On March 24, 2026, Sephora made its first move in AI commerce: the launch of its app inside ChatGPT. The launch gave customers access to personalized beauty recommendations and curated advice. Customers could also link their Beauty Insider account directly to the ChatGPT experience. More than 80 million active Beauty Insider members worldwide could connect their profile, preferences, and loyalty rewards to the AI interface from day one.
On June 3, 2026, Sephora became the first prestige beauty retailer inside Google Agentic Checkout. Shoppers could discover products, ask detailed beauty questions, build a routine, and complete checkout within Google without leaving the platform at any point.
On September 2, 2026, three days ago, Sephora announced its debut on TikTok Shop. The Sephora Drop Shop pilots in the United States from September 19, bringing exclusive monthly product drops, creator-led content, and TikTok LIVE shopping events where customers discover and buy without leaving the app. TikTok Shop’s beauty and personal care category generated $2.08 billion in sales in the first half of 2026 alone, making it the largest category on the platform.
Three AI commerce channels. Six months. ChatGPT for the shopper who wants expert advice. Google Agentic Checkout for the shopper in purchase mode. TikTok Shop for the shopper who discovers while consuming content. Sephora is not closing one loop. It is closing three simultaneously. In the category everyone said was too tactile for conversational commerce.
What Sephora Did and In What Order
The sequence matters more than the individual announcements. Sephora did not launch AI commerce by enabling checkout inside ChatGPT. They launched it by making their expertise and their loyalty program available inside the AI interface first. Anca Marola, Global Chief Digital Officer of Sephora, described the positioning precisely: “At Sephora we are passionate about assisting our customers with the best beauty advice and curation, wherever and whenever they want. Today this means piloting the Sephora app on new intelligent channels such as ChatGPT.”
The discovery layer came first. Recommendations, personalization, and loyalty integration launched in March. That step was not a pilot waiting for the real launch. It was the foundation that made the transaction layer viable. By the time Google Agentic Checkout launched in June, Sephora had established something more important than a checkout button inside an AI platform. It had established that the AI interface could give beauty advice as good as or better than the search experience. This was possible because it had access to the customer’s own preference data through their Beauty Insider profile.
Consequently, the checkout followed naturally. The customer had already received a recommendation they trusted. The loyalty rewards they expected were already connected. The friction of leaving the platform to complete the transaction was the only remaining barrier. Google Agentic Checkout removed it. TikTok Shop then added a third behavior: discovery through entertainment, with purchase available in the same moment of engagement.
Why Beauty Was the Hardest Category and Why That Matters
The Tactile Problem Every Analyst Named
Beauty has always been the category cited as most resistant to digital commerce. Foundation shade matching requires seeing the product on the skin. Fragrance cannot be transmitted through a screen. Skincare routines require understanding specific conditions that vary by person and by season. The in-store beauty advisor, the physical sample, and the sensory trial have been the core of Sephora’s model since the company launched. When analysts evaluated which retail categories would be slowest to adopt AI-native commerce, beauty appeared near the top of every list.
The Data Architecture That Solved It
Sephora had already solved the tactile problem before the AI commerce announcements. Color IQ matches foundation shades to individual skin tone. Skin IQ assesses skin type and concerns, while Fragrance IQ maps scent preferences through a structured quiz. Moreover, Beauty Insider captures purchase history, brand affinity, and product ratings across every transaction since the account was created. The 80 million active members in that loyalty program represent one of the richest first-party preference datasets in retail.
The result is that when a customer asks ChatGPT to recommend a foundation for dry skin and links their Beauty Insider account, the AI does not start from zero. It starts from years of that customer’s color matches, purchase patterns, and product ratings. That is a better recommendation than a search engine or a first-time beauty advisor can produce. The AI did not remove the tactile barrier. The data infrastructure Sephora had been building for years already addressed it.
Three Channels, Three Shopper Behaviors, One Strategy
The ChatGPT, Google Agentic Checkout, and TikTok Shop launches are not three separate commerce experiments. They represent a deliberate map of where different types of shoppers are making purchase decisions in 2026 and a systematic decision to be present and transactable in all three.
ChatGPT: The Expert-Advice Shopper
The shopper who opens ChatGPT to ask a beauty question is looking for expertise. They want a recommendation they can trust. By integrating Beauty Insider from day one, Sephora made the ChatGPT experience more personalized than a generic search. The recommendation is informed by the customer’s own purchase history, skin type, and color profile. That shopper does not need to be convinced of Sephora’s credibility. They need to be met where they already are.
Google Agentic Checkout: The Purchase-Ready Shopper
The shopper using Google Agentic Checkout has already decided they want to buy. The friction they are trying to eliminate is the redirect. As I described in the AI retail discovery and transaction analysis, the friction of leaving the conversation is the primary conversion killer in AI-driven commerce. Google Agentic Checkout removes that friction entirely. Sephora being the first prestige beauty retailer in that environment means they captured that shopper before any competitor did.
TikTok Shop: The Discovery-While-Scrolling Shopper
The TikTok shopper is not in purchase mode when they open the app. They are consuming content. The Sephora Drop Shop converts that content consumption into a purchase moment through exclusive drops, creator content, and TikTok LIVE events. Artemis Patrick, president and CEO of Sephora North America, described the rationale directly: “We already have an incredibly active beauty community on TikTok, so we are excited to take the next step and pilot the Sephora Drop Shop on TikTok Shop.” Furthermore, the $2.08 billion in beauty TikTok Shop sales in the first half of 2026 confirms that this shopper behavior is already at scale. Sephora is not creating a new behavior. It is entering a channel where the behavior already exists.
