Key Takeaways: 

• Millions of shoppers are turning to AI chatbots for product discovery and purchases.
• Brands can no longer ignore how they appear in AI-generated responses.
• Victoria Beckham Beauty is auditing AI responses,  optimizing content across Reddit and other key LLM sources.

With 900 million weekly active users on ChatGPT, and millions more on Anthropic’s Claude, Google’s Gemini, and other competing AI chatbots, search and discovery are beginning to migrate to what’s known as agentic commerce. This massive pivot in shopping patterns is leading Victoria Beckham Beauty and others in the industry to overhaul their digital strategies.

Bettina Goesele, Chief Digital Officer at Victoria Beckham Beauty, said that her team conducts regular searches using popular large language models (LLMs) to see what information the chatbots link to. In the earlier days of social media, platforms like Reddit were a popular source of training data for LLMs, so Victoria Beckham Beauty put more effort into engaging that community with accurate brand information. More recently, Google’s video-sharing service YouTube has become popular as a product discovery tool.

During a panel discussion in New York hosted by e-commerce giant Shopify, Goesele said, “Things are moving so quickly with these LLMs that the sources that were being used a few months ago are not going to be used now.”

Victoria Beckham Beauty has begun embracing agentic commerce by making its products available for search and purchase via ChatGPT and Microsoft Copilot. “It’s very early days, we’ve only been on it a month, but we’ve already noticed that the vast majority of traffic coming through these new channels is new customer traffic,” said Goesele.

One-third of US consumers use AI to research general topics, according to recent research published by consulting giant McKinsey, which reported that Gen Z and millennials were far more likely than baby boomers to use AI when shopping for beauty or personal care products. This movement is upending the business model for beauty manufacturers and retailers yet again, who faced similar technology-led disruptions when e-commerce, omnichannel, and social commerce gained traction.

“The way that AI works, I give incredibly detailed prompts, annoyingly detailed, but then AI breaks that down into many different search queries and fans out those searches,” Sandy Jeong, Field Chief Technology Officer at Shopify,  told BeautyMatter about her own personal searches for Korean beauty products. “It’s essentially running dozens of different searches for you and then intelligently collating that information into a recommendation. It’s like everyone has a personal shopper in their pocket now.”

In December 2025, Shopify debuted Agentic Storefronts, which can help brands get discovered and sell directly on chatbots. Within the beauty category, recent agentic commerce announcements include the launch of a ChatGPT app by LVMH-owned Sephora in March and a partnership between Ulta Beauty and Google to build a new AI shopping assistant called Ulta AI in April.

Shoppers using AI chatbots are more highly informed than those using general internet search queries, and are also far more motivated to spend, according to Jeong.

While a significant segment of American consumers start their online shopping journeys on Amazon or Google, nearly one in 10 are now kicking off their search with ChatGPT or another AI chatbot, according to research published by Bain & Company. The management consulting giant also reported that AI accounts for up to a quarter of referral traffic for some retailers. 

Beauty industry experts acknowledge that such swift movement to LLMs can feel intimidating. Renan Goupil, a digital and omnichannel consultant whose clients include beauty giant L’Oréal and fashion conglomerate Kering, said the luxury category has thus far been “a bit shy about AI.”

“Luxury is not selling a product; they’re telling a narrative, and they want full control over that narrative,” Goupil shared during the panel discussion. “They see AI coming and think LLMs could tell their potential consumers nonsense about their brands. It’s very scary.”

Paige, a Los Angeles-based apparel retailer, isn’t quite ready to embrace agentic commerce, but it’s not off their radar. “I think it’s the next channel for us,” Tawny Brown, Senior Vice President of Omnichannel at Paige, told BeautyMatter.

More urgently, Brown said the company has focused on consolidating all physical retail and e-commerce IT systems into fewer vendors, including Shopify and Google. “We got rid of all the noise, all the mess, and really simplified things,” said Brown.

Externally, Brown said her team is formalizing Paige’s generative engine optimization (GEO) strategy, which is when brands create and structure their content so it can be easily discovered on LLMs like ChatGPT, similar to the playbook they had to develop for search engines like Google. “It’s a big undertaking, but if your data is not all organized, you are back to manual redundancies,” said Brown.

“AI is ultimately a data story,” said Jeong. “AI needs really good data to be able to understand your brand, who you are, and to understand your products.”

Paige’s in-house development team is also conducting searches within ChatGPT, Claude, and Gemini to see what responses those LLMs generate. These insights will inform how Paige structures its data on the back end. Internally, Paige is also using AI to help automate some simpler customer service inquiries, like the status of a return or tracking an order. Brown said AI is now handling about 30% of those questions.

Goesele said Victoria Beckham Beauty developed an AI council to explore various AI tools and settled on three for internal use: Claude for analysis, Gemini for workplace tasks, and cloud storage provider Snowflake for data. “Having that data infrastructure in place allows you to spin out agents quickly,” Goesele said.

“Whether you like it or not, your brand is already being talked about in these AI channels,” said Jeong during the panel discussion alongside Goesele and other technologists. “It’s not a question of whether consumers will look for you there. It’s a question of, what are they going to experience about your brand when they do find you there?”

Jeong’s advice for beauty brands just starting to wrap their heads around a GEO strategy: Prioritize the creation of accurate and descriptive product details, pricing, and inventory that AI systems can scrape, and make that information available on public websites, Reddit, and other sources.

Beauty brands should also promote authoritative third-party voices, according to Jeong, which can include media news stories and recommendations from dermatologists and other leading experts. LLMs prefer these sources because they deem them more credible.

Ultimately, for the beauty brands just beginning to wrap their heads around a GEO playbook: consolidate and organize your data, set up the proper governance to understand what information can and cannot be provided to LLMs, set up a team that will monitor how the brand shows up on AI chatbots, and empower that team to consistently update the information that’s being pulled by the LLMs for both accuracy and to inspire purchases.