Creating a personalized perfume using AI requires addressing three challenges simultaneously: transforming a person’s description into a formula, presenting the resulting scent before it’s smelled, and accurately producing that formula at the point of sale. SCircle combines a digital perfumer named Aura, a visualization of the perfume, and a machine with capsules that precisely mixes the blend in minutes. In chapter 67 of Toque de Ingenio, Álvaro Suárez explains how this system was built, why it originated within an appliance group, and what he learned during its first three weeks of sales.
Guest: Álvaro Suárez, who was introduced in the episode as the CEO of SCircle. Interview published: November 11, 2025. Episode: 67. Duration: 58 min.
In this episode:
- How can AI translate memories, places, and reference scents into a formula?
- Why 34 compatible capsules replace the 4,000 ingredients a perfumer needs.
- What did the data from the first week of sales reveal, and what changes did the team implement?
The conversation was recorded using the Scent Creator machine on the table. Edgar requested a perfume with a wood and leather scent, intended to evoke the memory of a rock concert, and it was created during the interview. SCircle had been selling in the United States for three weeks, so the figures and plans referred to that specific time.
A digital perfumer who begins with the brief.
Álvaro is part of a historical comparison. For centuries, only prominent individuals had access to a perfumer – an artist who memorizes ingredients and is able to imagine how a blend will smell. He explains that a perfumer is familiar with approximately 4,000 ingredients approved by the International Fragrance Association.
SCircle has recreated that fragrance using artificial intelligence. Aura, their digital perfumer, draws on knowledge of the brand’s ingredients, information on perfumery available online, and the formulas created by users. The more perfumes people create, the better the system becomes.
“You now have a personal perfumer in your pocket, thanks to your phone”.
Álvaro Suárez 06:20.
The interaction mirrors the process a perfumer would undertake. The user describes ingredients they like or dislike, perfumes they find inspiring, or stories and places. Álvaro illustrates this with an example – a cove on the Costa Brava: the system tends to add a hint of sea breeze with pine, reflecting the character of that location.
According to his account, that correlation is valid: the knowledge exists, but it’s scattered among perfumers and doesn’t converge in any one place. AI can help to structure it.
Reducing 4,000 ingredients down to 34 compatible capsules.
It was technically impossible to offer 4,000 ingredients to each individual. Working with the fragrance house that helped develop the concept, the team concluded that the best approach was to create pre-made blends, which could be combined and distributed across the fragrance wheel – the system used by perfumers.
There are 34 different blends, based on two types of fragrance components. The foundation consists of partially-created perfume structures that form the olfactory pyramid, incorporating notes that evaporate quickly, such as citrus fruits, and others that last longer, like woods. The accords are rich ingredients that, on their own, have an incomplete scent. A perfume combines one, two, or three of these base notes with any of the accords desired.
There’s another important chemical decision. A traditional perfume is a 20% concentration that is mixed with alcohol and fixatives, and requires several days of maceration. In SCircle, everything is encapsulated, so the perfume is ready to use immediately, without the waiting time required by other competitors.
Teaching someone to recognize a scent before smelling it: the scentboard.
Perfume is the beauty category that has lagged behind in the digital space, according to Álvaro, because only people who are already familiar with a particular fragrance tend to buy it online. SCircle is attempting to break down two barriers simultaneously: online sales and personalization.
The tool is a scentboard, a collage of photos representing the perfume. An algorithm calculates how the 100 or 150 ingredients in the formula combine, and it displays only the dominant notes – those that a non-perfume expert would perceive. The size of each image indicates how much of each ingredient will be detected.
Álvaro identified that aspect of the software as the most crucial. Before starting the project, they explored numerous ways to present a fragrance – with their own ideas, those of the perfumery partner, and suggestions from agencies – and then tested it in stores and with perfumers from other companies.
The aim is to close the loop: ensuring that what is seen aligns with what is sensed. This allows for iteration, much like a perfumer working in real-time – removing leather, increasing vanilla, and recalculating. This work User experience design explains that the first session will last approximately 20 minutes.
A machine that makes the perfume in the shop.
