AI is becoming beauty's new storefront

Whether someone asks ChatGPT for the best mascara for sensitive eyes, Gemini for skincare for rosacea, or Perplexity for the best long-lasting foundation, AI is increasingly becoming part of the purchase journey. The brands these models recommend won't just influence consumer decisions—they'll define the next generation of beauty discovery.

That's why we wanted to understand what these AI models are actually recommending.

At Foundation, we analysed responses across ChatGPT, Gemini, Perplexity and Claude to identify which beauty brands consistently appear, where those recommendations come from, and, ultimately, what separates the brands AI trusts from those it overlooks.

The results challenged a few assumptions.

The first is that AI simply recommends the biggest brands.

It doesn't.

Some of the world's most recognisable names appear consistently, of course, but so do brands like Tower 28 and Saie. They don't have the budgets of MAC or Benefit, yet they repeatedly earn recommendations alongside them.

To me, that's one of the most encouraging findings from the research because it shows AI isn't simply measuring share of voice. It's measuring credibility.

These brands have built passionate communities that genuinely advocate for them. They're talked about, reviewed, debated and recommended by real people, and AI models recognise those signals.

The same pattern appears elsewhere.

Aquaphor and O'Keeffe's consistently perform well for highly specific searches such as lip care. Makeup by Mario and Natasha Denona compete comfortably alongside global brands like L'Oréal and Charlotte Tilbury.

What they all have in common isn't necessarily scale.

It's authority.

Some have clinical credibility. Others have category expertise. Others have become synonymous with a particular type of product. They own a space rather than trying to dominate every space.

That's a lesson many brands should pay attention to.

The second thing that stood out was where AI is actually getting its information.

Editorial coverage dominated our findings, accounting for 44% of all citation sources. Brand-owned content represented just 7%.

That tells us something important.

AI trusts what other people say about your brand far more than what your own website says.

The publications appearing time and time again were remarkably consistent: Vogue, Allure, Who What Wear, Cosmopolitan, Byrdie and Marie Claire.

For beauty marketers, that's a significant shift.

We've spent years talking about owned content strategies, but in an AI world, third-party validation has become one of the strongest visibility signals available. PR doesn't become less important because of AI—it arguably becomes more important.

The other standout finding was Reddit.

Across every model we analysed, Reddit emerged as the single biggest individual citation source.

That shouldn't really surprise us.

Beauty has always been driven by recommendations. Consumers have long trusted real experiences over advertising, and AI appears to be following exactly the same logic.

Authentic community discussion has become machine-readable trust.

Perhaps the most fascinating part of the research, though, was how differently each AI platform behaves.

ChatGPT leans heavily towards premium editorial publications. Brands with strong PR strategies perform particularly well because the model consistently references established beauty media.

Perplexity is much stricter. It recommends fewer brands, draws heavily on Reddit and retailer websites like Boots, and appears to reward brands with strong community engagement and retail presence.

Claude surprised us the most. It surfaced 24 brands that appeared nowhere else in our research and references a much broader mix of niche publishers, specialist blogs and SEO-driven content. For smaller beauty brands, that potentially creates opportunities that don't yet exist elsewhere.

Gemini takes a different approach again, placing much greater emphasis on official product information, structured data, clinical studies and authoritative organisations. Editorial still matters, but so does the quality of your product information and the trust signals surrounding it.

The important point isn't that one model is better than another.

It's that AI visibility is no longer one single discipline.

Brands will increasingly need to understand how different models build trust and tailor their digital presence accordingly.

The brands winning today all have three things in common.

They own a category instead of trying to compete everywhere.

They earn coverage from the publications AI consistently trusts.

And they build genuine communities that advocate for them.

None of those things can be bought overnight.

As AI becomes another front door to product discovery, visibility will belong less to the loudest brands and more to the most credible ones.

For beauty marketers, that's both the challenge and the opportunity.

Because while algorithms are changing, trust is still what determines who gets recommended.

‍