AI product discovery: from websites to conversations
AI product discovery is the shift from browsing static catalogs and search results to finding, comparing, and buying products through conversational AI systems that understand intent, context, and preferences in natural language. This shift matters because it turns search, recommendation, and checkout into one continuous dialog, dissolving the old boundaries between marketing content, product pages, and shopping carts. Beauty and travel are becoming early proof points: L’Oréal is expanding its CreAItech platform through a foundational partnership with OpenAI to support product discovery and creative production at scale, while Omio integrates OpenAI models across its engineering operations to accelerate travel product development and launch booking interfaces. The message is blunt: if brands keep treating AI as a side tool instead of the primary storefront, they will lose the interface where customers now start their journeys.
L’Oréal’s CreAItech: AI-native retail, not a gimmick
L’Oréal is not experimenting at the edges; it is rebuilding how beauty products are discovered and promoted around AI-led interfaces. The company is expanding its generative AI production system, CreAItech, through a foundational OpenAI integration that gives teams access to the latest models inside its own stack. More importantly, it plans to plug Maybelline’s virtual try-on technology directly into ChatGPT, so shoppers can test looks and shades without ever visiting a brand site. That is conversational commerce in action: the store comes to the dialog, not the other way around. L’Oréal will feed up-to-date product information into OpenAI so AI answers reference its notes when users ask about products, a clear bid to control AI product discovery rather than leave it to scraped reviews and forums. CreAItech has already cut production costs by 40% and powered 50,000 marketing assets—proof that AI-native operating models, not standalone tools, are where the economic payoff sits.
Omio’s travel booking AI: conversational commerce with real stakes
Where L’Oréal is turning AI into a virtual beauty counter, Omio is turning it into a travel agent that never sends you to a clunky booking form. Omio has embedded OpenAI Codex across its entire software development lifecycle—research, planning, coding, testing, review, and maintenance—mandating use as standard practice. Internal analysis shows the technical effort to build specific products now sits at roughly 20% of previous levels, with projects that once took a full quarter for several developers now handled by a single engineer in about a month. That speed is not an internal vanity metric; it fuels faster testing of new travel booking AI features. Omio launched one of the earliest conversational travel booking interfaces in 2023 by connecting OpenAI models to its proprietary inventory. Travelers can describe complex multimodal routes in natural language; the models interpret the request, query real-time transport data, and return directly bookable itineraries across trains, buses, ferries, and flights. This is conversational commerce built on real infrastructure, not a chatbot bolted onto legacy funnels.
Why this is happening now: discovery is leaving traditional e-commerce
The timing is no accident. In beauty, discovery behavior has shifted so sharply that traditional e-commerce is becoming a second stop, not the starting point. Gartner’s Greg Carlucci links L’Oréal’s OpenAI move to declining website traffic and a rise in AI-led search, noting that nearly one in five consumers already use generative AI tools to find information. L’Oréal is positioning the deal around AI product discovery, not just creative speed, and is already testing paid ads in ChatGPT for brands like CeraVe and Garnier. In travel, the baseline experience has long been broken: legacy booking forced users to jump across multiple websites, compare modes manually, and assemble itineraries by hand. Omio’s conversational interface is a deliberate break from that, treating AI as the primary interface layer mediating between users and thousands of transport providers. Both companies assume the same future: discovery and decision-making move inside AI dialogs, while brand sites become supporting infrastructure.
The new playbook: embedded AI, faster cycles, and blurred funnels
The deeper pattern is not about any single model; it is about how brands are rebuilding their systems and organizations for AI-native experiences. Omio built custom internal connectors so OpenAI tools can work directly with proprietary data inside developer environments, which slashes time spent on basic retrieval and lets engineers move straight to execution. Management is now pushing Codex into non-technical functions so the entire organization runs on the same AI foundation. L’Oréal’s CreAItech takes a similar stance: it assumes different models for different tasks, mixing tools from several providers and OpenAI rather than betting on one engine. The company poured about €1.5 billion (approx. RM7.37 billion) into tech in 2025 and aims to equip roughly 10,000 marketers to build assets quickly and switch tools by brief. The larger signal is clear: brands that treat AI as a one-off plugin will be outpaced by those who treat it as the operating system for discovery, trial, and purchase. In the next wave of conversational commerce, the winning storefront is not a website—it is the AI interface your customers talk to first.
