Machine learning has infiltrated nearly every aspect of consumer living, and the world of olfaction is no exception. But how does the cold, hard, data-fueled approach of AI stack up against the romantic, emotionally intertwined world of handmade scent?
The last industrial revolution of perfumery came about in the early 1900s with the discovery of synthetic aroma molecules, and AI is now increasingly on the uptake, bearing the potential of the fragrance industry’s digital revolution equivalent. Forced to shift all retail operations completely online seemingly overnight, perfumery has awoken to a dire need to overhaul its digital presence. Extending this to the production realm had remained a more nascent aspect, although AI is now being increasingly employed by startups and manufacturers.
Machine Learning-Driven Customization
AI, beyond being a mere tool, now constitutes the core of fragrance-focused startups. Smart perfume brand NINU offers customizable fragrance options enabled through a coordinating app that lets customers choose their fragrance for the day based on occasion, scent preferences, and intensity desire. NoOrdinaryScent creates perfumes based on image submissions from customers which are then interpreted by an AI into fragrance compositions.
The potential of this technology not only offers additional variables in terms of how fragrances are made, but can be harnessed to create new ingredient possibilities. Osmo, a startup backed by Google Research and an initial $60 million funding round, is using generative artificial intelligence and an odor map (built to predict the smell of a molecule based on its structure) to create new aroma molecules.
Givaudan and Tmall Innovation Centre launched the AI-powered co-creation service Creatogether, following on from the founding of the T-Lab Source Innovation Laboratory, which enabled end-to-end fragrance product development in as little as four weeks. The service utilizes a previous Givaudan launch, the machine learning-driven fragrance assistant Carto, which suggests initial formulas to the perfumer in order to strengthen the performance of the end product, utilized in the creation of Etat Libre d’Orange’s She Was An Anomaly fragrance, created by perfumer Daniela Andrier.
Creatogether is a five-step tool that spans the creative research process, product creation, sampling, launch strategy, and marketing for hands-on, immersive, and speedy operations—bringing together the previously separate realms of the perfume lab and brand office. “Our common goal is to bring agile solutions to the market by connecting consumer insights with industry knowledge and expertise, so as to provide users with instant, accurate and personalised fragrance products,” explains Tian Qi, Head of Tmall Innovation Centre.
It's not the first time Givaudan has invested in this space. In 2021, the Swiss manufacturer had acquired AI tech company Myrissi, which has created “a patented AI technology capable of translating fragrance into color, patterns and images, relevant to the consumer as well as predicting the end consumer’s emotional response.” The program aids the manufacturer's visual storytelling amidst the increasing growth of e-commerce trade in the notoriously difficult to optically or scripturally capture world of olfaction.
Competitor Firmenich used ScentMove neurotechnology, which applies fMRI brain scans of a test subject smelling different fragrances to create more nuanced emotional connections and contexts through its products. “To drive emotion through scent, one solution does not fit all. By applying our proprietary scientific and consumer research, we can help customers create stronger emotional impact, with contextualized solutions that reflect country and category preferences,” says President of Global Perfumery Ilaria Resta.
Firmenich also launched Scentmate, an AI-enabled platform which recommends bespoke fragrance solutions to briefing input in minutes, as well as simplifies order fulfillment, shortening the period of launching a creative vision to market. The tool also offers safety and regulatory legislation guidance based on geographic markets. "Scentmate allows customers to match their creative ideas with the ideal, winning fragrance," comments Alfonso Alvarez-Prieto, Scentmate General Manager. "We offer an intelligent, user-friendly platform, bolstered by human creation, to easily co-create the perfect fragrance that will drive consumer preference, while dramatically increasing speed to the market."
In November 2020, the manufacturer had already unveiled AI-created laundry care fragrances—a liquid detergent with yuzu and fruity floral scent options, a gourmand fabric spray, and citrus powder detergent (all tailored towards preferences of the Chinese fragrance consumer market). “AI is now systematically empowering our perfumers to achieve their very best creations by identifying potential solutions across a complex combination of parameters, including olfaction, format, dosage, safety and sustainability,” states Resta. The fragrance manufacturer is tapping into the power of AI for “augmented creation, prediction and optimization to provide bespoke sensorial experiences faster and more efficiently than ever before,” according to SVP of Creation and Development in Global Perfumery at Firmenich, Odile Pelissier.
IBM’s Philyra had concocted fragrances for the Brazilian market in collaboration with Symrise in 2018. The two fragrances Egeo on Me and Egeo on You were created on behalf of O Boticário, the second-largest Brazilian beauty company, and specifically created with millennial consumer audience data. “Philyra does more than serve up inspiration—it can design entirely new fragrance formulas by exploring the entire landscape of fragrance to discover the ‘white spaces’ in the global fragrance market,” IBM research scientist Dr. Richard Goodwin explains. Master perfumers still work alongside the machine to give the formula the final tweaks, but with personalization in product offering being a key market growth factor, and the machine’s ability to scan hundreds of thousands of formulas and thousands of raw materials in mere seconds, AI-driven customization does have its efficiency merits.
At the 2022 World Perfumery Congress, the two companies unveiled Philyra 2.0. The updated version incorporates a new range of eco-friendly ingredients to reduce the negative environmental impact of product creation. "With Philyra, we can find new inspirations, I use it as creativity booster or to modernize iconic classics. The beauty of Philyra 2.0 is that I can now also develop fragrances with renewable as well as biodegradable raw materials," comments Pierre Gueros, Senior Fine Fragrance Perfumer. With the upgrade of Philyra, Symrise now aims to combine creative fragrances with high sustainability values. In doing so, the company is redefining perfumery for the modern age. "We are opening a new era where a successful fragrance will result from the alchemy of perfumers’ experience, intuition and creativity combined with Philyra’s algorithm”, adds Viola.
Consumers get excited about fragrance AI
AI is gaining popularity as a differentiator for their brand positioning. "Early adopters show great interest in how AI influences the creation of fragrances," says Nathalie Benareau, Fine Fragrance Perfumer. Paco Rabanne recent men's fragrance launch, Phantom, was created in partnership with International Flavors and Fragrances (IFF) using an AI program that predicted the emotional effects of the different ingredients in the formulation. The end result uses an overdose of the styrallyl acetate molecule, plus lavender, as well as green, citrus, and woody aromatics to boost energy and self-confidence.
While mass-market options tailored towards pleasing the masses may be a no-brainer application for AI’s data-driven approach, the individual stories, experiences, and narratives driving other fragrance concepts cannot be replicated by a machine. Unless that machine can live, feel, and experience the world just like a human being, that is. Google Brain’s research team may have taught an AI to smell, but it has yet to parallel our own noses. As a human species we have 70% shared olfactory receptors (meaning those parts of our smelling abilities are the same), but 30% are unique to the individual.
Perfumery may technically come down to formula, but it should (ideally) not be formulaic, for there are creative deviations from the norm that only the human hand can administer, and that’s what keeps the genre exhilarating. “Without the human, all you get is a blend. If you told a machine you like blue, red, and pink, and then you asked the machine [for a painting], you would never get a Mondrian,” exclaims perfumer Francis Kurkdjian. So far from fearing automation, perfumers should see it as a chance to elevate their craft, to dream even bolder (olfactory) dreams.
