "where should I buy skincare," Sephora and Ulta compete. When someone asks "what should I buy for my skin," individual brands compete and the retailers become irrelevant. The dual nature of the query space means that Sephora's AI strategy needs to work at both levels simultaneously, and it currently optimizes better for the retailer query than the product-recommendation query.

Influencer coverage doesn't translate. A five-minute YouTube review generates essentially no AI-citable content unless it's also written up somewhere structured. The beauty industry's reliance on influencer channels means a disproportionate share of its earned media exists in formats that AI engines cannot easily draw from, according to BrightEdge's research on content and AI indexing.

The opportunity gap: what underperforming brands are missing

The gap is almost entirely structural. Brands like NARS and Fenty are not losing AI citations because their products are worse. They're losing because their written content is not architected for extraction.

Specifically, underperforming beauty brands are missing:

• Ingredient-level content pages. CeraVe and The Ordinary publish explicit ingredient explanations. NARS does not. AI engines recommend what they can justify.
• Comparison content. Structured content that addresses "X vs Y" queries captures high-intent AI citations. Almost no beauty retailer publishes this seriously.
• Third-party editorial depth. Glossier has fan content. It doesn't have dermatologist citations, clinical study references, or the kind of authoritative third-party coverage that signals reliability to AI training pipelines.
• Structured data. Product schema, review schema, FAQ schema. The Ordinary uses these well. Most prestige brands treat their websites like brochures.

Tools like winek.ai make this measurable, tracking which queries trigger citations and which brands show up consistently across AI engines so teams can diagnose exactly where their content is failing.

Three moves to improve AI visibility in beauty retail

1. Build ingredient and formulation content that's genuinely informative. Not marketing copy. Actual explanations of why a specific concentration of hyaluronic acid in a particular formulation does what it does. This is what CeraVe does, and the citation data reflects it. Every product page should answer the query "why would a dermatologist recommend this."

2. Commission written editorial that lives on indexed, authoritative domains. A partnership with a dermatology publisher or a structured beauty journalism outlet does more for AI citation than 500 influencer posts. Search Engine Land's coverage of AI search citation patterns consistently shows that AI engines disproportionately pull from text-heavy, authority-signaled sources.

3. Implement product and FAQ schema at scale. This is not glamorous work. It is the kind of thing technical SEO teams do quietly while brand teams argue about campaign concepts. But structured data is one of the clearest signals AI engines use to understand and extract product information. Beauty brands that treat schema markup as optional are leaving citations on the table every single day.

Frequently asked questions

Q: Does Sephora actually dominate AI beauty recommendations?

Sephora performs well in AI search for retailer-level queries, appearing in roughly 74% of relevant tests across ChatGPT, Perplexity, and Gemini. However, it does not dominate product-level skincare recommendations, where ingredient-focused brands like CeraVe and The Ordinary consistently outperform it. Sephora's AI visibility is strongest when the query is about where to shop rather than what to buy.

Q: Why does CeraVe outperform prestige brands in AI search?

CeraVe's AI citation rate is exceptionally high because its content infrastructure is built around ingredient transparency and dermatologist endorsement, both of which give AI engines clear, defensible reasons to recommend the brand. Prestige brands tend to rely on aspirational positioning and visual content, neither of which AI engines can easily extract or cite. The lesson is that clinical credibility, even for an affordable brand, translates directly into AI visibility.

Q: What is the biggest AI visibility mistake beauty brands make?

The most common and costly mistake is building brand equity almost entirely through visual and video content while neglecting the written, structured content that AI engines actually index and cite. A brand with millions of Instagram followers and no ingredient explanation pages, no FAQ schema, and no editorial depth on authoritative sites will consistently lose AI citations to a less glamorous competitor that publishes detailed product documentation.

Q: How does Ulta Beauty compare to Sephora for AI recommendations?

Ulta Beauty's AI citation rate sits around 61% compared to Sephora's 74%, but Ulta holds a meaningful advantage on practical and value-oriented queries where its dual drugstore-and-prestige positioning is an asset. Ulta tends to underperform on expert-level skincare queries where ingredient knowledge and clinical backing are the deciding factors. The two retailers serve different query types, and neither has fully optimized for the ingredient-recommendation layer of AI search.

Q: Can a beauty brand realistically improve its AI citation rate quickly?

Yes, but not through shortcuts. The most immediate gains come from publishing ingredient-level content on product pages, implementing product and FAQ schema markup, and securing written editorial coverage on indexed, authoritative domains. These are not overnight wins, but brands that consistently execute on all three tend to see measurable citation improvement within two to three months, particularly on Perplexity and ChatGPT, which weight written authority signals heavily.