Estée Lauder built 240 custom GPTs to mine 75 years of beauty data

Estée Lauder Companies isn’t just dabbling in AI—it’s gone all in. The prestige beauty giant has deployed more than 240 custom GPTs across its portfolio of brands like Clinique, La Mer, and Aveda, all running on ChatGPT Enterprise. The immediate driver? Protecting what SVP Raheel Khan calls their “most valuable asset, which is 75+ years of data,” while speeding up the kind of grunt work that used to eat entire afternoons.

The company’s approach is refreshingly structured for a space that often gets lost in vague promises. Their internal GPT Lab, a cross-functional group, runs development like a sprint. A business user, a subject matter expert, and a technical lead collaborate in rapid cycles—designing, prepping data, building, and testing. Kingsuk Chakrabarty, Director of Enterprise Architecture, says they prioritize GPTs based on a simple matrix of high value and quick build times. No sprawling year-long IT projects here.

Two tools stand out. The Fragrance Insights GPT lets Yuan Zhan’s team ask complex questions in plain English and instantly comb through massive consumer survey datasets, a task that previously required hours of manual cleaning. The Clinical Trial Data GPT extracts efficacy stats—like the immediate moisturization improvement percentage for a specific serum—from thousands of reports with a single query. Even copywriting and vendor snapshots have their own GPTs now.

The whole thing started with a company-wide call for ideas that drew over a thousand employee submissions. That organic demand, rather than a top-down mandate, is what makes this rollout notable. When lab members can move from a two-page use case brief to a deployed GPT with a user guide in ten weeks, you’re not just talking about efficiency gains. You’re watching a legacy company rewire how its institutional knowledge actually gets used, one sprint at a time.