‘GEO’ Is Beauty’s New ‘SEO’

Beauty brands want in on “the ChatGPT glow-up.”

From Reddit to TikTok, influencers and beauty shoppers are describing the myriad ways that AI is dictating their beauty purchases, from creating entire skincare routines to advising which makeup shades go with their skin tone. Embracing the “ChatGPT glow-up” trend, they’re uploading photos and asking detailed questions to let the AI platform plan their entire beauty regimen.

With an estimated 700 million weekly users on ChatGPT alone, beauty companies are ramping up their investment in trying to make large-language model (LLM) AI platforms recommend their products — and get the facts right. They’re in largely uncharted territory. A more complicated practice than SEO, the process of refining online information about a product to make it more likely to be picked up by an LLM is so new it hasn’t been universally named — the top two terms are “generative engine optimisation” (GEO), and “answer engine optimisation” (AEO).

“This is the beginning of the beginning,” said Anurag Banerjee, co-founder and chief executive of consumer insights firm Quilt AI, adding that most beauty companies are early in their understanding of GEO. “Honestly, nobody’s doing this even remotely well.”

The stakes are high, however, as emerging data indicates that LLMs are now a significant part of many shoppers’ path to purchase. According to February 2025 data from AI e-commerce startup Nosto, 52 percent of Gen-Z shoppers prefer asking AI for skincare recommendations over Google or Amazon. As a result, brands are upping their marketing spend on AI search, changing their websites and focussing on credible sources they’re told can boost their chances of being mentioned. They’re also contending with inaccuracies and a public that doesn’t often care to fact-check.

“The [purchase] funnel has been disrupted, and ChatGPT is at the beginning of that funnel,” said Tara Loftis, the global president of skincare at Galderma, which has adopted GEO for its dermatological brand Cetaphil that will be rolled out to the more premium label Alastin next.

Estée Lauder Companies is exploring GEO for channels including ChatGPT, Gemini, Perplexity and others, confirmed a spokesperson for its technology and AI initiatives. Its February 2025 earnings call said AI search visibility was a priority for Clinique and Origins. L’Oréal has been testing generative AI discoverability in ChatGPT and Google since 2024, according to its earnings call at the end of last year.

“This is not just a ‘nice-to-have’. It’s an essential,” said Loftis.

Visibility Through Credibility

The use of ChatGPT for beauty recommendations has gone mainstream, with magazine articles titled “I Let ChatGPT Decide My Skin-Care Routine for a Month” and the ChatGPT glow-up gaining steam on social media. In a May 2025 TikTok post that earned over 1 million views, influencer Marina Gudov said she uploaded photos of herself and every makeup product she owns to ChatGPT, asking it to analyze all the colours and tell her what to keep and throw out. Over on Reddit, a member of the group “30-Plus Skincare” proudly showed off before-and-after photos of her cleared acne eight months after asking ChatGPT to design her skincare routine.

“Generative” queries, like those asking for general recommendations on a particular skin condition — and that might yield a variety of answers, including products, ingredients and therapies — make up 37 percent of ChatGPT searches, according to AI marketing consultancy Profound. Galderma’s internal research found over half of Gen-Z consumers use ChatGPT as their first point of recommendation for skincare.

Banerjee said that about 20 percent of beauty queries are asked by “GPT pros” who “tweak their prompt, change it and edit it. By the 10th version of the prompt, they’ve got what they need.”

Only 6.1 percent of questions are “transactional,” or asking about specific product names, according to Profound. But ChatGPT’s addition of shopping links in May has started to change that. The UK trade group Institute for Practitioners in Advertising reported in its IPA Bellwether report that consumers’ use of AI for shopping has increased by 47 percent so far this year.

To reach these consumers, experts say brands should focus on non-brand domains, industry publications and detailed content to rank higher. For the Estée Lauder Companies, this means increasing its third-party citations on both established publications and sites like Reddit and YouTube.

To increase chances of brands turning up when asking ChatGPT to list out personalized routines, the expert consensus is that traditional media is crucial. The datasets that power major LLMs like ChatGPT pull from media sources, such as articles in magazines or scientific journals, but tend to ignore influencer reviews on TikTok and Instagram. Experts say that gaming the system, as many do with SEO through backlinks on Google, is not possible with AI.

“With this, you can’t really cheat,” said Sally-Anne Stevens, the founder of PR agency B. The Agency, which is working with tanning brand Bondi Sands and K-beauty label Beauty of Joseon on GEO strategies.

But non-media sources like Reddit and YouTube also have a major impact.

“Reddit is becoming a very important source of influence when it comes to GEO,” said Benjamin Lord, the founder of brand accelerator Antidote, which works with labels like Nars and conglomerates including Unilever and L’Oréal. According to Antidote’s data, Reddit is the top source of citations on ChatGPT answers, showing up in 40 percent of them, followed by Wikipedia.

How to Optimise Accuracy

For beauty brands, showing up in recommendations is half the battle. The use of unvetted sources like Reddit posts can lead to errors, while even fact-checked articles are subject to AI “hallucinations.”

When Profound conducted a test that fed a series of skincare questions into generative AI platforms including ChatGPT and Perplexity, they discovered that the answers only yielded 50 percent factual accuracy.

“There were times when it said a moisturiser was a sunscreen,” said Josh Blyskal, Profound’s head of answer engine optimization strategy and research. “It was like, ‘Oh yeah, it has SPF. Don’t worry.’ It didn’t.”

Seekers of skincare and makeup advice are not promised factually accurate answers, but a study by CivicScience found that 54 percent of U.S. adults say AI brand recommendations are as or more trustworthy than Google results. They’re also not necessarily checking sources: a July 2025 Pew Research survey found that only 1 percent of people clicked direct citations on Google’s AI results, and only 8 percent click the Google links under it, nearly halving the percentage of people that click links on AI-free Google searches.

To improve accuracy, AI-focused strategy firms are offering website overhauls with redesigns to remove ChatGPT-unfriendly JavaScript coding. Cetaphil undertook multiple changes, including rewording site content, product descriptions and even photo descriptions to up its chances of being picked up and accurately described by generative AI. ELC is also prioritising an effort to ensure product descriptions show up accurately in LLMs.

As the technology rapidly evolves, and efforts to manipulate it become as sophisticated as today’s SEO economy, consumer trust in GEO recommendations will be tested. Paid marketing — which is currently absent from LLMs like ChatGPT — may soon influence generative answers. OpenAI has been considering how to incorporate advertising, according to a 2024 Financial Times report.

“People want to hear a genuine recommendation that’s not just from a brand -owned advertisement,” said Loftis. “They want to make sure that they’re not being led to purchase by something that’s paid. And AI seems to be the tool of the moment right now.”

Sign up to The Business of Beauty newsletter, your complimentary, must-read source for the day’s most important beauty and wellness news and analysis.