Tiffany Iser
For many people, the aisle of personal care may be the last place you associate with machine learning. But for UK-based Unilever, decades of hair protein data and over 500 Live AI projects have transformed these shampoo and skincare shelves into proven land. Building the foundation has been like a “travel” for years, according to Tiffany Iser, R&D for the company’s beauty and happiness in North America. The company has given decades of hair protein experiments to understand, for example, the individual and regional nuances of textured hair types and straight hair types. Unilever boasts one of the highest concentrations of global material chemistry robots at its facility in Liverpool, UK.
The company’s R&D Digital Transformation Journey worked with scientists to provide structured data captures, supplying everything from early development experiments to pre-production confirmation batches to central data lakes. We are also working to keep our bench tools away from manual data entry, connecting them directly to the lake. “As well as engineers responsible for expanding their formulations to prepare for production scales, they’re moving forward for years to secure digital twin pilot plant facilities,” Yizar said. This allows data from scaled production simulations to be directly correlated with those experimented by scientists years ago on bench scale. According to Yizar, the ambitions are expanding even further. “Our factories are on the same journey, and we’ll look at ways to integrate them further and get a full closed loop,” she said.
This helps correlate what we often saw as our first hypothesis 3-5 years ago with how we deploy it at a large scale
The task of connecting decades of experimental data across different scales and even incorporating post-launch consumer feedback is changing the way Unilever approaches science. In fact, the company’s digital muscles played a role in the 2024 launch of the Vaseline Provitab3 Serum-Burst lotion. “Deleting data silos helps to unlock more of the “AHA” moments where you can see the data and results generated by scientists in the lab in Trumble, Connecticut, and compare it with results from scientists in labs in Mumbai or Shanghai that have been generating and verifying,” she said. This approach “gives the confidence to implement innovation in the market or see if it is inconsistent, and we need to dig deeper.” Digital focus also extends to tracking consumer signals, helping companies measure satisfaction with a particular product. “If there’s that gap, it allows us to respond fairly quickly,” she said.
Unilever Scientists at SXSW: Expanding the Role of AI
In addition to a recent interview with R&D World, the Unilever article in March 2025 summarized highlights from a 2025 presentation by SAM SAMARAS (SVP of SVP, Personal Care R&D) and Tiffany Iser.
Unlocking Complex Biology: Yizar and Samaras explain using AI to “solve and explore vast datasets” from human biology, comparing them to finding the “butterfly” in Chaos theory. In other words, small changes that could lead to large product breakthroughs. AI reduces wet labran. Over the years, “there’s been an exponential change in the way we can understand complexity,” Samsamaras said. Now, the company can do that in just a little bit. And machine learning “helps mathematically model the rest using very sophisticated statistical methods,” she added. Meet a diverse range of needs: Unilever’s Polycultural Center of Excellence (PCOE) is mentioned. AI and data analysis shed light on understanding and development of products tailored to melanin-rich skin and textured hair. Microbiome Mastery: The company uses extensive Microbiome data, one of the world’s largest, to employ pioneering innovations from AI like whole body deodorizers, understanding the unique biology of sweat and odor across a variety of body regions. As of 2024, the company had registered over 100 microbiome patents. Nanotechnology Case Study: Through Nanotechnology, Dove’s micromotion technology development exemplifies how Unilever uses analytics and refined science with digital tools to create a “clearly superior” product experience. Human ingenuity, AI capabilities: Both scientists emphasize that AI and machine learning accelerate research, handle complex modeling, and free human scientists for the “type of creativity that only the human mind can bring.”
Source: Unilever News, March 20, 2025
These “aha moments” have a distinctly human touch at Unilever’s Liverpool facility. There, robots with names like Shirley and Ariana work with over 250 R&D specialists. “Shirley and Ariana are my best robots for hair care,” Iser said. In the lab, robots can help you redefine your workflow.
