In May 2019, a leading Brazilian beauty and perfume brand, O Boticário, introduced two new fragrances named Egeo On Me & Egeo On You. Apart from their fascinating names and interesting ingredients like combining notes of fruits, flowers, woods, caramel, and even condensed milk, what set them apart and took the internet by storm was how they were created.
These perfumes weren’t created by humans but were created using Artificial Intelligence or AI. Yes, we did have AI-generated perfumes long before ChatGPT went mainstream. The AI technology behind these perfumes was IBM’s Philyra, a smart system that was trained to analyze thousands of fragrance formulas, customer preferences, and even their molecular compositions to generate unique scent combinations.
This wasn’t a one-off experiment, but the scents became a huge hit among both men and women, signalling a massive shift in traditional perfumery that relied a lot on intuition and craftsmanship. The combination of AI and aromas meant more benefits, such as:
- Personalized formulations based on individual preferences.
- Faster research & development time through predictive modelling.
- Lesser waste of resources by optimizing ingredient combinations before physical testing.
But today, it's evolved more than that - it can predict how ingredients will interact before they’re even mixed and can even create a perfume based on a person's digital history. What’s more? Let’s find out.
Data-Driven Formulation – Role of AI in Scent Creation
Before AI, perfumers who were also known as ‘noses’ relied on their skill and intuition to craft scents. Fragrances typically follow a three-layer structure:
- Top notes: Light, fresh scents like citrus.
- Heart notes: Floral or spicy elements like cinnamon.
- Base notes: Deep, lasting aromas like vanilla or musk.
Creating a perfume from the base required countless trials and ageing, all of which made the process slow and cumbersome.
But thanks to AI, that's not the case anymore. AI tools like IBM's Philyra and Symrise's AI-powered fragrance assistant along with Machine Learning (ML) and Deep Learning identify patterns in molecular structures, predicting how different ingredients will interact. This dramatically reduces trial and error, allowing perfumers to focus on refining the final formulation rather than starting from scratch.
Digital Scent Avatars – Personal Fragrances
One of the best uses of AI is the level of hyper-personalization it allows. Imagine an advanced AI system that can create a unique fragrance just for you based on your mood, search history, shopping habits, and even social media activity.
Too good to be true? Well, a Dutch startup named EveryHuman is already making this a reality.
How Exactly Does AI Understand Perfumes?
EveryHuman’s AI tool understands and analyzes fragrances by using their ‘custom scent database’ built from scratch. First, they collaborated with expert perfumers worldwide to distil thousands of fragrance elements into 46 distinct “scent building blocks.” Each block contains nearly 2 to 20 raw fragrance materials, starting from classic notes like rose and vanilla to more abstract scents like ‘metallic’ or ‘depth’.
When users visit an EveryHuman location or access the platform online, they start by filling out a detailed questionnaire. Instead of simple multiple-choice answers, the questions dive deep into personal preferences, asking about their style, favourite colours, dream destinations, and even emotions tied to scents. The AI analyses the answers and considers how long a person takes to respond, which could indicate which choices they feel strongly about or are unsure of.
After processing, the system generates three unique fragrance options in real-time, each tailored to the individual’s personality and sensory preferences. Customers who like one can purchase it immediately or order it online. If they want adjustments, EveryHuman offers a fragrance coaching session where they can provide feedback, refine their scent, and even influence how the AI continues to evolve in its formulations. This dynamic approach ensures that no two perfumes are alike, making every fragrance as unique as the person wearing it.
Ethical & Privacy Concerns
While AI is revolutionizing the fragrance industry, it also raises important ethical and privacy concerns. Personalized scent creation relies on collecting vast amounts of user data, including mood patterns, shopping habits, and even emotional responses. This level of data collection poses questions about how consumer information is stored, shared, and protected.
Additionally, as AI-generated fragrances become more sophisticated, there’s a risk of losing the human touch in perfumery, where traditional craftsmanship and intuition play a vital role. Striking a balance between innovation and ethical responsibility will ensure AI-driven perfumery remains creative and consumer-friendly.
Future Scents – What is Next?
AI is not just personalizing fragrances – it’s redefining their purpose and the process of how we experience them. Some emerging trends include:
- Olfactory Augmentation: AI-designed scents are used for well-being, such as boosting focus, enhancing relaxation, or even improving sleep. Companies like International Flavors & Fragrances (IFF) use AI to develop functional fragrances that interact with brain chemistry to promote emotional balance.
- VR-Integrated Fragrances: AI is also bringing fragrance into the digital realm, with companies like OVR Technology developing scent emitters for virtual reality (VR), making online experiences more immersive. Imagine digitally sampling a perfume before purchasing it online — AI is making it a reality.
- Adaptive Fragrances: Wearable scent technology can now respond to emotions and environmental changes in real-time. Startups are developing mood-sensing devices that detect stress levels and automatically adjust the intensity or composition of a fragrance, creating a truly personalized and responsive olfactory experience.
With AI slowly engulfing almost every area of our lives, we ought to rethink the very essence of a fragrance. From a bottled-up liquid scent to a dynamic mix of data, tech and emotions, AI is redefining personalization.
But does this level of personalization improve our sensory experiences, or does it blur the thin line between thoughts, intuition and algorithms?
What happens to the memories and nostalgic connections we attach to certain smells?
If AI can do it - from predicting, analysing and even formulating perfumes, what happens to the artistic side of the process of making perfumes?
If AI can predict what we like before we even smell it, are we still exploring scent, or are we just being guided by data?
