L'Oréal debuts at WAIC with 3CE GENBA generative AI beauty advisor, pushing AI across the beauty chain

L'Oréal China made its first appearance at the World AI Conference (WAIC) in Shanghai, bringing a batch of AI applications rather than just beauty products. The centerpiece was 3CE GENBA, a generative AI beauty advisor aimed at offline consumers, with popular tech blogger He Tongxue on site to demonstrate how AI turns consumer needs into solutions.

3CE GENBA is L'Oréal's first generative AI application for offline consumers. Built on the brand's knowledge base and L'Oréal AI models, it lets shoppers describe their needs in plain language and receive tailored product and makeup recommendations without understanding the company's complex product system. The booth also featured PPD, an interactive professional hair-styling magic mirror that lets consumers preview different hairstyles and makeup looks before committing to a change.

Huang Bingrong, L'Oréal's chief growth officer for North Asia and China, said consumers no longer wait for brands to provide answers — they actively describe needs, compare options and let AI join the decision. That shift forces brands to become more professional, precise and personalized, and underpins L'Oréal's "AI for Beauty" philosophy: AI's value is not to define beauty for consumers, but to understand differences and expand choices.

The push extends into product development. L'Oréal's Hair Digital Twin model moves part of the trial-and-error of formula screening into the digital world, letting researchers simulate how different formulas affect different hair types before selecting candidates for real experiments — at up to three times the speed of a traditional lab. The Longevity AI Cloud, built on computational biology, can analyze more than 260 longevity biomarkers.

On the content side, L'Oréal launched CreAItech, a generative AI beauty content lab debuting in North Asia, using a three-step workflow: AI first generates ideas across colors, materials and shapes, brand teams then inject consumer insight, brand spirit and aesthetic judgment, and AI finally absorbs repetitive administrative work. Huang warned that without guidance, AI-generated output tends toward homogeneity.

Delivery and operations are changing too. L'Oréal uses POC and small, fast iterations — validating direction with small tools before refining based on business feedback — and its AI customer service accuracy climbed from the low 80% range in early internal testing to above 90%. The company has also introduced AI quality control at its Shangmei factory in Suzhou and built an intelligent packaging creative center and an advanced intelligent operations center.

The deeper change is organizational. In the past month, more than 70% of L'Oréal China employees took an AI capability self-assessment, and more than 73,000 employees worldwide have completed generative AI training. The Beauty Tech Agentic Platform, built over the past year, gives staff a unified technical base that can quickly integrate external large models while applying technical guardrails to keep AI applications responsible and trustworthy.

During WAIC, L'Oréal also signed a strategic cooperation memorandum with Volcano Engine to bring in local AI ecosystem capabilities, and launched its second "Young Talent · Tech Fun Song" beauty tech hackathon to draw university-age developers into building AI solutions around real beauty needs.

Taken together, what L'Oréal brought to WAIC is a systematic rewrite of the beauty industry's people, products and places by AI. Huang predicts that after shelf e-commerce and social e-commerce, the next consumption scenario could be "agent e-commerce," where brands must make product information, brand knowledge and service capabilities accurately recognizable and callable by AI agents — a question every traditional consumer brand will eventually face.