Unilever uses AI to cut beauty product development time from months to days

Unilever cuts product development time from months to days using AI

Unilever's Beauty & Wellbeing division is using AI for product development to compress formulation cycles and respond to consumer trends faster than competitors. The company analyzes over 1,000 external data sources monthly-social media, search queries, retail data, competitor activity-to identify what consumers actually want.

The speed gains are substantial. Concept-to-R&D-brief time has dropped from months to days. Formulation cycles that once required five or six rounds now need one or two. Claims generation is 75% quicker. Consumer insight analysis happens 60% faster than before.

How the process works

Unilever's scientists combine trend data with the company's own R&D databases-ingredient libraries, formulation trials, sensory tests, consumer studies, packaging specs-to design products that match what people are searching for. An AI tool called the "R&D Assistant" connects over 150,000 scientific documents spanning more than a century of research, letting scientists query findings in natural language across multiple countries.

The company also uses virtual cohorts: AI-generated sample groups built from Unilever's microbiome datasets. These digital representations let researchers test how specific demographics-defined by age, skin type, hair type, location-would respond to new formulas and claims before physical testing begins. The system can analyze around 2,500 virtual subjects at once, cutting both time and cost from early-stage R&D.

This doesn't replace real-world testing. Instead, it accelerates the research phase by letting scientists access and explore vast datasets faster than manual analysis allows.

Products already in market

Pond's Skin Institute's Hydra Miracle range emerged from this process. The core ingredient, Cera-Hyamino™ technology, combines pro-ceramides, hyaluronic boosters and GAP amino to strengthen skin barrier function. Clinical results showed 78% more hydrated skin from day one and a 100% hydration boost. Without AI analysis of microbiome data, scientists said they wouldn't have spotted the ingredient connections on their own.

Dove's Damage Therapy range followed a similar path. Scientists used robotic tools and advanced measurement to examine over 100,000 data points on hair properties at nanoscale. The research identified how formulas could penetrate hair fibers effectively. Unilever filed five patents for the active ingredients and rolled out Damage Therapy across 35 countries in Q3 2025, contributing to Dove's double-digit growth that year.

What this means for product teams

For Unilever's 4,500 researchers, the shift goes beyond speed. Jason Harcup, Chief R&D Officer of Beauty & Wellbeing, said structured data, AI and human creativity are redefining what's possible in R&D. The combination lets teams move from spotting a consumer trend to launching a science-backed product at a pace that was previously impossible.

The approach addresses a real market pressure: 87% of beauty product discovery now happens on social platforms, and 91% of US adults aged 30-54 say they're more ingredient-aware than before. Consumers want proof, and they want it fast. AI-driven product development lets companies keep pace with that demand.

For product development professionals, the lesson is practical. AI data analysis tools can compress research cycles and surface insights humans might miss-but only when they're connected to real business data and paired with expert judgment.