Mining consumer voices to create new lotions of petroleum jelly
One of the repeated complaints on sites such as Amazon, Reddit (particularly R/Skincarediction), and product feedback sites is lotion that made the skin feel “greasy” and “heavy.” Unilever has decided to dig into broader feedback on the subject. The company’s initial consumer survey confirmed that the issue is actually a broad barrier, but its extensive insights alone were not practical enough for product development. “We were able to use consumer digital voice to get more specific about not only their general insights, but also about a specific cohort of consumers who are bringing that insight at a particular time,” Yizar said. Logically, such complaints were more closely correlated with hot, humid environments and summer consumers.
The Unilever R&D team then turned to what Yizar calls broad product profiling. “We see sensations as a multifactorial experience. We can profile over 50 different attributes of a particular lotion and model where those products are in those different attributes,” she explained. The Vaseline team used this mapping approach to identify opportunities. “Our Vaseline team did it in earnest and specifically identified this white space, which overlaps with its light, sensory, fast absorbing product,” says Yizar. The team “using modeling and mapping tools, it allowed the formula to define the blanks that can develop winning products.” Its technical brief became the ProvitAB3 Serum-Burst lotion, launched in 2024.
Unilever’s Provitab3 Serum-Burst Lotion
The actual test came in a hot, humid day in New York City at a product launch event. “It was the first time I’d been to a Vaseline launch event where people were enslaving lotions to themselves, because it was just lovely sensible, very fast absorbing, and very refreshing,” recalls Yizar. The Serum-Burst Formula dealt with a permanent “second layer coating” complaint that emerged in the first consumer survey.
Global R&D Lab Lessons
Yizar highlights that by breaking down data silos across the Global R&D Lab, for example, in Trumbull, Connecticut. Unilever’s Mumbai or Shanghai was now able to identify both confirmation and conflicting results from experiments performed in parallel at various sites. Even competing data can be useful. This can help you answer questions such as “Is this a patentable space?” “Is this a place that may be “aha” for us and not “aha” for someone else yet? In short, the ability to disassemble silos and discover more is “very exciting right now, and can’t be done without digital tools.”
This strategic approach to R&D has even been mined for its own insights, and is reflected in the company’s substantial patent portfolio, a figure of over 20,500 worldwide.
Robot Shirley and Ariana are working on a boring job
At Unilever’s Material Innovation Factory in Liverpool, a team of over 250 R&D experts work with unique colleagues with names such as Shirley, Ariana, and Gwen. But she calls them the “best of hair care,” but they are not human. They are robots. “What they really unlock for me or for the team there is to cut the time for some of the daily tasks we need to do biological research, scientific research in hair care,” she explains.
Take Shirley, for example. The robot “washs hair samples on a large scale, something that a regular scientist can even clean 10 times a day,” says Yizar. Shirley can process up to 120 samples every 24 hours. Her counterpart, Ariana, supports detailed hair fiber assessments, while Gwen focuses on analyzing the form. This high-throughput, consistent work is “fully integrated into the information supply chain in R&D,” says Yizar.
One of the beauty robots from Unilever in Liverpool
Robot assistance is directly converted to tangible R&D gain. For example, the Tresemmé Color Radiance Booster Range uses a technique Shirley identifys through sustained hair washing experiments. “We’re thinking about a multi-year landscape to reach pigeon bioprotein care, so by leveraging that type of robotics capabilities, we were able to get deeper and move a little faster when we proved the technology that came into the product,” Yizar said.
The Robots and Data Lakes represent only the current stages of Unilever’s digital transformation. Following a massive cloud migration using Microsoft and Accenture, the company uses Microsoft’s Azure Quantum Elements. This is a system aimed at accelerating the discovery of new chemicals and materials. Yizar points to the discovery of unique molecules that support its sustainability and microbiome programme.
Unilever also gained access to IBM’s Quantum Network through its £210 million Hartree Center Collaboration with IBM in the UK. This will provide access to IBM Quantum Computers and development tools to over 150 organizations.
Pull the beat push
Through its digital R&D initiative, Unilever is creating a more titered closed loop to seamlessly connect decades-old bench data with real-time customer feedback to continuously improve products and drive innovation. The value of this integrated approach has become very clear and, as Yizar emphasizes, achieved momentum. “The advantage from lab bench chemists to us to our leader-level users is that we really could see this adoption being adopted.
