Fashion and Artificial Intelligence: Technologies geared towards more sustainable manufacturing practices The 12 November 2024 - Première Vision Paris - Denim Première Vision - Première Vision New York

When it comes to aligning production with demand to avoid overproduction and boost the creative potential of designers, a number of artificial intelligence tools exist, like those of Livetrend, Heuritech and Refabric. For the garment manufacturing field, however, many hurdles remain.

Garment manufacturing is an energy-intensive process, particularly in factories lacking advanced energy-saving solutions: sewing and many intermediate or final stages (folding, ironing, packaging) are still mostly manual, and depend on energy networks that are often old, non-electrified, or powered by coal and gas. According to some studies, these assembly stages can account for up to 20% of the fashion industry’s total emissions.¹

During the production process, even the most experienced manufacturers are regularly confronted with production errors, inevitably resulting in fabric offcuts.

Due to stringent quality and compliance requirements, on average, 10-15% of fabric² can end up as waste during the garment assembly process

Added to this is packaging, often made of single-use plastic, and transportation – with garment manufacturing often outsourced to regions with lower labor costs – further increasing the carbon footprint of a product and the volume of waste generated.³

Faced with these challenges, artificial intelligence can provide concrete solutions to optimise every stage of production, from sourcing (such as FabricDNA) to the garment’s end-of-life, while facilitating communication between players and collecting essential data for traceability and extra-financial reporting. Thanks to machine learning, predictive analysis and automation, A.I. enables more efficient resource management, reduces waste and emissions, and improves product quality and time-to-market. From generating creative options through prompts to visualizing 3D prototypes, from body measurement solutions to optimized patterns, discover some of the key ways in which A.I. is transforming the entire garment manufacturing process.

Optimising pattern-making for maximum precision and material efficiency, realistic grading, and accurate material estimates

When it comes to creating virtual prototypes, A.I. has made major strides, generating 3D garments in motion that closely resemble real-life clothing, either using prompts, patterns or even sketches, as with Newarc.ai. Established 3D garment visualization platforms such as CLO3D or Browzwear help designers preview how clothes will look, drape and move on different body types and in various materials.

But beyond 3D simulation, these software programs above all enable designers to visually verify a pattern in 3D space to anticipate the rendering of the final product, starting from digitized 2D patterns (DXF files or a compatible format). This reduces the need for multiple physical prototypes and opens up an almost infinite field of exploration, while considerably reducing material waste (paper, muslin, fabric) and saving time and money.

Six Atomic (which partnered with Lectra in September 2024) takes streamlining the workflow to another level. Easily integrated into Browzwear’s VStitcher and CLO, this solution reduces design time from hours to minutes, and sends patterns straight to manufacturers. This combination of speed and accuracy results in correct physical samples on the first try, giving manufacturers a major edge in the highly competitive fashion industry.

A “base size” of a pattern typically generated through this process – often based on dummies tailored according to a brand’s target clientele in terms of morphology, such as those supplied by Alvanon – can then be accurately graded using software and sent to cutting machines, like those from Lectra, a major industry player.

In addition to automating the traditionally manual nesting task, A.I. impresses with its ability to anticipate pattern placement with unrivaled efficiency, providing estimates as close to reality as possible, using software such asGerber Technology’s AccuMark or Optitex’s Marker Making.This level of precision, including recommendations for enhancements that are barely visible to the human eye, help companies avoid the pitfalls of over- or under-ordering of materials, a costly problem in the textile industry. Lectra’s latest product, Valia Fashion, enables precise estimation of required fabric quantities, while synchronizing and streamlining operations via “digital twins”: each machine transmits its activities in real-time to the cloud, accessible to the entire production chain. This unprecedented responsiveness allows for near-immediate adjustments to meet specifications precisely, reducing fabric waste and improving production lead times. This saving in time and materials helps to reduce costs and, consequently, boost margins or offer more competitive pricing.

One of the major challenges remains the variety of morphologies. Drôme-based start-up Euveka tackles this problem with patented technology which adapts cuts to the actual body shape of customers, reducing unsold stock and brands’ carbon footprint, as well as returns. The company’s mechatronic robot, made from a material that mimics the suppleness of skin, measures the pressure exerted by the fabric on the body, switching from one morphology to another in less than 90 seconds.

The headache of production planning and the complexity of handling textile materials

“Highly digitized automated micro factories for clothing and fashion will be one important building block for on-demand local-for-local production which will make the wasteful long lead time offshore production model obsolete.”
Lutz Walter, Textile Innovation expert, Director for Innovation and Skills at EURATEX

La confection est historiquement une étape qui nécessite beaucoup de main-d’œuvre et qui est difficile à optimiser sans affecter la qualité. Comme l’explique Sophie Benson dans The Interline , magazine spécialisé dans la technologie pour la mode : « Les compétences subtiles et l’intuition des petites mains de la mode, couturières et fabricants en matière de manipulation des tissus, constituent le socle précieux pour l’apprentissage des techniques circulaires. »

Sewing historically represents a labor-intensive stage, and one that is difficult to optimize without affecting quality. As Sophie Benson notes in The Interline , a magazine specializing in technology for fashion: “The nuanced skills and intuition for handling fabric that garment workers have is a valuable foundation for learning circular techniques.”

It is crucial to preserve this know-how and expertise in textiles (so rich in diversity!), especially for circular approaches that promote sustainability, such as repairing and upcycling, which are also key pillars of extrinsic durability.

