• AI visibility

Guide · AI search

How do beauty retailers turn AI product recommendations into sales?

By being the store an assistant names when a shopper asks what to buy and where to buy it, with prices, stock and loyalty value it can state correctly. Beauty shoppers already ask AI for product advice in large numbers, and the big retailers are building their own routes into ChatGPT and Google’s AI. This guide follows the path from an AI answer to a beauty order, and shows what a multi-brand retailer can do to be on it.

By Underneath Team Guide Updated

The short version

  1. Beauty shoppers already use AI: 49% have received beauty product recommendations from AI platforms such as ChatGPT, Claude, Gemini and Copilot (NielsenIQ (opens in a new tab)), and Criteo found (opens in a new tab) that 57% of beauty shoppers who use AI assistants say the recommended products influence what they buy.
  2. The big retailers have moved: Sephora launched an app inside ChatGPT with its loyalty perks, and Ulta Beauty began agentic checkout in Google’s AI Mode and the Gemini app, drawing on its 46+ million members (Google Cloud and Ulta Beauty (opens in a new tab)).
  3. Ulta reports “double the conversion and intent” from shoppers who find its products through Gemini and ChatGPT (Reuters (opens in a new tab)).
  4. Amazon is the rival to beat: it sold an estimated $8.1 billion of beauty products in the first quarter of 2026, up 13%, and says shoppers who used its Rufus assistant to buy spent 80% more over the holidays (Glossy (opens in a new tab)).
  5. Beauty is a discovery category: up to 53% of purchases go to a brand the shopper had not bought in the previous 12 months, which is exactly the moment an AI answer can steer.

Who buys beauty online, and what is a shopper worth to a retailer?

Repeat buyers who replenish and discover, so a shopper is worth a stream of baskets, not one order.

US beauty is large and still growing. Circana reports (opens in a new tab) that prestige beauty retail sales grew 4% to $36 billion in 2025, while beauty sales in mass retail rose 5% to $72.7 billion. In the first quarter of 2026, prestige sales grew 6% to $8.1 billion and mass sales grew 7% to $18.1 billion. Online keeps gaining: McKinsey’s State of Beauty 2025, as summarized by RetailBoss (opens in a new tab), puts ecommerce at 26% of global beauty retail sales in 2024, up from 15% in 2019, and expects 31% by 2030. Circana adds that hair is the only prestige category where online already accounts for the majority of retail sales.

The retailers that sell this category depend on loyal members. Ulta Beauty (opens in a new tab) grew net sales 8.9% to $3.0 billion in its second quarter of fiscal 2026, helped by its acquisition of Space NK, and highlights its Ulta Beauty Rewards loyalty program in describing the business. Members return to replenish favorites, try new launches and redeem points, so the value of winning a new shopper sits in the baskets that follow the first one. Prestige houses face their own version of this, covered in how luxury brands win high spenders.

Where does AI already sit in the beauty shopping journey?

At the research stage, between a social video or a need and the choice of product and store.

The survey evidence is consistent. NielsenIQ reports that more than half of shoppers are interested in AI tools to help manage their shopping and that 49% have already received beauty recommendations from AI. Criteo’s Health & Beauty Pulse, based on surveys across six markets, found that 38% of shoppers already use AI assistants when shopping for beauty and personal care, and 57% of them say AI recommendations influence what they buy. NielsenIQ data cited by Glossy puts beauty-related searches on ChatGPT at over 1 billion a week.

Beauty journeys are long and wide. Criteo found that US shoppers browse an average of 19 makeup products and 12 fragrances before buying. Glossy’s reporters describe today’s path as fragmented: a TikTok, a Reddit post, a store visit to try products, then perhaps a checkout on Amazon. An assistant compresses that research into one conversation.

Across all US retail, Adobe found (opens in a new tab) that AI traffic converted 42% better than other traffic in March 2026, and in June that visitors referred by AI generated 41% higher revenue per visit. Those figures are not beauty-specific, but they match what Ulta reports for its own AI visitors. Outdoor gear shows the same contest between brand and specialty retailer for the store link in an answer, as our guide to outdoor gear shortlists in AI answers shows.

What do beauty shoppers ask AI before they buy?

Category, comparison and “where should I buy it” questions, often with a budget or hair type attached.

