Recently, Amazon has presented a concept called the future of retail stores. Its name is Amazon Go, and it is an offline grocery store that has no checkout process. You simply take the things you need, go out of the store, and the whole list of your items’ cost is taken from your Amazon account. This option is for you if you hate standing in lines, or searching for your credit card somewhere in your bag. It has no cashiers. Lots of people are amazed by this idea; however, let’s see if Amazon has everything neat.
There is only one Amazon Go place, and it is available only for Amazon employees for testing. This place is in Amazon’s hometown – Seattle. The store is not grand (1800 sq. ft.) and has lots of different food, from cheese to crackers. According to Amazon, to start shopping, you must have their account and download their app and go. The store for everyone else is not opened yet. The first opening was planned for 2017. You can subscribe on their website to get notification when the first store available for everybody is opened.
Amazon doesn’t give many details about testing by their employees. According to the brief information, the system is empowered with computer vision, has lots of different sensors, and works due to deep learning technology. The sensors all over the store work in combination with artificial intelligence. Their cooperation lets them know where the customer is looking and helps to recognize even those labels shadowed by some objects. That is basically all the information about the tech behind Amazon Go.
Don’t think Amazon tries to present something that fails as something big. Professionals say all the technologies involved into Amazon Go could actually work successfully due to the level of artificial intelligence technology, nowadays. Only Amazon knows for sure how it works, and we believe they can create something amazing.
As you might know, retail companies are ready to do a lot of things to track and understand the patterns of behavior of people in the store. They involve different technologies, and Amazon is no exception. Sensors in the store that see where people look are needed to get the information and to combine it in a statistic report form. It helps them aim different products at people better, change the layout of the production in the shops to increase sells, and it tells them the best place to locate advertising. If they have organized data on how their customers move though the store, on average, which products and places are most popular, they can not only increase their income, but also present better reports to the stakeholders.
Amazon Go is innovative in how it allows, not only gathering average and statistic information, but gathering individual offline purchase information. They can track customers through their WI–FI address and make smartphones something like a key to real-life purchases. However, it is hard to attach the product that leaves the store to a particular person in real life, and we haven’t figured out it yet. It might become the number one problem for Amazon Go and for other stores that would like to follow similar path. As it is hard to make something work with 100% accuracy, Amazon might be exposed to stealing, or in the opposite situation, Amazon can sign a person to the product he or she has never taken from that store. So, it is interesting in the way they will sort it out.
The shopping process is not something logical or calculated. Shoppers can move through the store for a long time, changing their purchases, placing them back, and taking them again. There are lots of food items, like fruits, that are weight–priced. If you shop with the family, only one will probably have to pay for the whole cart, so money is not withdrawn from the account of each member who visits the store or has an Amazon account on a smartphone. As you can see, it is very complicated.
However, Amazon is a well–developed organization that knows how to work in retail, how to tackle problems, and organize space and storage. They will probably have AI as helpers in different situations.
Specialist from AI startup Dextro – David Luan says Amazon probably has computer vision to see the direction of what customers look at. They also probably have trackers on the shelves, an app that helps to identify the person in real life with an account user (QR code technology is probably involved), and some kind of RFID to verify the items. However, the most important talent Amazon has – it is organized and works as a mechanism. They can sort out things easily identified through different ways, like cameras and sensors, and it will critically save them from lots of problems. Amazon controls everything, as it identifies the person who walks in, follows the process of purchase, controls space with sensors, and they probably know how to make it work. If something happens, they are in a beta-testing mode and can foresee lots of usual outcomes.
As for now, we must leave all the supposition aside and wait a little bit to see this future in action.
Continue reading about Technologies