Meanwhile, artificial intelligence is beginning to automate certain tasks, from sewing to packaging. Sewbo, for example, uses A.I.-controlled robotic arms to precisely assemble garments by temporarily stiffening them. Although still in its infancy, this technology could enhance efficiency for certain types of garments, reducing labor costs and energy consumption. This automation does, however, raise social concerns. Excessive use could lead to massive job losses in developing countries, while the concentration of technology in a few large companies could accentuate economic inequalities.

At the same time, A.I. is transforming supply chain management, enabling real-time global monitoring of sales and demand, while optimizing factory operations, delivery times and inventory management (e.g. autone) through resource allocation. This interconnection between all players offers unprecedented agility and significantly reduces time-to-market.

Finally, platforms like Manny streamline order management and production planning in real time, based on available skill sets and delivery deadlines, enabling instant flexibility. The concept of “micro-factories” embodies this approach, with factories in Portugal standing out as exemplary models known for their speed, agility, flexibility, and eco-responsibility, alongside European players such as Rodinia (long time partner of Isnurh), which can help fashion brands cut costs, improve production and time-to-market considerably.

“Combined with highly digitized and integrated workflows comprising processes such as printing, cutting, sewing, piece-dyeing, folding, pressing and packaging, the emergence of highly automated micro-factories for efficient local on-demand production comes within reach. Such automated processes can also greatly benefit professional care, repair or remanufacturing as well as used textile sorting and disassembly.”
Textiles of the Future, Partnership under Horizon Europe, Strategic Research & Innovation Agenda, ETP

From the yarn to the finished garment: the promise of innovations in apparel manufacturing

While 3D knitting, or wholegarment knitting, is already an established reality, particularly with Shima Seiki machines that knit complete garments, woven garments until now still required cutting and sewing

But the revolution in 3D weaving is underway, with companies like Unspun and their Vega technology, which began with pants woven directly into shape, notably for Eckhaus Latta. In a similar vein, the start-up WEFFAN founded by Graysha Audren, is reinventing textile design by engineering the garment directly into the fabric using jacquard, cutting down on production stages, waste, and carbon footprint. By weaving two shirts with 50% less fabric, it reduces cutting waste by half. WEFFAN functions as a 3D-woven garment translator, proposing adaptable garment blocks in a design library, allowing brands to infuse their unique design identity into each piece, turning production into a genuine collaboration that spans the entire supply chain, from material selection to design to disassembly.

In the field of sneakers, Swiss sports brand On recently revolutionised the complex assembly of shoes with its Lightspray technology, based on the spraying by a robotic arm of a polymer that adheres directly to the sole, thus avoiding numerous cutting and assembly stages, reducing manufacturing waste and energy consumption. The result is an ultra-light running shoe with no glue or seams

While the integration of A.I. into apparel manufacturing opens up promising prospects for sustainability and improved communication throughout the supply chain, there are still limitations.

First and foremost, the high cost of implementing these solutions represents an obstacle for smaller manufacturers. So why don’t brands join forces to support this industry transformation? As Maxine Bédat, Executive Director of the New Standard Institute, points out: “The vast majority of carbon emissions in fashion come from the supply chain, which fashion companies share. No one brand is in a position to cover the costs of decarbonising their supplier, when their competitors would stand to benefit from it. With the Fashion Act⁴ in place, all major brands would be required to meet reduction targets and would thus be equally incentivised to step up and address their share of the burden in their respective supply chains.”

Il est également important de noter que, si d’un côté l’IA est particulièrement énergivore (en énergie et en eau pour refroidir les serveurs⁵), de l’autre, une partie de l’industrie n’est ni électrifiée ni digitalisée. De nombreuses marques restent éloignées de la compréhension de leur chaîne d’approvisionnement et des parcours amont de leurs produits, notamment au niveau du TIER 2, et la vraie transparence est encore rare⁶.

It’s also important to note that, while A.I. is particularly energy-intensive (both in terms of energy consumption and water usage for cooling servers* http://arxiv.org/pdf/2304.03271), a large part of the industry remains neither electrified nor digitalized. Many brands are still far from understanding their supply chains and the upstream journeys of their products, especially at the tier-2 level, and true transparency remains

It’s also important to note that, while A.I. is particularly energy-intensive (both in terms of energy consumption and water usage for cooling servers⁵, a large part of the industry remains neither electrified nor digitalized. Many brands are still far from understanding their supply chains and the upstream journeys of their products, especially at the tier-2 level, and true transparency remains rare⁶.

Digital investments should thus prioritize the digitisation of the supply chain to collect vital data and improve connections and communication between players, helping to identify key challenges and implement the most effective improvement measures, ultimately driving better traceability and transparency.

“For suppliers to remain competitive on speed, quality, price and compliance, objective, verifiable and real-time data is key. At Smartex, we call this type of data ‘Golden Data.’ ‘Golden Data’ enables an accurate understanding of the full cost of a product, a factory’s free capacity, clear quality metrics and environmental and social impact. In a lot of ways, this is the first domino that needs to fall in a factory’s modernization journey since it is a key enabler for improvement.”
The Modern Textile Factory, Smartex X The Interline

“Data should lead to an insight and an insight should lead to an action which is better informed than the one you took before.”
Matthew Dwyer, Patagonia, at the Textile Exchange Conference 2024

Optimising pattern-making for maximum precision and material efficiency, realistic grading, and accurate material estimates

The headache of production planning and the complexity of handling textile materials

From the yarn to the finished garment: the promise of innovations in apparel manufacturing

OTHER ARTICLES FROM THE FASHION MEETS AI SERIES