Artificial intelligence is constantly learning, these include the art of perfumery. In 2017, artificial intelligence was able to compose music, which lead to the same question arising in the art of perfumery. Can the computer invent a new scent?
In theory, if a large database containing every formula in perfumes, past and present, were to be created, an algorithm could curate the most popular scents in order to programme the ‘perfect fragrance’. In these scenarios, we could say that machine learning or data science has the ability to predict a variable on the basis of existing data that has accurate information. In an artistic profession like a perfumer, these technologies have long been incorporated into the creative process, as Francis Kurkdjian states formula creation, and artificial intelligence would automatically determine the quantity, the price, and the current regulatory status. These tools are like giant virtual assistants that would allow us to carry out certain tasks more quickly.
How does artificial intelligence help the process of scent creation?
Consumer tests
Hypothetically, a brand sends a brief for a perfume aimed at 40-year-old urban women that has to be very popular in the United States without falling flat in France and Germany. The perfumer and evaluator consult the database to find the accord of the best matches to the demands. The algorithm makes productivity gains possible by capitalising on the composition work of all the perfumers within the same house, in order to free up time for the creative process. The raw information acts as the basis for the perfumer’s work to find the best fragrance that would appeal to a certain target market.
Personalised recommendation
Online clients specify whether they want their perfume to be feminine, masculine, unisex, subtle or intense, to include certain ingredients or to exclude others, and the algorithm recommends the creation which best matches their request, identified simply by a number. Samples and a small bottle can be sent to them, so they can get to know the brand, and the consumers can easily send the box back if they want to exchange it. Brands are able to have the algorithm developed to include pyramids for perfumes currently on the market, to help beginners identify their favourite ingredients.