Artificial intelligence, machine vision technology … What will the "future store" bring?

  Wal-Mart, a retail giant, recently unveiled its brand-new "Future Store". According to foreign media reports, this "future store" is actually a testing ground for emerging technologies, including cameras and interactive displays that support AI. The store is a working concept called Intelligent Retail Lab (IRL for short). IRL is a concept designed by Store No 8, a technology incubator of Wal-Mart, which runs a number of enterprises to test new ideas in the retail industry.

  From Amazon Go, an unmanned convenience store in Amazon, to Wal-Mart’s "future store", the exploration of new technologies such as artificial intelligence and big data in the retail industry continues to bring surprises and reverie. How futuristic is Wal-Mart’s "Future Store"? Is the future of retail really coming?

  The core is to improve operational efficiency.

  As early as 2016, Amazon took the lead in opening an Amazon Go experience store in Seattle, adopting "black technology" such as computer vision, deep learning and sensor fusion, which was refreshing. In recent years, various unmanned stores in JD.COM and Suning have also blossomed everywhere in China, but the retail giant Wal-Mart has given a different answer.

  According to Wal-Mart, this store is one of the busiest Neighborhood Market stores in Wal-Mart, which contains more than 30,000 items and can test technology in real environment.

  Similar to Amazon Amazon Go convenience store, a camera is installed on the ceiling of the store. But unlike Amazon Go, the cameras and sensors in Wal-Mart’s entire store are monitoring the shelves, instead of recording which goods shoppers choose, so that consumers can pay for these goods.

  Mike Hanrahan, CEO of IRL, said that this technology will enable employees to interact more with customers rather than replace jobs. Therefore, this store still retains the traditional cashier. The store covers an area of more than 50,000 square feet and has more than 100 employees.

  The camera here will monitor the inventory level to determine whether the staff need to take more meat from the cold storage to replenish the shelves, or whether there are some fresh items that have been on the shelves for too long and need to be removed. Artificial intelligence will help store employees know more accurately when and where to replenish products. This also means that customers will know that products and meat are always fresh and in stock when they arrive.

  "This ‘ Future supermarket ’ The main idea is to improve operational efficiency and serve staff, so that staff can have more time to interact with customers instead of spending time on inventory and sorting out goods. " Ren Feixiang, founder of Beijing Zhuoweizhi Technology Co., Ltd., which focuses on intelligent manufacturing and digital retail system research and development, told the Science and Technology Daily reporter.

  “‘ Future store ’ The biggest highlight is that technology has greatly reduced the tally cost of supermarkets, especially in Wal-Mart, which is a semi-‘ Labor-intensive ’ In the industry, the cost of tally is very high. If the tally is not timely or manual negligence, the goods will expire, and a series of problems such as overdue goods will cause cost increase or consumer dissatisfaction, and the products will be sold at a discount, which will increase the loss of enterprises. " Xiao Yi, the founding partner of Aoying Capital, also said that the product categories of a comprehensive lifestyle supermarket like Wal-Mart are complex, and it is difficult to accurately and timely understand real-time product information only by SaaS systems such as manual or simple ERP. Therefore, focusing on the back end of the "future store" to improve tally efficiency and warehouse circulation will inevitably reduce costs and indirectly improve consumer satisfaction.

  The calculation and storage of massive data is a challenge.

  Some analysts believe that IRL stores seem to have another purpose, that is, to help customers get used to shopping in an environment full of artificial intelligence technology. As Mike Hanrahan said, most people interact with artificial intelligence every day — — When chatting with customer service online, browsing videos and even searching Wal-Mart’s website, everyone is dealing with artificial intelligence to predict what we want. But even so, it is very likely that some customers who patronize Wal-Mart stores are not particularly familiar with artificial intelligence and may not be used to it.

  Therefore, the IRL store demonstrated these technologies.

  Cameras and other sensors in Wal-Mart’s "Future Store" can output 1.6TB of data per second, which needs large data centers to store. In the store, the data center is wrapped in glass, bathed in blue light and displayed to the public. But Wal-Mart said that the data center only stores less than a week’s data.