We wrote these sample beauty questions ourselves to show the range; none comes from real shopper logs:

  • Category: “Best hair oil for frizzy, color-treated hair under $40.”
  • Comparison: “Is this prestige setting spray worth it, or is there a drugstore alternative that works as well?”
  • Where to buy: “Where can I get this fragrance in the 50ml size in stock today, with the best rewards?”
  • Gifting: “A beauty gift set for a 30-year-old who loves clean fragrance, under $75.”
  • Routine: “I bought this serum; what moisturizer and sunscreen go with it?”
  • Loyalty and value: “Which beauty retailer gives the best birthday gift and points?”

For a multi-brand retailer, the third, fifth and sixth questions matter as much as the first. A brand only has to be named; a retailer has to be named as the place to buy, and then earn the add-on purchase. How brands get named for skin concerns is covered in our skincare brand guide.

How does an AI recommendation become a beauty retailer’s sale?

Through the merchant link: the assistant names a product, then a store, and the order lands where checkout is easiest.

The path has four steps, and the retailer can lose the shopper at each one:

  1. The product is named. The assistant suggests a handful of products. Glossy’s hosts found the size of that list varies widely: Sephora’s assistant gave three or four products, Ulta’s gave 11 or 12, and ChatGPT gave 15. NielsenIQ warns that AI experiences “could reduce that visible choice to only one or two highly personalized recommendations.”
  2. A store is chosen. OpenAI documents (opens in a new tab) that when a shopper clicks a product, ChatGPT may list merchants offering it, and “merchants are ranked based on factors like availability, price, quality, and whether they are the maker or primary seller.” The price shown first typically comes from the first merchant listed. A multi-brand retailer is competing with the brand’s own site and with Amazon for the same product. The brand side of that contest is covered in how ecommerce brands get chosen by assistants.
  3. Checkout happens somewhere. Sephora’s app inside ChatGPT lets shoppers use loyalty points and member perks such as free shipping and samples (Cosmetics Business (opens in a new tab)). Ulta’s agentic checkout lets shoppers complete eligible purchases inside AI Mode and the Gemini app. Retailers still prefer the sale on their own site: Ulta’s head of digital told Reuters it would rather customers complete transactions there. Reuters also reported that OpenAI ended its Instant Checkout tool in March 2026; FashionUnited (opens in a new tab) says the feature was scaled back, and OpenAI’s help page still describes it for some eligible merchants.
  4. The basket and the member. On the retailer’s own site, the first order can grow into a routine, a loyalty sign-up and later replenishment. That is where a retailer’s advantage over a general-purpose assistant lies.

What decides which beauty retailer an assistant sends shoppers to?

Platforms document product data, price and availability; the rest comes from surveys and our inference.

Documented by the platforms. OpenAI says ChatGPT considers structured product data from first-party and third-party providers, such as price and description, and other third-party content, and that review summaries are built from reviews on public websites. Its merchant ranking factors are availability, price, quality and whether the seller is the maker or primary seller. When merchants update prices, OpenAI notes there may be a delay before ChatGPT reflects them. Google and Ulta say the agentic checkout runs on the Universal Commerce Protocol, an open standard for agentic commerce.

Observed by retailers and researchers. Ulta reports double the conversion and intent from AI shoppers. NielsenIQ found that 84% of consumers are more likely to purchase when key product attributes can be easily compared, and 65% find personalized recommendations helpful.

Our inference. Because the brand’s own site is the “maker” of every product a retailer sells, a reasonable expectation is that retailers win the merchant slot on what they add: price and stock, exclusives, gift sets, samples, loyalty value and fast shipping, stated clearly enough for an assistant to read. Our article on what drives AI product recommendations shows that hard facts such as ratings, prices and reviews tend to decide the pick when assistants can see them.

Trust factors specific to beauty retail. Authentic stock from authorized brands, honest reviews, accurate shade and size options, return policies on opened products, and expert advice. Our article on fake reviews and AI recommendations explains why review quality matters to assistants as well as to shoppers.

What does a beauty retailer lose when AI sends shoppers elsewhere?

The first order, the basket that follows it and, often, the loyalty member.

