In the competitive fashion industrysucceeding with trends is crucial and supposes the difference between a sales success and a batch of unreasonable clothing. They count on SCMP As a Chinese startup is helping brands like Nike and GAP to make more successful decisions thanks to artificial intelligence. Analyzing social media data, online stores and consumption patterns, its technology is able to predict what will be fashionable before you see it on Tiktok.
The company. Zhiyi Tech was founded in 2018 in Hangzhouone of the most powerful technological centers in the country and where it also has its headquarters The giant Alibaba. The company focuses on offering artificial intelligence -based solutions that help clothing brands to predict what the upcoming supervent trends will be. In 2022 they got 100 million dollars in several financing rounds And today among its most notable clients are Nike and Gap.
Technology. Zhiyi’s flagship software (simply called Zhiyi) has been working since its inception and uses artificial intelligence to analyze large amounts of data. They presume that in their database they have more than 1,000 million images of materials or 100,000 million patterns of garments. They also analyze the designs that are best selling in stores, trends in social networks and monitor the sales of the rivals. With all this and through Deep-Learning and recognition algorithms with AI, they claim to be able to detect changes in the fashion industry and predict trends so that companies can design and plan more efficiently.
Local success case. The Chinese fashion chain Urban Revivo He is one of Zhiyi Technology clients and with which they have been working since 2020. According to Zheng ZeyuCEO of Zhiyi, in the first three years Urban Revivo managed to improve its Sell-Through rate (stock proportion sold compared to that received) from 60 to 80%, as well as improving its opportunity to create very successful products by 5%. The brand launches more than 100,000 designs per year, so we would be talking about 5,000 exceptionally successful designs.
From China to the world. In the beginning, the startup was focused on the domestic market. Almost all the sources he drank were sales and social networks of the country such as Taobao, Tmall or Rednote. In its opening to other markets, three years ago Overseas Exploration launched, a tool that was also aimed at Chinese companies, but already explored trends from the rest of the world.
Trendscopes It is one of Zhiyi’s first products with an focus on international customers. It was launched last year and with it trackean global sales platforms such as Amazon, Temu or Shein (China’s Great Baza in Fast-Fashion), In addition to networks such as Instagram and Tiktok. Among its most important international clients are GAP and Nike, which they help in decisions such as choice of color for sportswear or if exterior footwear must have a high cane.
A chatgpt that designs clothes. In addition to its analysis and prediction tools, Zhiyi also has a generative -called AI Fashion diffusion+ which allows to create “high quality designs and images from moodboards or sketches.” The tool offers a free test if you register so we have tried it, specifically the option of turning a sketch into a render, although it offers more like adding a garment to a photo of a model. When generating the image, it allows us to choose if we want it in color, the similarity with the sketch and add a descriptive prompt.
Test with Fasion Diffusion+.
The result is good in that it represents very well all the details of the pockets and zippers, but the improvable quality and remembers what Dall-e did in his first versions. We have asked Chatgpt and the result has been infinitely better, especially representing the photorealistic effect we had asked.
From left to right: sketch, fashion diffusion+ and chatgpt.
The company wants to turn this tool into a kind of chatgpt with which to create designs from Prompts With natural language, without having to start from more detailed images or drawings. They hope to launch it this month and promote it especially among its international clients. There are still no images of the new version, but seen the performance of the current one, there is work to be done.
Other fashion chains also bet on AI. The AI is reaching everywhere and this Chinese startup is not the only one that has occurred to the textile industry. Already In 2018 Inditex began investing in Big Data, AI and Analytics solutions For his businesses, he specifically associated with the Startup Jetlore that offers a prediction of the behavior of buyers.
Uniqlo also integrates AI in your business And they do it in an integral way, from the design process to the supply chain. Similar to Zhiyi, they use data analysis and collection technologies that allow them to detect new trends in order to advance to the market. They also apply it to their stores registering the movement of customers to define “hot areas” and thus better configure the space.
AI as a predictor. The generative AI is the most talk, but its uses go much further and more and more are the sectors in which it is acting as a prediction system. The most obvious example we have In advertisingwhere the algorithms have been wanting to guess what we will buy. The financial sector is also using AI to predict credit risk and for detect and prevent fraud. In science we have also seen how they use it to try predict the next epidemic. Perhaps the biggest challenge of the predictive AI is in health, where tools are already appearing predict diseases and customize treatments. Here ethical challenges are raised and the need for constant supervision and update of algorithms.
Cover image | Pixabay
In Xataka | Openai is going better than expected thanks to payment users. And that marks an evident direction
GIPHY App Key not set. Please check settings