"Some fashion companies launch a good 12 collections a year," Michel says. It is almost impossible to produce all designs manually. "It's a major stress factor for the designers."
For Michel, the solution is artificial intelligence (AI). In 2017, during her studies at the Berlin University of Applied Sciences (HTW), she founded her own research group and developed an AI that creates new designs autonomously. A year ago, she launched the software start-up Yoona Technology. Michel has already won over major fashion brands as clients, including German brands Ernsting's Family and Vaude.
Crunching the data
But can AI really do the work of a fashion designer? Michel explains that the software has to be trained with large quantities of data: drawings, patterns, cut pieces for it to work at all. Business data from previous collections of the respective company is also important - such as the colours, cuts, designs that sold particularly well. After analysing the data, the AI can crank out hundreds of new designs, colour concepts or even entire collections in a few seconds.
AI is no subsitute for the work of a designer, Michel says, but plays a supporting role. It can help designers create the basics for new collections. According to Michel, basics account for 70 per cent of what a given label sells. Influenced by new trends, they're updated and adapted for each collection. The Yoona AI is capable of performing this step.
"The designers themselves can concentrate on the 30 per cent where creativity is really needed," says Michel.
The whole process might also become more sustainable thanks to the AI, she says. The technology can calculate the precise amount of material needed based on the data of previous collections, preventing overproduction. The digitised designs also make the usual prototypes superfluous. All in all, this means less waste. In Germany alone, a good 400,000 tonnes of textile waste are generated each year.
Shopping advice, smart clothing, avatar models
Yoona Technology isn't the only tech company disrupting the fashion industry. Over the past few years, more and more start-ups have been entering the market with innovative solutions, according to business information scientist Ingo Claßen at the Berlin University of Applied Sciences.
Claßen says AI is already used in the field of personalised shopping. When we shop online, AI can suggest outfits to shoppers that match the items they have recently purchased on the site.
"Clothing itself can become intelligent," he says, referring to smart textiles that can alert you when you adopt the wrong posture, or that change colour depending on the outdoor temperature.
"Another exciting question is whether AI will be able to forecast styles and trends" based on the analysis of past styles, says Claßen, but adds that it's still too early to say whether that will ever really work.
The modeling industry is also in for a shake-up. 3D avatars are replacing flesh-and-blood models. The British agency "The Diigitals" already offers digital models exclusively. Yoona founder Anna Franziska Michel also uses avatars - at virtual fashion shows, for example.
"Designers can save an enormous amount of money this way," she says. Already in 2013, researchers at Oxford University predicted that modeling would be one of the professions that could be completely replaced by AI and avatars - both on the catwalk and in catalogues. Plenty of online stores already use virtual models. The Oxford researchers say real-life top models will keep working for well-known brands - but they will be an exception to the trend.
The Yoona founder is convinced that AI will and must become an integral part of the industry. "Our whole world is accelerating, including the fashion industry. Without AI, many labels and designers won't be able to keep up - especially small and medium-sized companies," says Michel, who hopes to sell her AI software worldwide.