AI turns beauty from browsing to ‘agentic’ shopping
AI beauty try-on and conversational assistants describe a new kind of shopping journey where generative models guide product discovery, simulate looks on a user’s face, and connect science-led skincare insights to real-time purchase decisions across digital channels.
L’Oréal’s new partnership with OpenAI is a clear statement: beauty will not wait for commerce platforms to evolve, it plans to build its own AI-first rails. At VivaTech 2026, the company and OpenAI announced a collaboration focused on AI-powered consumer journeys and AI-powered métiers from research to marketing, framed as a foundational pillar of L’Oréal’s “Transformative AI” roadmap. With ChatGPT counting more than 900 million weekly users, the decision to meet shoppers inside an AI interface rather than on a brand site is not a side experiment – it is where the traffic already is. Nearly one in five consumers now use generative AI tools to find information, and beauty discovery is following that gravity.
From virtual makeup preview to conversational advisors
The most visible change for everyday shoppers will be how they try and choose makeup. Maybelline New York is bringing its Makeup Virtual Try-On into ChatGPT, powered by L’Oréal’s ModiFace technology, so users can test looks in real time while chatting with an assistant. This blends AI beauty try-on with natural language search: instead of scrolling through a catalog, you ask, preview, refine and buy – all inside one conversational window. That is a radical shift from static “virtual makeup preview” widgets buried on brand sites.
L’Oréal plans to push richer product signals into ChatGPT so the assistant can answer with its own up-to-date product information when people ask about brands from Lancôme to Kérastase, instead of relying on scattered reviews and forums. A key element of the partnership is how L’Oréal products appear in AI-generated answers, turning search replies into a new form of personalized beauty shopping rather than generic recommendations. Beauty is quietly redefining SEO as “how you surface in AI chats” – and this deal is one of the clearest examples yet.
AI skincare research: mapping the microbiome for the next serum
Behind the glossy try-on tools sits a harder, more consequential use of AI: skincare R&D. L’Oréal is using GPT-Rosalind, OpenAI’s life sciences reasoning model, to map the skin microbiome at unprecedented scale and identify beneficial bacteria that can drive the next generation of natural, effective skincare, starting with La Roche-Posay. This is AI skincare research as competitive weapon, not a marketing gimmick.
If the project delivers, consumers will not only see tailored product suggestions but formulas born from pattern-finding across millions of microbiome data points. That tightens the loop between what science learns and what a chatbot recommends. When a conversational assistant can explain why a specific cleanser fits your skin ecosystem – with the same AI family that helped design the formula – personalized beauty shopping becomes grounded in evidence, not vague skin-type labels. The risk, of course, is overclaim: brands will need to keep explanations clear and testable, or trust will erode as quickly as the novelty wears off.
CreAItech shows AI is becoming the new beauty supply chain
L’Oréal’s CreAItech platform reveals how far AI is being woven into the beauty supply chain for attention. The company is expanding this in-house generative system with OpenAI’s latest models to produce still images and video for social and e-commerce placements at scale, alongside tools from Google, Adobe, Stable Diffusion and others. According to the company, CreAItech has already reduced production costs by 40% and supported the creation of 50,000 marketing assets – a volume impossible under traditional workflows.
This is less about novelty filters and more about control: AI lets L’Oréal route briefs to the model best suited to the job while keeping brand heritage intact. OpenAI’s latest model will power CreAItech to generate imagery and video that align with each brand’s history, augmenting human creativity rather than replacing it. With tech investment around €1.5 billion (about USD 1.98 billion, approx. RM9.1 billion) in 2025 and 70,000 staff trained in AI, the company is treating generative tools as core infrastructure, not a campaign toy.
Beauty sits at the crossroads of generative AI and commerce
The real story in this partnership is not that a beauty giant is dabbling in AI. It is that beauty is becoming a proving ground for “agentic commerce”: AI systems that can understand intent, advise, and nudge purchases across channels. L’Oréal is already testing AI-native advertising in ChatGPT with SkinCeuticals, CeraVe and Garnier, placing messages at the moment of intent rather than around it. At the same time, it is strengthening product discovery signals inside the assistant, so organic answers and paid placements live in the same conversational stream.
As website traffic drifts toward AI-led search, brands that do not design for this surface will be edited out of the conversation. L’Oréal’s move positions beauty retail right at the intersection of generative AI and e‑commerce personalization, where virtual makeup previews, AI skincare research and massive-scale content creation all feed into one thing: a shopping experience that feels less like a funnel and more like a dialogue. That is good news for consumers who want guidance, but it raises a hard question for every other brand category: if mascara can have its own AI agent, what excuse does your product have?
