In our last FashionTech piece, we looked at how the widespread consumer purchasing practice known as ‘size-sampling’ contributes to unsustainability in the fashion industry’s downstream. In this third post, we will move upstream to examine issues affecting the midstream and upstream, to form a holistic view across the entire value chain.
Several issues plague the fashion industry’s midstream, resulting in the waste of enormous amounts of energy and the discarding of unsold garments. One of the biggest issues is the difficulty in predicting demand and future trends, leading to brands taking a “bet and hope for the best” approach to production thus leading to over-production and excessive amounts of inventory. In addition, without the necessary systems to assist with production and demand forecasting, fashion companies also lack visibility into their supply chains, making it a challenge to ensure stock availability in high-demand areas. To make matters worse, supply networks span the globe, from China to Hawaii, consuming significant energy for transport and logistics – yet most of these garments end up in the landfill anyway.
What can AI do?
AI allows for improved trend forecasting & predictive analysis based on big data, and more efficient inventory management. According to A McKinsey report, AI can reduce prediction errors by as much as 50% and stock amounts by 20-50%.[1]
AI-based solutions like Heuritech, which uses visual recognition technology to analyse social media images are already enabling the predictive analysis of upcoming fashion trends. There’s also Stylumia, which uses one-of-a-kind ‘demand sensing’ machine learning algorithms to validate trends, predict demand and reduce wasted inventory.
AI can also optimise inventory demand planning to minimise overstock and streamline production. AI-driven demand forecasting algorithms can predict which products to purchase to meet upcoming trends as well as the volume of products to purchase.
With such game-changing benefits, many leading industry players are taking action. Inditex, the world’s largest fashion retailer, has committed to invest €2.7 billion in online capabilities and technology upgrades.[2] LVMH and Google Cloud announced a strategic partnership in 2021 to speed up innovation and develop new cloud-based AI solutions to enhance demand forecasting, inventory optimisation and personalisation.[3] Nike’s transformation has hastened in recent years with significant investments in digital technologies and information systems to improve demand forecasting, insight gathering and inventory management.
While consumer habits fuel an incredible amount of discarded garments, the commercial side of the fashion industry is no less responsible for the generation of textile excess. Experts estimate fashion companies produce 40X more textile waste than consumers.[4] A big part of this waste stems from the garment design and production.
Before an item gets sold in stores or online, brands create numerous iterations and samples in the design phase. Creating different samples often involves mixing natural and synthetic materials, resulting in these garments ending up in landfills as they are non-recyclable. 80% of samples go straight from the runway to the dumpster along with 15-20% of pre-production fabric.[5]
Materials and dyes create significant emissions and contaminate the local environment. Many archaic dyeing practices are particularly damaging to the environment due to their tendency to leak contaminants back into local river systems and farmland. Water consumption is also extremely high, with the industry using over 5 trillion litres of water every year to dye around 28 billion kilograms of textiles. [6]
In addition to dyeing, water is a critical solvent in various pre-treatment and finishing processes, including washing, bleaching, and scouring. All up, the industry contributes almost 20% of global wastewater. It takes 5,000 gallons of water to produce just one pair of jeans and a t-shirt.[7]
Beyond these issues, the manufacturing process is inefficient and produces considerable amounts of scraps and deadstock without a useful outlet. Additionally, a lack of efficient tools to carry out quality control at scale also results in dissatisfied customers down the line, resulting in more garments in landfills.
As if all this was not bad enough, it is the low-cost, fast fashion garments that require mixing natural and synthetic materials. These garments are often over-ordered (due to size sampling) and thrown away when they enter the reverse logistics process or because they are unfit for recycling.
What role can AI play?
AI-based solutions have a significant role to play in transforming the current design and sampling process. 3D fashion design software and virtual garment sampling can help to reduce the number of physical samples needed to launch a design, reducing material use and the potential for waste.
Tukatech’s TUKAcad and TUKA3D software enable designers to create unlimited digital patterns and 3D samples without needing physical samples. TUKA3D also offers a “fabric feel factor value” feature that provides a 3D sampling experience that resembles the tactile experience of a physical garment sample without generating garment waste
Tukatech has also partnered with Sowtex to create a 3D visualizer and design-lab solution. The Sowtex Design Lab combines Tukatech’s 3D Visualizer with life-like digital swatches from over 10,000-plus global textile manufacturers to drastically trim the design and sampling process timelines. Companies like Hugo Boss and Tommy Hilfiger have already begun implementing 3D sampling practices to reduce textile excesses produced in the garment design stage.
AI can also assist in cleaning up the fashion industry’s dirty and wasteful manufacturing processes in a variety of ways. It can simplify and improve the archaic dyeing process. For example, Datacolor‘s SmartMatch solution employs AI and ML to reduce the need for unnecessary dye correction and the energy and contamination that comes with it. The solution improves first-shot match rates by enabling users to better understand the interactions between materials and more accurately predict how a set of dyes will perform together.
Another area within manufacturing where AI can have a substantial impact is optimising the use of raw materials to reduce future environmental damage. H&M and Tommy Hilfiger, for example, have recently implemented AI and predictive technology to understand the environmental nuances of each raw material and input that goes into their apparel supply chains. Further, AI is able to 3D-model yarn patterns that simulating the flow and fall on designs, ultimately providing a more sustainable alternative to synthetic materials.
And finally, AI has applications for quality assurance at scale to reduce and even eliminate the disposal of poor-quality garments down the line. For example, Cognex ViDi has developed a vision-based platform tailored for fabric pattern recognition that can discern good textiles from others.
Evaluating AI’s potential impact on emissions
It is challenging to quantify the precise impact of AI on emissions emanating from the industry. We know it is significant. We also know AI’s role transcends the entire fashion value chain.
Every year, an astounding 65% of the 30+ billion items produced for the fashion industry end up in landfills,[8] either due to overproduction or users discarding them and fashion, as a whole, accounts for 10% of CO2 emissions[9]. This means that AI can help reduce global emissions by 6 – 7% by eliminating the overproduction, optimising inventory management and ensuring item-user match.
Such a reduction would equal the damage done by all aviation and shipping activities or half the damage caused by all residential buildings on earth combined – a huge win for sustainable FashionTech, not to mention the planet.
Learn more about us.
