Could AI hold the key to bringing fashion production closer to home?

Artificial intelligence could enable a new model of on-demand manufacturing that shortens lead times and reduces waste, boosting reshoring efforts. We’re not there yet.
Image may contain Accessories Glasses Adult Person Head Face Bracelet Jewelry Curly Hair Hair Happy and Smile
Photo: Vladimir Vladimirov

Sign up to receive the Vogue Business newsletter for the latest luxury news and insights, plus exclusive membership discounts.

Could artificial intelligence hold the key to reviving Made in UK?

Last week, the UK industry got a boost when British knitwear company John Smedley announced a £4.5 million investment in restarting its third-party manufacturing after a 40-year hiatus. However, the fact remains that less than 3 per cent of clothes worn in the UK in 2023 were manufactured domestically, according to Fibre2Fashion. The vast majority is still made in countries like China, Bangladesh, Vietnam and Pakistan.

Amid an onslaught of sustainability legislation that will demand much greater transparency in the fashion supply chain — alongside a recent pull back from China amid the US-China trade dispute — there’s a growing drive to bring manufacturing closer to home. But challenges remain for doing it at scale. Technology could play a transformative role, experts say.

“Unfortunately, the weavers, the spinners, the dyers — they don’t exist [in the UK] anymore,” says Simon Platts, CEO of circularity platform Recomme and former director of responsible sourcing at Asos. “The old days have gone, so it’s about how we use artificial intelligence and innovative technology to revitalise the industry.”

Elliot Barlow, manufacturing consultant at the UK Fashion and Textile Association (UKFT) believes AI has the potential to influence reshoring opportunities in the UK. He points to small-batch responsive manufacturing, which uses AI programmes to account for current stock levels while tracking market and consumer purchasing behaviour. Some brands and manufacturers are testing an on-demand model.

“AI can support this by calculating, optimising and reconfiguring workflows, bringing brands and manufacturers closer together,” Barlow says, adding that recent trials show that lead times can be reduced to as little as five to 10 days from order to completion.

These applications for AI are already being developed through various projects, such as those supported by the £1.8 million industry-funded Circular Innovation Fashion Network (CFIN), of which UKFT is a partner. However, Barlow believes it will be a couple of years before there is a sufficient critical mass of retailer-sized production orders going through the UK manufacturing industry to fundamentally determine whether AI can support reshoring at scale.

Reduced lead times, better forecasting

Academic institutions in the UK are driving research and development of AI and other technologies that could drastically improve supply chain efficiency, which would pave the way for an increase in manufacturing.

Among them, Future Fashion Factory is a £5.4 million research-and-development (R&D) partnership set up in 2018 to explore and enhance new digital and advanced textile technologies. It is led by the University of Leeds in collaboration with the Royal College of Art and University of Huddersfield, alongside industry partners. British mill Abraham Moon — one of the country’s last remaining vertical woollen mills — is working with Future Fashion Factory to scope out the feasibility of using AI across its operations. The primary focus is on developing a new AI-powered planning system that optimises production, rather than relying on spreadsheets and reports to decide what to produce, when and in what volumes, which often results in unnecessary waste.

There are two intersecting challenges that need to be addressed, says Stephen Westland, professor of colour science and technology at the University of Leeds. The first is reducing lead times and the second is improving forecasting. Traditional fashion supply chains are built on older models of production that can take up to 18 months from design to retail, he explains. “Because of that we have this trend forecasting industry that tries to predict what people want — but those processes are not very accurate. So lots of companies are making stuff that they can’t sell.” Westland points to colour forecasting as a prime example: “When I ask the 600 companies we work with, ‘What’s the single biggest reason you think you can’t sell things?’ Many of them tell me they’re making things in the wrong colour.”

Stephen Westland points to the colour forecasting industry as a prime example of the lack of accuracy within trend forecasting.

Photo: Berezko

Future Fashion Factory is working with companies like Heuritech, which analyses three million photos on social media daily using computer vision to better predict what consumers will be wearing for brands such as Dior, New Balance and Prada; as well as Unmade that uses AI to automate design and manufacturing processes to provide personalised, on-trend apparel. “Give people what they want, not what experts think they want,” says Westland, adding that AI is one of the best ways to do this. “AI will be used in almost all aspects of the supply chain. It’s making the whole thing so that data can flow along it very easily.”

The ultimate goal is to reach a model of “rapid agile” on-demand manufacturing, he says, which would drastically shorten the supply chain through highly digitised and interconnected production facilities. Under this model, the consumer purchases the garment, and then local micro-factories use advanced robotics and automation to make the garment in 18 days or less, according to Westland. “This is a complete disruption, because at the moment manufacturers make things and then later try to sell them. But through this model people can buy things online and then they’re made. And if you don’t make anything before you’ve sold it, there’s no dead stock,” he explains.

Westland predicts that in the next five to 10 years advances in technology will allow the creation of automated “smart factories” that utilise machine learning to continuously improve efficiency. Eventually, consumers will be able to co-design the exact garment they want, or at the very least, “retail will happen before manufacturing takes place”, he adds.

Heuritech analyses 3 million photos on social media daily using AI-based vision recognition software to better predict what consumers will be wearing for brands such as Dior (pictured).

Photo: Christian Vierig/Getty Images

What will it take to make this a reality?

More than half of fashion industry respondents to a 2022 Euromonitor survey said they planned to invest in cloud-based data collection tools, robotics and AI in the next five years. Though we’re still at the very nascent stages of this transformation. For this kind of innovation, a lot of investment and political backing is required — and funding has so far proved elusive.

There is some financial backing available. Future Fashion Factory is funded by the Arts and Humanities Research Council (AHRC), which has also awarded a £3.8 million grant to The Robotics Living Lab (RoLL), a fashion research facility that aims to help businesses develop high-value, low-volume garment production using agile, collaborative robotic technologies. Its new facility, set to open in June 2024, will embark on research into highly responsive, sustainable approaches for garment manufacturers as part of the UK’s reshoring effort.

CFIN, meanwhile, is backed by non-departmental government body UK Research and Innovation (UKRI). “But certainly there could be a lot more done in this area,” says Barlow of the funding available.

“R&D funding is crucial and the government needs to take the garment manufacturing sector seriously and invest,” says Susan Postlethwaite, professor of fashion technologies at the Manchester Fashion Institute, who heads up RoLL. She spotlights the Advanced Research and Invention Agency (ARIA) — an £800 million R&D funding agency created by the government and sponsored by the Department for Science, Innovation and Technology to develop technological breakthroughs — as another important initiative that could propel progress in this area.

A change of attitude is also needed. The goal should not be to return to mass manufacturing in the UK, but rather to use this opportunity to reset the way clothing is produced, says Postlethwaite. The focus should be on reducing the carbon footprint of fashion’s supply chain by manufacturing closer to home, and finding ways to reduce waste.

“Volume manufacturing is not desirable in a UK context. We buy more clothes per capita than anywhere else in Europe yet we are trapped in a throwaway culture where clothing is too cheap and we are entirely reliant on high street brands,” she argues. “The UK needs small-scale high-quality garment manufacture where the offer is local, more sustainable and the repair of cherished items becomes the norm.”

Comments, questions or feedback? Email us at feedback@voguebusiness.com.

More from this author:

How to define streetwear in 2024

Jewellery's next big opportunity: Divorce

The bumster is back on the runway. Will it sell?