Perfumery has always existed at the intersection of art and science, but today the question arises: has the second term really not yet supplanted the first? Or has artificial intelligence already firmly established itself in this niche? We spoke to perfumer Maria Golovina about how algorithms help generate fragrances and which brands (including local ones) have already implemented it.
PS: In unison with the raised topic, we decided to keep up with advanced technologies and turned to artificial intelligence for help in creating a cover for the material. We are still practicing scriptwriting (it worked on the eighth try), but the result is more than satisfactory. Script: “Draw a realistic picture of the process of creating beautiful aromas by a robot in a pink laboratory with flowers.”
perfumer Holynose, Nose Republic, Don’t Touch My Skin, author of the telegram channel “Musky Muscovite”
Elle reports: The perfume market is going through better times today, worth $45.85 billion in 2022 and expected to grow to $69.25 billion by 2030. This is generating interest in studying the influence of scents on human psychology and behavior. The sense of smell, unlike other senses (taste, touch, vision and hearing), whose signals go first to the thalamus and only then to the head, has the shortest path and is sent directly to the bulb located in the front part of the brain. brain. Thus, fragrances immediately evoke vivid memories, strong reactions and nostalgia. Here it was discussed whether perfume is really capable of influencing the state of health and what functional perfumery is in general.
@holynoseparfums @donttouchmyskin @holynoseparfums
Of course, in this situation, major perfume brands will rely more on artificial intelligence tools, because only then will they be able to offer customers the necessary amount of individual scents. Let’s break everything down into its components: when and how neural networks began to be used in the industry and what manufacturers they were from.
Swiss perfume company Givaudan has been using mechanical algorithms to decipher complex scents for more than a decade. At the same time, Masha notes: it is difficult to say when they began to introduce artificial intelligence into developments. “These innovations occur behind the closed doors of giant companies and often remain in the territory of corporate secrets. There may be a number of AI-powered food and perfume flavors on the market before they are released to the general public. But I remember the first time I heard about Carto, Givaudan’s artificial intelligence assistant, in the context of Etat Libre d’Orange’s She Was An Anomaly fragrance in 2019,” Masha shared.
Carto digital assistant
@scenthunters
“Artificial intelligence is based on data sets that the main market players (Givaudan, IFF, Symrise, Firmenich and others) have and that relate to the preferences of existing consumers and sales leaders. From this information new formulas are outlined. In addition, the results of scientific research on the emotional and physiological reactions of people to certain aromatic components are used (such research is also sponsored by large manufacturers). Advances in this area were useful, for example, for the launch of Paco Rabanne Phantom in 2021,” shares Masha.
Paco Rabanne Ghost Fragrance
Last January, the aforementioned Givaudan announced Myrissi, a technology that can transform aromas into color patterns using AI. It is currently only available in Western Europe, the US, Brazil and Japan. The creation of the platform became one of the stages of the strategy for the implementation of digital technologies until 2025.
The mechanism of her work is as follows: Myrissi establishes associations between aromas and colors, after which she creates special mood boards for perfume brands. In other words, the algorithm is designed to gain a deeper understanding of consumers’ emotions. The challenge of technology is explained by Maurizio Volpi, president of Fragrance & Beauty: “Since e-commerce remains the main communication channel, we must satisfy the need of customers to want to listen to fragrances even at a distance and through digital screens. the telephones. This was not possible before Myrissi. With it, we invite you to “feel” images and colors that evoke the same emotions as the perfume.” So, with the help of Myrissi we feel colors and “smell” emotions.
Feel colors, smell emotions with Myrissi™
Myrissi by Givaudian @holynoseparfums
Another company that uses consumers’ personal data is No Ordinary Scent. To generate smells, analyze the three most “emotionally charged” photographs of the user. They could be vacation, wedding or birthday photos. The machine extracts emotions from images (no matter how elegant they may seem) and associates them with olfactory notes. This results in a new, and most importantly, individual formula, which can be compared to a “fragrant wardrobe” that emphasizes individuality and changes with you every day.
NOS Emotiontech Workshop NOS Emotiontech NOS Emotiontech
“As an independent perfumer working in a small laboratory, I (like my comrades) prefer to dream of more basic things for authors who work in corporations. This includes a gas chromatograph with a good database, software that helps you easily calculate the cost of a formula and much, much more,” Masha responded.
@holynoseparfums @holynoseparfums
“And you also had the opportunity to play with AI at a basic level thanks to the EveryHuman Algorithmic Perfumery project. These are the same robotic collectors adapted for public use. They bottle several versions of perfumes, the formulas of which are created using digital algorithms based on personal questionnaires. The client answers 20 questions and the machine, in turn, creates three sketches of perfume formulas, which, if you wish, you can study, edit, name in your own way and, if you don’t like it, take the test again. ” says Masha.
EveryHuman Algorithmic Perfumery “Little Factory” YSL Fragrance Selector Device
EveryHuman co-founder Anahita Mekanik notes that the first perfume is created from answers to psychological questions, the second combines the user’s mental reactions with their demographic data (age, gender), and the third is based on customer interaction with the test itself. For example, how much time the customer spent on the trip. Thus, the tool can generate more than 500 billion variations.
Fictional fragrance bottles @takiekrasivieflakony Fictional fragrance bottles @takiekrasivieflakony
“Givaudan has Carto and Myrissi, Symrise has Philyra, Firmenich has Dr. Nez. By the way, with the help of the services of the latter (and, of course, the main perfumer chosen for this project from the corporation’s portfolio), the Russian company Faberlic created the Futuria fragrance in 2021. Manufacturers can turn to actors larger, those that already have this technology. In other words, any whim and any data is on your money,” says Masha.
Faberlic Futuria Fragrance Faberlic Futuria Fragrance
“No. As in other areas, artificial intelligence does more quickly what would take a human years. Another question is how its integration will affect the sales of new products, that is, increasing these indicators is what companies usually aim for. transnational perfumery and cosmetics companies. In the end, without human experience and without control over it, the meaning of the work of algorithms is lost,” Masha shares.