While many chief marketing officers are still working out how to fold AI into their workflows, beauty marketers have quietly become the reference implementation. Brands like e.l.f. Beauty, L’Oréal, and Kosé Corp. (owner of Tarte and Decorté) are using AI for data mining, content versioning, chatting with loyalists, and product discovery — and have restructured teams to test and train on it. Mark Abraham, who leads the marketing, sales and pricing practice at Boston Consulting Group, calls them a “lighthouse” for other categories. “What used to be digital attacker brands are now agentic attacker brands,” he said. “They will threaten some of the established brands that move slower to adopt these end-to-end agentic campaigns.”
The data-mining layer came first. Two years ago, e.l.f. built an “elf-luencer” model to answer customer comments online — a “no comment left behind” mission, per Ekta Chopra, chief technology and AI officer. Trained on five years of comments, it generates responses in the brand voice, with human approval before anything posts. More recently, e.l.f. wired a cultural-insights agent into Slack for its consumer insights team; it has just moved from pilot to production. Abraham says AI can compress audience research from six months to two weeks.
The chatbot layer is producing the eye-catching numbers. L’Oréal’s Beauty Genius assistant dispenses recommendations from the company’s portfolio. Tarte’s new gift finder converted shoppers at five to seven times the rate of average site visitors, said Candace Craig Bulishak, Tarte’s president and chief marketing officer — evidence, she argues, that conversational shopping drives purchases rather than just inspiration. Tarte is now working on showing up in LLM searches, the same show-up-early instinct that made it a TikTok Shop adopter three years ago. “Now we are seeing AI-driven discovery becoming an emerging acquisition channel for us,” Bulishak said.
Content creation is where the brands go carefully. E.l.f. knows its community doesn’t want AI-generated models, and the industry has watched the “AI slop” backlash. So the AI does efficiency work instead: generating thousands of assets from the style guide, “versioning” one visual across paid media and email formats, experimenting with motion. L’Oréal runs an internal generative platform on tools from Google, Adobe, and Seedance.
Chopra’s advice for other brands: don’t bolt an agent onto an existing process. “The moment you reimagine the work, you clear the way to see the power of AI and if you don’t do that, that’s where 90% of agentic projects are failing,” she said. Her governance posture rejects both extremes — “let it rip” and lock-it-all-down. “Don’t choose one lane or the other — one can kill innovation and the other can create a governance risk.” New structures are emerging around the tools: e.l.f.’s “tiger teams,” the “super user groups” BCG found inside some CMO organizations, and talk of “marketing engineer” roles.
“The brands that will win are the ones that use technology to become more personal and more relevant and more connected to their consumers,” Bulishak said.
Read more at Ad Age.