A lot of change has been compressed into the last year – from the shift in workplace practices to the surge of online shopping.  Adapting to this new and evolving environment has amplified the need for fashion brands and retailers operate efficiently, sustainably, and with as little risk as possible.  Needless to say, global supply chains have been disrupted in major ways by the pandemic, and by ongoing geopolitical tensions, causing many factories to halt productions, disrupting the flow of product, and exposing both brands and suppliers to risk on a scale they have never encountered before.

All models of fashion retail have suffered, but those that rely on high turnover of new styles at scale, on a seasonless calendar, have been particularly impacted: large volumes of stock on shelves and in transit went to waste as consumers swapped sparkly dresses for sweat sets. And that same shift in consumption patterns has also caused detrimental effects beyond retail and in the supply chain, leading to cancelled orders, sunk material costs, sweeping layoffs, and other negative outcomes.

Long accustomed to treating production as a tap that could be turned on at will, the fashion industry is now being forced to switch gears, and to look at its supply and manufacturing partners as finite resources whose capacity needs to be managed efficiently and, in the very near future, automatically.  Where once brands of all shapes and sizes, across every segment, envied Zara’s short value chains, now they are looking jealously at AI-powered supply chains like Alibaba’s new Xunxi AI-driven manufacturing ecosystem.

Because if the future of production is going to keep pace with the unpredictability and volatility of the outside world, then an agile, on-demand approach to sourcing and manufacturing could soon become essential, rather than an experiment.

The Journey to Industry 4.0

From the Industrial Revolution to the robotics being used in manufacturing, logistics, and distribution today, applying technology to automate manual processes is not a new phenomenon.  But the Industry 4.0 label captures a much larger and more consequential combination of technologies than ever before: hybrid cyber-physical systems, powered by AI, that both consume and generate huge volumes of data that has the potential to transform the way we think about production.

An example: traditional apparel production has always been driven by historical sales and fashion forecasts. Designing, sampling and shipping – it all typically takes up to a year before stock hits the shelves, making the act of manufacturing far removed from the point at which the idea for a product was conceived. Today, instead of relying on WGSN and other data sources to make near-term predictions on what to design and manufacture for tomorrow, brands are instead facing the need to produce for the present, and discovering the potential of reducing waste, eliminating risk, and maximising margins by only manufacturing what is actually needed.


But this, too, is not a new possibility. As far back as 2017, Amazon had secured a patent for automated on-demand manufacturing (ODM) where clothing is only produced after an order has been placed.  And this year that investment came to fruition with Amazon’s customised Made For You line.

Allowing customers to configure or customise their garments, though, is only one approach to producing on-demand. Brands should now also be looking towards harnessing IoT and AI to analyse real-time data from a range of different streams, and to translate those insights into automated decision making for demand planning. Such is the case with Finesse, a vertically integrated fashion house that uses Natural Language Processing and deep learning across community platforms to predict the next viral trend and optimise distribution, rather than basing its style decisions off catwalk concepts, influencer looks, or direct customer demand.  This is one further step ahead.

However a brand approaches the on-demand production, though, the common requirement is for demand – whether predicted by a machine learning mode, or actual from customer orders – to be met by smart, automated production methods.  The idea behind smart manufacturing composes mainly of computer integrated solutions and robotics that uses analytical insights to adapt rapidly. A combination of hardware such as robotics, sensors and software are deployed together to provide visibility and monitoring of production in real-time. Manufacturing-focused machine learning models are then trained for automation, to build efficiency in creating a more flexible and technical workforce. And although this may impose a threat to human labour, especially in a time of volatile demand, it is undeniable that computers can make calculations beyond the ability of a human, and the use of AI can enable both brands and their suppliers to make better decisions, improve efficiency, safeguard agility, and decrease costs. 

How Alibaba’s “New Manufacturing” puts these principles into practice.


Alibaba has already cracked the code for “New Retail” (a digitised, consumer-centric retail value chain), but the technology giant has now gone a step beyond by automating its entire supply chain. Their proprietary “New Manufacturing” ecosystem (dubbed the Xunxi technology) takes in cloud computing, AI algorithms and embeds Internet of Things (e.g. connected cutting machines and sewing systems) to develop an end-to-end manufacturing supply chain that can realise the vision for demand-driven production.