The Scent Creator is, according to Álvaro, a kind of alchemy lab with capsules inside. Each capsule contains a small electronic component, its specific blend, and a number. Out of the 34 blends in the catalogue, the machine can accommodate 12, and a perfume typically requires four or five of these capsules.
The application that controls the machine reads each cartridge and how much remains, alerting you if a cartridge or bottle is missing, and changing the light when it’s ready. A single sample takes one minute; a large bottle, three. The sample uses between 0.5 and 0.8 ml, compared to the usual 1.7 or 2 ml, and doesn’t produce any residue.
“Everyone who has worked with hardware knows that hardware is demanding”.
Álvaro Suárez 50:23.
Álvaro admits that he didn’t experience that part as intensely: the hardware was coordinated by the appliances group. His team defined needs and specifications, and it was a team with expertise in Electronic engineering The process of assembly and construction built the machine.
The machine operates as a production unit – in retail stores for partners in the United States and Mexico, and as a small-scale manufacturing facility within logistics centers, where an online order is printed, packaged, and shipped. The small bottle costs $25, and the 20ml bottle costs $95.
From innovation project to startup: buying what’s left over.
Álvaro is an industrial engineer, and he spent his first eight or nine years working for Balay, a major appliance manufacturer with its headquarters in Zaragoza. He eventually moved to the innovation department in Germany, where a project began in 2017, originating from the idea of Javier Sanz, a chemist from Zaragoza.
The innovation committee rejected the project for six months. Álvaro presented it each month with a different approach, until he brought to the meeting a recorded interview with a former executive in the beauty sector in New York. The project was subsequently approved as a household appliance.
This led to a joint venture with a perfumery partner, with 20 people in Barcelona and a shop in Paris. He explains that COVID and disagreements between the partners eventually led to its closure. Álvaro later bought the inventory, moulds and software, and acquired the patents in 2024 when his non-compete period ended.
Following all of that, he contacted the former group president, who had retired, and they formed SCircle in March 2025, bringing back most of the previous team.
Launch, measure, and adjust within three weeks.
SCircle entered the US market because its marketing partner is American and a perfumer, and due to the fact that the market for niche perfumeries is doubling among people under 25 each year.
The pre-campaign week, the last one of September, yielded no results. The team used Microsoft Clarity to track user journeys on the website and interviewed those who had developed the product formulas. They discovered that people didn’t know what was being sold – there wasn’t even a picture of the bottle. They corrected this, and the first orders started arriving on October 1st or 2nd.
Three weeks later, Álvaro was speaking about hundreds of perfumes sold and thousands of formulas created. Half of those who join create a digital perfume and an account, with an average session length of 15 minutes.
When asked about his biggest professional mistake, he cites a previous product project where he waited too long to launch it. His key learning is not to delay the release, as it never seems like enough – feedback from the initial orders was used to make changes the following week.
What can another development team learn?
The SCircle case offers four key lessons for products involving chemistry, hardware, and software:
- First, design the consumable. The 34 capsules and their specific combination logic dictated the machine, not the other way around.
- Show before you manufacture. A reliable representation of the outcome reduces the risk of buying something that cannot be tested.
- Measuring the actual distance traveled by the user. A simple mistake, such as not demonstrating how to use the bottle, can actually halt sales.
- Distinguishing between hardware and software rhythms. The device requires a certain level of maturity; the software is updated with each order.
At the time of the interview, SCircle was selling in the United States, opening stores in Mexico on November 12th, and preparing for its arrival in Europe in early 2026. Álvaro was also developing smaller versions of the machine for beauty ambassadors and influencers.
At i-mas, we develop products that integrate mechanics, electronics, consumables, and software, from prototyping to manufacturing. If your project requires dispensing, measuring, or producing something at the point of sale, Tell us about your project. or see our product design and development.
To continue reading: How to develop an automated fresh food dispensing machine for retail..
Source of article: Interview with Edgar Guerrero on Toque de Ingenio with Álvaro Suárez. The excerpts cited link to the minute of the conversation. The recommendations for other projects are a summarised editorial from i-mas.
