• Kyodo


Apparel companies hit by the coronavirus pandemic have been tapping artificial intelligence technology to boost sagging sales, using it to predict the designs and colors that will come into trend.

One provider of such technology is Tokyo-based startup Neural Pocket Inc., which runs a system that automatically collects big data related to clothing on various fashion websites, Instagram and other social networking services to analyze day-to-day changes in trends.

The company predicts the next hit products by analyzing data on colors, designs and clothing lengths.

Such an approach using AI technology is new in the apparel industry, which generally relies on the experience and instinct of designers to predict trends.

But as the coronavirus pandemic takes its toll on retailers, demand is growing to efficiently predict the next hot items in fashion.

Neural Pocket said in April that its analysis revealed a change in trend colors amid mounting concerns about the pandemic. Its data suggested that clothing in whitish colors, which is known to impart a sense of security, was becoming popular.

The firm has tied up with an apparel group company of major retailer Aeon Co. to support its development of new items for the upcoming fall and winter.

Another AI venture, Sensy Inc., has developed software that analyzes data collected at “point of sale,” or the place where customers pay for products or services at stores, including information such as customers’ past purchases and the frequency of their visits.

Sensy provides such consumer spending data to TSI Holdings Co. and other major apparel operators, which will then use the data to optimize staff allocation at stores and develop new products.

“The apparel industry as a whole is becoming increasing conscious of costs due to the impact of the coronavirus pandemic,” said Sensy CEO Yuki Watanabe.

Zozo Inc., operator of major online clothing store Zozotown, said it began using AI technology in June to automatically form customer lists for sales promotion correspondence after analyzing such things as past purchases and tastes.

“We hope (such mails) will efficiently contribute to actual purchases of the products,” said Ryuji Noguchi, a senior official at Zozo Technologies Inc.

The coronavirus pandemic and the declaration of the state of emergency from April to May caused many apparel stores to temporarily close. Many customers remain cautious of visiting shops due to fears of virus infections, according to a department store industry group.

Consulting firm Roland Berger Ltd. estimates that sales in the domestic apparel industry in 2020 will fall up to 30 percent from a year earlier.

In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.