  "Wal-Mart’s way to improve efficiency is to collect, analyze and sort out shelf information with AI technology and machine vision technology, which will generate massive data, which requires very high computing power and storage capacity, and requires a special data center." Ren Feixiang said.

  Xiao Yi believes that it may not be the key to cultivate consumers’ consumption habits in the artificial intelligence environment, and the fundamental role of new technology application is to improve the operational efficiency of supermarkets. He analyzed that supermarket shopping is simply divided into three scenes: people, goods and stores, which correspond to the three processes of picking, selling and storing. "Artificial intelligence is a technical support covering the whole system of people, goods and stores, and it must play the role of commander in chief in such a complex scene as Wal-Mart, from big data processing of image acquisition to analyzing products — Inventory turnover replenishment — After the final settlement and even after sales, the participation of artificial intelligence is gradually accumulating more and more data. "

  "The only regret is that Wal-Mart said that their data is only stored for one week, which limits the analysis and processing of all data. I believe that if the database can be expanded, the efficiency of technology in the supermarket scene will be improved in all directions, involving supply chain, inventory, tally, settlement, after-sales and other links." Xiao Yi said.

  The two directions have not yet been perfectly combined.

  What will the future supermarket look like? Two and a half years after launching the first unmanned store, Amazon has now opened 10 Amazon Go stores in Seattle, Chicago and San Francisco.

  In a report released in September 2018, Amazon said that it has considered opening as many as 3,000 Amazon Go stores by 2021. Although Mike Hanrahan is dismissive of "dazzling elements", Wal-Mart has also opened a store without cashiers through its Sam member store.

  Pull your eyes back to China. At the end of 2017, more than 10 of the more than 40 unmanned shelf enterprises closed down. At the beginning of 2018, "GOGO Xiaochao" announced that it would stop operating, becoming the first enterprise with unmanned shelves in China to close down. Neighborhood Convenience, which was regarded as a dark horse in the unmanned convenience industry as early as its establishment, closed more than 160 stores in Beijing on July 31, 2018. But at the same time, JD.COM and other giants have not stopped exploring the new retail field.

  Forward-looking Industry Research Institute released "2018— The Analysis Report on Market Prospect Forecast and Investment Strategic Planning of Self-service Vending Machine Industry in China in 2023 shows that at present, the number and maturity of unmanned retail stores in China are still lacking, and the conditions for users to enter the market are not perfect, so the user scale has not yet been realized. With the promotion of internet technology and the penetration of the concept of unmanned retailers, the user scale and transaction volume of unmanned retail stores will usher in a blowout.

  "The core of retailers’ pursuit of emerging technologies such as artificial intelligence is to reduce costs while improving consumer satisfaction. The development of unmanned supermarkets also follows this concept. One way is to improve the efficiency of consumer checkout at the front end and reduce the cost of cashiers. One is to improve the turnover rate of goods, reduce losses, reduce costs and increase consumer satisfaction like Wal-Mart." Xiao Yi said that the domestic focus on the development of cashier is based on the advanced nature of mobile payment, while the United States is the originator of artificial intelligence and other technologies, and has strong scientific research strength. "So everyone will develop different ‘ Future store ’ Or the supermarket model. "

  "In short, everyone is trying to apply unmanned or efficient retail in various scenarios in all dimensions, and at the same time adding artificial intelligence technology to precipitate raw data." Xiao Yi said that in the later stage, it depends on whether the giants can combine various supermarket scenes to form a comprehensive artificial intelligence supermarket, instead of using artificial intelligence as a single application to reduce labor costs.

  "Some future supermarkets that have appeared before are committed to improving the customer’s consumption experience, such as fast payment and face payment. Wal-Mart’s new attempt is to improve operational efficiency, which is also the two directions of new retail in the future, but no giant has perfectly combined the two. "Ren Feixiang said.