Unilever

Unilever is one of the world’s biggest consumer products businesses. Around three-and-a-half billion people use a Unilever product each day, and its brands are behind everything from beauty products and washing powder to ice cream, sauces, and snacks. But the pressure to lead in each market is relentless.

“To keep ahead in our product categories, and be influential in our markets, we have to keep developing products that stand out by meeting consumer’s needs with concrete benefits and value propositions,” explained Kumar Subramanyan, Director of Digital R&D, Unilever. He continued, “we realised that, as in other spheres of business, harnessing and analysing data can speed up this process and help us make good decisions consistently. We wanted to use data to look beyond annual product and brand planning cycles to come up with better ideas, more quickly that unlock the consumer delight that we strive for.”

Unilever asked us for support, recognising the experience we had working across the business, leading a number of ambitious data and analytics transformation initiatives including COVID reporting to help safeguard Unilever’s global ecosystem.

Using the power of AI to meet consumers’ needs

For a product to be a hit, several characteristics have to chime with consumers’ needs. A cleaning or beauty product, for instance, needs the right fragrance and packaging, while delivering on the functional promise. And all products need persuasive and valid claims about benefits, in language that connects with consumers. In a traditional process, teams of researchers comb the market, social media, trends research, and a host of other sources for insights about what works today and make inferences about what might succeed tomorrow.

The tool draws on Unilever’s collective understanding of consumer behaviour, market research, and expertise in product formulations and ingredients to recommend the best combinations to match each product brief. Future versions will also be able to predict the commercial performance of product launches.

PA Data, AI, and Digital Transformation expert, Yash Lakhwani explained: “If a product brief for a new product launch asks for a moisturising soap, for example, Delphi gives ranked recommendations for ingredients and formulations, and packaging, tailored to the region where the product will launch. Based on data from product reviews, social media, and competitors’ products, it also generates product claims that are differentiating in communicating the benefits.”

To ensure delivery excellence, we embedded programme governance and established a cross-capability team of experts that included programme management, people and change experts, data scientists, engineers, developers, and other technology providers to drive momentum. In what was a global undertaking, we brought together Unilever business teams and partners across the United States, India, and the United Kingdom. It was therefore critical that we fostered a ‘one team’ ethos to ensure alignment and ownership for the journey ahead.

A human-centred approach to accelerating AI adoption

To reimagine Unilever’s relationship with AI, our people and change experts designed a change management strategy targeted at driving adoption of Delphi. This meant co-creating an AI vision and communicating and engaging with Unilever’s people to gain buy-in and commitment.

It was vital when upskilling people to ensure that Delphi was empowering them by enhancing their capabilities to provide them with control and an understanding of how Delphi AI arrives at decisions or recommendations. The teams at Unilever are continuously monitoring the programme’s progress and the users share success stories and encourage ways of working that adapt as the solution evolves. This approach was crucial to developing, building, and launching Delphi in just 18 months. It also made sure teams were receptive to the change.

Generating better ideas in less time

DelphiAI launched with laundry products in India and beauty products in the US. Word about its effectiveness soon spread throughout the teams at Unilever, and the Health and Wellness products group in the US began working with it later in 2023. In 2024, the goal is to scale Delphi through more business teams, with nine already set to start using it.

The tool promises to revolutionise product development, says PA’s Richard Chamier, “R&D and marketing teams have always had forecasting, trends research, and other data but this tool shows what that all means for product innovation by giving specifications that teams can develop at pace.”

The collaborative process that produced DelphiAI has already inspired new thinking within the R&D group in Unilever’s approach to data to make better, more informed decisions when driving product superiority. Kumar Subramanyan, Director of Digital R&D, Unilever said: “The Delphi experience has underlined the benefit of ecosystems of experts coming together to innovate with us. This includes our new virtual digital lab called DataLab, with 12 technology and analytics partners, including PA. Because DelphiAI is a shared platform different functions can access, it’s also helping our teams collaborate more.”