Visions episode 3: Artificial Intelligence (AI) in Fashion: A Strategic Tool for Enhanced Precision The - Première Vision Paris - Denim Première Vision - Première Vision New York
Visions is a series of forward-looking articles that explores the world of tomorrow — its contours, models, and challenges. Through the eyes of experts from various fields, the Visions series undertakes an exploration of our societies, adopting multiple viewpoints to address several questions: How can we imagine the future today? What emerging trends or subtle signals should we pay attention to identify and anticipate new models and ways of living, creating, producing, and consuming? By gathering observations, analyses, and future-oriented insights in fields ranging from multidisciplinary creation to economics, sustainability, and new technologies, Première Vision aims to inform, inspire and provide key resources for creative Fashion professionals
AI is already well integrated into the business operations of fashion companies, notably driving improvements in segmentation and pricing strategies. However, it is still relatively underused in creative studios.
In the fashion industry, as in society at large, AI is shrouded in myths with a veil of uncertainty surrounding its practical applications. While today no AI tools exist that can replace designers and creators, it is being progressively integrated into various aspects of the industry to improve efficiency, precision, and impact. Clearly, this cannot be achieved without reinforcing the legal framework, especially regarding copyright issues, and without incorporating the idea of frugality to curb the energy consumption tied to its use. On a broader scale, AI comes at a cost both economic and ecological. While these costs are not out of reach, it’s worth noting that the brands most proficient in these technologies are primarily large corporations: the major e-commerce players (Zalando, Veepee, Farfetch), followed closely by luxury groups.
Paul Mouginot, artist and the co-founder of stabler.tech, has a unique background that straddles engineering and art (under the pseudonym aurèce vettier). He is an expert in both AI and the fashion industry and is scientific advisor to the Fédération de la Haute Couture et de la Mode. In 2016, he developed an AI-based competitive benchmarking technology, which was acquired by the online retail platform Veepee two years later. Prior to that, traditional programming required manual data entry into computers to get a result.
“AI offers a new way of interacting with machines, based on algorithmic training, enabling the machine to reconstruct the path to the solution by itself.”
Paul Mouginot
Fashion’s New Obsession with Technology
But what meaning will fashion brands draw from the world of technology? Throughout history, fashion has looked to other creative realms for sources of inspiration, notably art. Examples include Elsa Schiaparelli who collaborated closely with the Surrealists, such as with her iconic lobster dress, designed with her close friend Salvador Dalí, or the couturier Paul Poiret, who was also a painter and artist.
“This interest is reflected not only in collaborations, such as Lacoste and Minecraft, but also in the creation of games by brands, like Enigma, [an immersive digital game] created by Louis Vuitton on Discord”.
Fast-forward to today, however, and it is technology that is the key tool for the fashion industry to strengthen its influence, according to Mouginot. The blending of creative disciplines, together with new technological developments, is unlocking new realms of expression. As an artist, Mouginot, through his entity aurèce vettier, seeks to move from creative ideas to physical objects. To do so, he harnesses the potential of AI by feeding it personal data to generate new forms, which he then translates into artisanal techniques, creating sculptures and tapestries. It is with the tangible nature of this work and its rich narrative that we gain a deeper understanding of the tool’s potential to pave the way forward.
AI Already Well Integrated into the Business Models of Fashion Companies
Mouginot highlights three key ways AI can be utilized in the creative industries. First and foremost, AI will enable the segmentation of large data volumes by sorting them based on complex criteria. “This is incredibly useful in cases that require image recognition, such as when a brand or manufacturer wants to compare different collections,” he says adding that it also applies to the programming of machine tool production. “More and more cutting machines now have integrated AI systems that reconfigure cutting patterns to optimize material use“, he says.
AI is already a powerful tool for companies specialising in pricing and forecasting, especially regarding demand. When analyzing an infinite amount of data, it first helps refine competitors’ pricing strategies. When launching a product, “AI integration allows for a deeper understanding of customers: not only based on age, occupation, or gender, such as ‘a female executive from Marseille,’ but by identifying ‘clusters,’ diverse customer groups who exhibit similar behaviors. By inputting a combination of personal data and external factors such as weather conditions, the system can more accurately forecast demand and fine-tune supply accordingly.”
These two applications are now widely recognized in the industry, whether in textiles or beauty, with price adjustments made to the nearest cent. Given the resources needed to control these technologies, small and medium-sized enterprises (SMEs) are training in them at their own pace. The key issue for these businesses is to balance costs, enabling them to leverage AI without requiring excessive capacity or necessarily having to hire dedicated teams.
Emerging Professions
AI is also driving changes in internal company structures. Just as Mouginot has a career that straddles technology and creativity, fashion companies are bringing data scientists into their teams to work across different departments. “Ideally, engineers shouldn’t work in isolation but collaborate with other department heads,” says Mouginot. To enhance fluidity within organizations, new roles are emerging that combine a deep understanding of business and technological challenges. “Translators” act as intermediaries between data scientists and product managers, while “architects” act as liaisons between those setting up physical computer components and data scientists, who adjust the equipment according to cost factors.
Generative AI: An Inspirational Tool Still Underutilized by Brands
In creative terms generative AI is primarily used by creative studios as an inspirational tool, generating ideas based on the processing of large amounts of data.
“By feeding the machine with a vast array of inspirational data collected from various sources, like a moodboard, we create a refined synthesis of data that supports the creative process, but the tool can also be personalized by inputting only the brand’s archives, for example. As noted, most companies apply AI within their business frameworks and are more cautious when it comes to generative AI tools when the model relies on a broad range of data, due to concerns over copyright and intellectual property rights“.
Paul Mouginot
Multimodality and Frugality: the Next Frontier for Tomorrow’s Innovations
Looking ahead, the main challenge will be to personalise AI models using a company’s own data in a cost-effective way for, as Mouginot points out, it currently costs millions annually to operate a private LLM (Large Language Model). The aim, however, is to prioritize frugality — inputting a smaller set of data but using a more targeted corpus to better meet the initial demand.
Generative AI is expected to grow and tools will evolve to diversify the range of data beyond the words given to the machine — also known as ‘prompts’ — by incorporating more images, sounds, sketches, and photos. This will better correspond to the creative fields where this kind of AI is expected to be applied. This concept is called multimodality, but it is still in its early stages and remains very costly.
Ultimately, the goal is for each company and individual to find their own way of using the tool. From the earliest stages of production to pricing decisions, AI can be a powerful strategic ally. In terms of competitiveness and results, it helps boost competitiveness, especially when establishing collection plans. It also serves as a tool for driving the transformation of industry players toward greater sustainability “by experimenting with new materials and improving traceability within value chains,” concludes Mouginot. “Fashion’s adventures in technological affinities still promises multiple surprises ahead”.