The Three Things Sephora Got Right That Most Retailers Are Still Getting Wrong
They Integrated Loyalty Before Completing Checkout
The critical question in AI commerce is not whether you can sell inside a third-party platform. It is whether you retain the customer relationship when you do. By making Beauty Insider the centerpiece of the ChatGPT experience, Sephora ensured that every interaction inside the AI interface fed back into their own loyalty data. The customer relationship stayed with Sephora even as the discovery and eventually the transaction moved to OpenAI’s and Google’s infrastructure.
By contrast, a retailer who enables checkout inside an AI platform without loyalty integration is completing a transaction without building a relationship. The sale happens. The customer data goes to the platform. The next AI recommendation will not include the purchase history that should inform it. That history is not connected to the AI’s memory of that customer.
They Sequenced Discovery Before Transaction
The 72-day gap between the ChatGPT launch and the Google Agentic Checkout launch was not a delay. It was a deliberate sequencing decision. Sephora launched in ChatGPT without checkout. That meant the first months of the experience were about building trust in the AI recommendation, not completing transactions. By the time Google Agentic Checkout launched, Sephora had already established that AI-native beauty advice worked. The transaction layer completed a journey the discovery layer had already made trustworthy. Most retailers attempt to do both simultaneously. They ask customers to trust a recommendation and complete a purchase in the same moment. The recommendation has had no time to prove itself.
They Solved the Personalization Layer Before Moving to AI Channels
The Color IQ, Skin IQ, and Fragrance IQ tools were not built for the ChatGPT launch. They were built over years as part of Sephora’s core digital experience. By the time the AI commerce announcements arrived, Sephora had a structured data layer that could express each customer’s preferences in a format the AI could use. Most retailers attempting to enter the AI commerce layer use product catalogs optimized for keyword search. Their customer data sits in systems that cannot feed a conversational AI in real time. As I described in the Generative Engine Optimization analysis, the product data quality and the customer data architecture determine how well the AI recommendation performs. Sephora had both before the launches.
What This Means for LatAm Retailers
Sephora operates more than 2,500 stores in 35 countries and 36 e-commerce websites. Its LatAm presence is significant. However, the lessons from Sephora’s AI commerce execution apply most directly to LatAm retailers who are earlier in the loyalty program maturity curve. The quality of your AI commerce experience is directly proportional to the quality of your customer preference data. Furthermore, a loyalty program that captures purchase history, product ratings, and category preferences is not a marketing tool. It is the data infrastructure that makes AI-native personalization possible.
Additionally, TikTok Shop is already active in several LatAm markets, and the social commerce behavior the Sephora Drop Shop is designed to capture exists in the region through TikTok, Instagram Shopping, and WhatsApp commerce. LatAm retailers who are building loyalty programs and structured product data in 2026 are building the foundation that will make their AI commerce execution in 2027 and 2028 possible across all three channels Sephora is now operating simultaneously.
The Honest Question Every Retailer Should Answer
What Sephora Had Before the Launches
Sephora closed the full AI commerce loop in six months across three channels. But the work that made those six months possible took years. The Beauty Insider program, the Color IQ and Skin IQ tools, the structured product attribute data, and the AI-driven demand forecasting infrastructure were all in place before the ChatGPT app launched in March. The speed of execution was possible because the foundation was already there.
What Your Organization Has Before Its Launch
Beauty was the category most analysts said would be last to work in conversational AI commerce. It closed the full loop in six months across three channels. The reason was the data infrastructure underneath it. What does that imply for your category? Three assets determine readiness: customer preference data, loyalty integration, and product data architecture. When those three are in place, the AI recommendation becomes trustworthy enough that a customer will complete a transaction on the strength of it.
If those three assets are not yet in place, the right action is not to delay the AI commerce strategy. It is to start building them now. The retailers who have them when the AI commerce infrastructure matures in their category will close their own loop quickly. The ones who build the foundation and the launch simultaneously will take much longer. They will produce a worse customer experience and wonder why the results did not match the announcement.
Sephora showed that the category does not determine the speed of execution. The data infrastructure does. ChatGPT, Google Agentic Checkout, and TikTok Shop are three different doors into the same store. The retailer who has the loyalty data, the product attribute structure, and the personalization layer ready will open all three doors quickly. Every other retailer will open them slowly, and their customers will already have built habits somewhere else.
If you are evaluating your customer data architecture, loyalty integration strategy, or product data structure for AI commerce readiness, connect with me here or reach me on LinkedIn. I am happy to walk through the framework we use across the U.S. and Latin America.
Adriana Rivas is a retail technology executive and AI strategist, and the founder of a U.S.-based hardware company specializing in self-service kiosks, POS systems, electronic shelf labels, and digital signage deployed across the United States and Latin America. She is the award-winning author of How to Implement Self-Service Without Failing (Amazon #1 Hot New Release, Silver Nonfiction Book Award 2025) and recipient of the Gold Stevie® Award, Thought Leader of the Year 2026. She is also recognized by Thinkers360 as the #7 Global Thought Leader in Retail and a Certified Master Expert in Retail.