Unilever's Beauty & Wellbeing division is embedding artificial intelligence across its entire product development pipeline, slashing formulation cycles from five or six days to just one or two and cutting claims generation time by 75%. The EUR 12.8 billion business unit, home to brands like Dove, Vaseline, and Dermalogica, is betting that AI-driven speed and precision will define its next generation of beauty products.
Jason Harcup, chief R&D officer at Unilever Beauty & Wellbeing, said recent advances in generative AI and large language models have "significantly expanded what's possible." The company now uses AI to track consumer sentiment, buzz, and search terms in real time, analyzing more than 1,000 external data sources each month. "This always-on intelligence feeds directly into the product development process, allowing Unilever's R&D teams to act on trends quickly," Harcup said.
AI-designed products already on shelves
Two recent launches show how the approach works in practice. The Pond's Skin Institute Hydra Miracle range uses Cera-Hyamino technology developed by mining 30 terabytes of microbiome data with AI-led digital tools to identify effective ingredient combinations. Dove's Damage Therapy range draws on more than 100,000 data points on hair properties, collected through advanced measurement tools, robotics, and AI analysis.
Vaseline Originals took a different route. The campaign launched two products - the Vaseline Brow Tamer and Vaseline All-in-One Primer & Highlighter Jelly - directly on TikTok Live, tapping into what Harcup described as "enduring beauty rituals emerging across digital communities." The company's AI tools had flagged the trend signals early enough for R&D teams to move from concept to brief in days rather than weeks.
Virtual testing at scale, not as a replacement
Unilever has built an AI-powered agent that connects more than 150,000 scientific documents from decades of internal research. Its AI-created virtual cohorts can analyze microbiome datasets from roughly 2,500 subjects simultaneously. The goal is not to replace real-world testing. Harcup stressed that virtual cohorts are "complementary to real-world testing" and always overseen by in-house scientists. They let researchers understand potential outcomes earlier, but human validation remains essential before any product reaches the market.
The company has trained more than 40,000 employees globally through AI learning programs and hands-on workshops. "We are committed to the responsible and ethical use of AI and have clear responsible AI principles in place, maintaining human oversight and accountability, and ensuring scientific expertise remains at the centre of decision-making," Harcup said.
Shifting scientists toward higher-impact work
Automation of repetitive tasks is freeing Unilever's 4,500 researchers to focus on creative problem-solving. "AI isn't just a time-saver; it's changing how we discover, collaborate and innovate," Harcup said. The company plans to invest further in predictive science, digital twins, agentic AI, and personalized consumer experiences.
Harcup sees the real opportunities at the intersection of real-time consumer signals and scientific data. "Consumer insights are at the heart of everything we do, while science allows us to translate those insights into superior-performing products," he said. "AI helps connect those two worlds by bringing together external consumer signals, proprietary research and scientific data to identify unmet needs and guide formulation, claims and product design with greater precision."
Why this matters for product development professionals
Unilever's approach signals a shift in what speed-to-market can look like when AI handles the heavy lifting of data mining, trend detection, and early-stage modeling. For product developers, the takeaway is practical: the 75% reduction in claims generation time and the collapse of formulation cycles from nearly a week to a day or two come from connecting consumer intelligence directly to scientific datasets. The bottleneck is no longer data collection - it is how quickly teams can extract actionable insights and act on them. Building that connective layer between consumer signals and R&D systems is where the next competitive edge will come from.
Your membership also unlocks:
