3 Questions for Paul Mouginot
Co-founder of Stabler.tech

Paul Mouginot, a graduate of CentraleSupélec University and the ESCP Europe business school, is an expert in digital culture and innovation and teaches those subjects at the Institut Français de la Mode. This “engineer-artist” also created the aurèce vettier studio in 2019 and is now starting a new adventure with Stabler Tech, a business specialising in data extraction. In the following conversation he talks about data’s use and usefulness in the world of fashion. 

How can artificial intelligence help fashion brands? 

An increasing number of fashion brands are using artificial intelligence (AI) in different ways. The fact is that AI – and more generally data – can be used to boost every stage in a fashion house’s value chain. Starting with creative tasks, it can applied to create moodboards built on digital signals such as texts, images, or sound. These elements can be used to feed models, especially generative adversarial networks (GANs), which then generate a kind of distillation that is neither cultural appropriation nor copying-pasting from other designers’ work. They are, nevertheless, complex tools to create, because you have to collect an enormous amount of information. 

That’s the core business of my new company, Stabler Tech, which I founded with my associates Anis Gandoura and Romain Hévin. AI can also help create more cutting-edge collection plans using data on sales, stock, and even information about competitors’ offers. Collecting data on a very large scale and a recurring basis is already a relatively complicated activity, and treating this information to extract pertinent guidance is even more difficult. 

Until now, this process was only available to big companies who had the means to set up the necessary infrastructures. But today, thanks to technologies like no-code and open sourcing and to dedicated online platforms, we’re seeing the widespread availability of these technologies. This is turn gives brands a lot of confidence in establishing forecast calculations for sales or in modelling demand curves for their products. AI is also a way to improve customer experience on ecommerce sites by helping generate more sophisticated product descriptions that are more enjoyable to read, thanks especially to algorithms like GPT-3. Recently in another sector I used this algorithm to write a book of poetry. 

Where do French businesses stand in terms of data treatment? 

Many businesses made good use of the difficult period during the pandemic and lockdowns to look into this topic. They were able to take the time to get equipped to manage data with data bases, product information management (PIM) apps, and data science tools. This let them store data and start to pull out helpful information. Some of those companies had been behind in this domain, but many of them were able to catch up. Of course, they’re not all at the same level; that depends on their size and available means, but there has been some interesting progress. Outside the fashion world, the Leroy Merlin brand, for example, has already announced in their stores that every day they collect 20,000 prices to verify their offer is competitive. They seem to have their own extraction technology, but smaller structures today have to turn to external experts for this type of mission. 

Data is important since it efficiently captures market evolutions and consumer expectations, needs, and demands. We see this with the Chinese brand Shein who can use data to launch a capsule collection at the first weak signal of a trend. But thankfully, everyone doesn’t want to be a Shein! Businesses with more eco-friendly ambitions can also make use of tools in this domain. 

Being well informed in fashion means, for example, knowing what the best sellers are and avoiding over production. And that brings peace of mind. Brands can also develop ranges that meet demand and then, at the same time, finance capsules or fun little pieces and keep creativity alive in specific ranges. Nothing will ever replace human awareness and, above all, we must not become slaves to a medium or technology. 

Your studio aurèce vettier combines algorithms, artificial intelligence, and artistic creativity. Can this be an inspiration for fashion brands? 

Questions about the link between technology and creativity have a long history. In the 1950s the composer Iannic Xenakis used Le Corbusier’s blueprints to write the scores for some of his operas. In the 1960s Sol LeWitt used an algorithmic approach in his paintings. aurèce vettier explores the consequences of this foundational research in art history and makes use of recent tools to find new forms. Even the studio’s name, aurèce vettier, comes from that process. It was generated from the names of the artists, authors, and scientists – men and women – who have inspired me for a long time. 

This way of working can be applied in a fashion design context. In fact, I’ve had the opportunity to work with brands in the sector. For Gucci’s 100 year anniversary I created an imaginary perfume bottle and described a fantasy fragrance for the event’s retrospective book Future Proxima. Other important projects are also happening, especially in the design world. In my opinion, the most interesting collaboration in fashion right now is the one between the American artist Robbie Barrat, known for his use of AI, and Acne Studios. The project involved the brand’s men’s winter 2021 collection. The designs were amazing with asymmetric forms and surprising textures. What’s key in this type of approach is that the presence of AI not be noticed. It should be used to encourage and enhance ideas. I’m not an engineer who believes in the power of a perfect machine that knows and solves everything. I believe that algorithms should, above all, stimulate us, build bridges, and help us find new paths.