The new normal is that your next customer might never see a search results page. Instead of typing “best budget foundation” into Google, they’re asking ChatGPT or Claude “what’s the best budget foundation for oily skin?” and buying whatever comes back first. That shift is already well underway.
Foundation’s new AI SEO Beauty Guide (July 2026) found that over half of UK AI users now search with AI several times a week, and 27% do so daily.
For beauty brands, this is a visibility question with a completely new rulebook. There’s no bidding system, and your follower count barely moves the needle. So, we ran the numbers to find out what drives beauty AI SEO, and what quietly does nothing.
What we did
We built 20 prompts around the way real shoppers talk to AI, things like “how do I get a glowy base without looking greasy,” then put each one to ChatGPT, Gemini, Perplexity and Claude, three times over.
Every brand named, its position, and every source cited went into the dataset: 1,335 brand mentions and 690 citations.
The full findings live in the guide, but here’s what surprised us most.
The same brands keep winning, and position sticks
AI recommendations isn’t the lottery.
The rank correlation between runs sat at 0.75 to 0.77, so a brand that ranks well one week almost always ranks well the next. e.l.f. was the standout, named in 17 of 20 prompts and appearing as the first recommendation 40 times, three times more than any other brand. The first slot matters more than any other, because that’s the answer people read and act on.
Visibility comes in categories, not one big list
This finding has the biggest implications for beauty AI SEO. There’s no single beauty leaderboard. e.l.f., Maybelline and NYX own the budget and dupe conversation. ILIA, Kosas and Tower 28 own clean and vegan. Charlotte Tilbury, NARS and Giorgio Armani own luxury.
A brand that dominates one is close to invisible in the others, because the models match each question to the associations they’ve built around that specific topic. A brand that tries to win everywhere wins nowhere.
AI is reading the press, not your product pages
The models overwhelmingly cite third parties writing about brands rather than the brands themselves.
Editorial and glossy coverage made up around 44% of citations, the “best X for Y” listicles in titles like Vogue, Allure and Who What Wear. Those roundups map almost perfectly onto how people phrase questions to AI. More unexpectedly, Reddit was the single most cited source in the whole dataset, just ahead of Vogue and Allure, because beauty subreddits generate the peer-to-peer discussion models treat as genuine consumer signal.
What strong beauty AI SEO looks like for your brand
The takeaway is to be selective. ChatGPT and Gemini reward glossy editorial.
Perplexity leans hard on Reddit and retailer pages like Boots.
Claude tends to reflect a content footprint built over years.
Pick the category you can realistically own, land in the roundups your customers are effectively quoting through AI, and build content that answers real questions about skin, ingredients and routines rather than just selling.
Get that right, and AI visibility stops being a mystery and becomes a measurable, repeatable channel you can tie to activity, which is exactly what our services and digital PR teams help beauty brands do.