Beauty is unusually open to switching. Criteo found that only 34% of shoppers say they rarely switch from their preferred beauty brands, and up to 53% of purchases come from shoppers buying a brand they had not bought in the previous 12 months. Each of those discovery moments is a chance for an assistant to send the shopper to a different store. Amazon’s own claim that Rufus users spent 80% more shows what a retailer gains when the assistant is its own, and what others risk when it is not.

There is also a data cost. Glossy’s reporting stresses that much of the value of AI shopping for retailers is knowing how a shopper started, compared and chose. When the research happens in someone else’s assistant and the checkout happens on Amazon, the retailer loses both the sale and the insight. We have not seen a study that measures how much beauty revenue retailers lose this way, so treat the size of the loss as unmeasured.

How does GEO work for a beauty retailer?

It makes your store the easiest correct answer to “where should I buy this”; it cannot guarantee being named.

Generative engine optimization (GEO) for a multi-brand beauty retailer usually covers six pieces of work:

  1. Accurate, fast product data. Keep prices, sizes, shades and stock current in the feeds assistants use, so the merchant list can name you. In our study of AI pricing accuracy, which covered software rather than beauty, only 61.9% of the plan prices assistants quoted were fully faithful to the official page.
  2. What only you offer, in plain words. Exclusives, gift sets, samples, loyalty points and services such as in-store testing are reasons to choose a retailer over the brand’s own site; state them on product pages, not only in banners.
  3. Answers to routine and comparison questions. Expert guides on routines, hair types and fragrance families give assistants something to cite. Our article on the product content AI shopping assistants prefer explains what kind of copy helps.
  4. Reviews that hold up. Encourage detailed, verified reviews on product pages; assistants summarize public reviews.
  5. Brand-partner consistency. Make sure your product names and descriptions match the brand’s, so an assistant recognizes it is the same product you sell.
  6. Measurement by question and surface. Track category, comparison and where-to-buy questions separately across ChatGPT, Gemini, Perplexity, Copilot and Google’s AI features, and connect them to orders, basket size and loyalty sign-ups. If a new launch you carry is missing from answers, why ChatGPT misses new products walks through the usual causes.

Which beauty retail questions can’t the evidence answer yet?

It does not show how often AI answers send beauty shoppers to one retailer rather than another.

The beauty surveys measure use and influence, not which store a shopper chose. Ulta’s “double the conversion” comes from the company itself, and Amazon’s Rufus figure is Amazon’s claim to Glossy. The Amazon beauty sales estimate comes from an agency’s data, not Amazon’s filings. Adobe’s figures cover all US retail. Glossy’s comparison of assistants was one reporter’s test, not a study. OpenAI and Google document what they consider, but not how they weigh one retailer against another for the same product. Checkout inside AI tools is also in flux: Reuters says OpenAI ended Instant Checkout in March 2026, while FashionUnited says it was scaled back.

Where should a beauty retailer start?

Start by checking which store assistants name for your best-selling categories, and why.

A useful first step is an audit of the category, comparison, routine and where-to-buy questions your shoppers ask, across the main assistants, with the merchant each answer names and the price it quotes. That shows where you lose the sale to a brand site or a marketplace, and which fixes would move the most orders. If you would like us to run that audit with you and plan the work, tied to orders, basket size and loyalty sign-ups, get in touch with our team. On our generative engine optimization service page you can see how that audit leads into feed accuracy, clearer loyalty and exclusive details, and routine guides an assistant can cite.

Frequently asked questions

Do beauty shoppers buy what AI recommends?

Many say it influences them. In Criteo’s survey, 57% of beauty shoppers who use AI assistants said recommended products influence what they ultimately buy.

Should a beauty retailer sell inside ChatGPT or Gemini?

It depends on the trade-off. In-chat checkout removes friction, but retailers such as Ulta prefer purchases on their own sites, where they keep the customer relationship and data.

Why would an assistant send a shopper to the brand’s site instead of ours?

OpenAI lists “whether they are the maker or primary seller” among merchant ranking factors. Retailers need to make price, stock, perks and services clear to compete.

Does Amazon have an advantage in AI shopping for beauty?

It has its own assistant, Rufus, and Amazon says Rufus users spent 80% more. Other retailers are partnering with ChatGPT and Google to compete.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work