As personalisation continues to be a strong trend,  production becomes difficult to justify at scale, since it is normally capped by minimum order quantities and production flexibility limits. Yet, Alibaba’s model is specifically geared towards online SME’s that produce a variety of personalised goods at scale, offering as few as 100 pieces per item with only a 10-day lead time from production to delivery – an unbeatable contrast when compared to the months-long wait that has defined manufacturing for decades. 

Taking cue from Zara’s playbook, production efficiency is achieved by optimising the lay plan from the same fabric. Alibaba’s technology arranges items from the same material to be processed together even if from different orders and brands. The fabric cutting machine can determine how to fully utilize an entire sheet of cloth, enabled by the system’s AI. Even the raw materials are coded and assigned with a unique ID and QR code to enable traceability.

Throughout the entire production stage, the system provides transparency, making the status of every production line visible to every worker, and giving managers the ability monitor the entire workflow remotely on their computers and even mobile phones. The finishing touches are picked and packed by robotic arms and transported around the plant with autonomous guided vehicles.

According to the World Economic Forum, Alibaba’s “New Manufacturing” model has reduced the need for apparel merchants to hold inventory by 30%, shortened the delivery time by 75%, and even cut water consumption by 50%.  These are all major improvements to the metrics that any apparel producer will be benchmarked on – internally and by customers, investors, and regulators – making the investment case for a new approach to production clear.

Redesigning the supply chain

Even with the worst of the pandemic hopefully behind, and retail set to reopen in the UK and elsewhere very soon, fashion’s biggest burden remains excess inventory – both in terms of lost revenue potential and brand reputation. It is estimated that over 350,000 tonnes of wearable clothing end up in UK landfills each year – the equivalent of £140 million worth of stock – which is waste that both the companies creating those products and the environment do not need.

Under an ODM model, where stock is only produced when sold, smart manufacturing further presents the opportunity for retailers to rethink manufacturing in a way that completely avoids the need to create inventory in the hope that it will sell, assisted by AI.  The use of autonomous solutions such as robotics and AI-powered insights empower manufacturers with precise knowledge in preparation and forecasting based upon various factors in market trends, weather and the likes, enabling the “see now, buy now” culture of today’s fickle and impatient customers. 

The level of efficiency achieved with an AI-enabled supply chain is said to deliver over 65% effectiveness in reducing risks and lowering overall costs (Research and Markets, 2021). Cobalt Fashion’s lifestyle start-up 22 Factor offers designers a virtual sample of their garment, shortening from sketch to market by 92% compared to traditional processes. Migrating from traditional old formulas to new manufacturing models, retailers can forecast in advance top-selling items.

Yet another pressing matter within the supply chain is the availability of human labour. Factories were gravely affected by the pandemic, in addition to unforeseeable geopolitical tensions and local protests – causing a loss of manual labour and sparking a humanitarian crisis.  Where typically brands outsource production abroad for cheap labour costs, the issues of ethicality, sustainability, lack of manpower and long lead-times can be mitigated with technology, although brands must be careful about completely abandoning traditional production methods, and diversity in the supply chain is likely to be beneficial.

With smart manufacturing on the horizon, retailers are bringing back some production to domestic grounds, in hopes their technology investment and digital transformation initiatives will eventually offset the negatives to drive a smarter future.

Key Takeaways

The need to shift to new normal begins with the change within the supply chain: moving beyond seasonless models to shorter value chains that bring production closer to home, either physically, digitally, or both. Digitalisation is never an easy transition, and it can be a costly one too. And while this may be the largest hindrance to digital transformation, Supply-Chain-as-a-Service solutions are becoming increasingly popular as a way of removing that capital expenditure barrier, and that new industry is predicted to reach $15.5 billion USD globally by 2026. Case in point, Alibaba’s Xunxi technology and H&M’s Treadler have both now opened up their supply chains, offering their solutions as a service to brands of all sizes, and allowing them to leverage the efficiencies and economies of scale in smart production. 

Like in many use-cases and scenarios across all industries, technology should not be seen as disposal or substitution of human labour – but rather a tool that enhances the ability of humans to effectively make decisions, backed by real-time insights and accurate data forecasts to reduce risks and optimize costs. 

The impact of a pandemic has not just brought upon a crisis but more so revealed an underlying problem with the archaic and traditional ways of manufacturing and supply chain. To stay ahead of the curve, a digital model will become the lifeline for traditional factory operators who are struggling to compete. Rather than learning to react quickly, businesses need to switch gear from defence to offence – eliminating risk by redesigning their supply chains completely, rather than just artificially protecting ways of working that may already be becoming obsolete.