How Colgate-Palmolive Is Bringing AI-Enhanced Scents to Retail
Colgate-Palmolive detailed how it’s having success in using artificial intelligence and predictive analytics to bring new product scents to shelves.
The CPG last year launched its first full line of shower gels that made deep use of AI from a fragrance house partner, the company shared in a blog post. With the goal of blending ingredients to match scents with emotions like happiness and love, they leveraged consumer reactions to colors, images, and smells to develop the Palmolive Aroma Essence line of shower gels.
The process involved billions of data points that also included brain scans, according to the company.
AI algorithms predicted a certain percentage of the specific emotion enhancement that the fragrance was created for, said Rafael Giganti, Colgate-Palmolive senior director of fragrance development. The company then followed up the findings with a consumer data panel to confirm the predictions.
The line of gels, which includes “moods” like Forever Happy, Ultimate Relax, Alluring Love, and Sweet Delight, are currently available at retail in Europe. The company intends to use AI to further refine the products, and it expects the fragrance model will find more innovative fragrance combinations as it ingests more data.
“As we continue to refine our AI tools and gather more data, the potential for creating even more personalized and effective fragrances becomes limitless, not just for Palmolive but for our entire brand portfolio,” said Giganti in a statement.
AI’s Fragrance Future
As consumer demand for new products ramps up, CPGs are eager to showcase their success in using AI to expedite the innovation cycle. The use of AI to develop and enhance scents has garnered particular interest from science research communities and consumer goods organizations alike given the complexity of the process.
Procter & Gamble last year began using a new fragrance design platform within its product innovation process; like Colgate-Palmolive, the CPG expects the technology to help bring products to market more quickly.
How AI can help: In one research study, Google researchers used graph neural networks to analyze a database of 5,000 molecules with known scents, training models to link molecular structures to odors. In testing, AI models were better at identifying scents than humans for more than half of the tests, and they used the AI to create a database of scents for 500,000 new compounds that had been smelled before.
One of the researchers, Alex Wiltschko, went on to co-found AI scent solution provider Osmo. At the recent Axios AI+ Summit in New York, Wiltschko said that computers will one day have a sense of smell, and he highlighted AI’s role in helping preserve scents that are disappearing because of raw material shortages.
