Amazon GO is a fully automated convenience store allowing its customers to buy their products in person as opposed to its conventional e-commerce nature. However, in comparison to normal retail stores, you do not require to interact with cashiers or a self-checkout to purchase items. Amazon can do so using a new type of shopping style they accommodate using what they titled as Just Walk Out technology where the process accurately described by the name: Walk in, pick up your items then walk out.
Billing is made directly towards your linked amazon account on your smartphone application, shortly after you leave the store.
Supporting the facility of the process uses a seamless collaboration of the most advanced technology used for shopping, which also is the same type used in driverless cars for monitoring and responding to objects. Amazon has named this as Computer Vision: The process of allowing machines to see what is in front of them and determine what an object is and to detect when an item has been taken from a shelf by a costumer who has taken it.
This is combined with their Sensor fusion: Combining data from many sensors weight sensors in the shelves to track individual products and deep learning algorithms.
Further support for the process is then required by a network of their information system (IS) architecture, ranging from the data gathering foundation of Transaction Processing systems (TPS) to the automated decision making of Expert systems (ES). Customers entering the shop require to go through turnstiles which can be bypassed by scanning a QR code from their Amazon GO app. To ensure this process would simply require a TPS to record the data of who enters the store and associate that person to the Amazon account which is then stored and modified as they shop. As the customers proceeds through the shop, a combination of different IS would be required to monitor the roaming customers and the items they interact with. Importantly, to aid their Just Walk Out technology, is an Organisational Information System (OIS) to allow seamless interaction between their sensors and an expert system for the AI, which decides the actions of customers, whether an item has been taken from a shelf or returned and by who, supporting the entire technology architecture. Customers wanting to leave can simply just walk out without having to use the app again. The sensors detect this movement, which again requires an ES to recognise and an OIS to relay between each tech and a TPS to store the data.
Although many IS is already required to run the store, it is only one side of the framework and can be inherently seen as a huge TPS system for the management side of the business to help further refine the system. Data gathered from customers shopping include everything associated with your Amazon account from personal details such as your name and address and financial details such as your card and bank information as well as your online shopping habits. With the addition of the store, it would then be capable of storing faces to their respective account holders. However, they claim the cameras only use infrared sensors to detect and track heat signatures. Data gathered in store is used to provide efficiency and increase profits. Amazon can track the number of people using the store at their respective times and how long each customer shops for to create reports on their peak hour performances. Additionally, they are able to keep track of the products in purchased in the shop to produce insights on items that are in demand and get rid of stocks that only take up space and use this to create a profile of the population demands of the town. Businesses can take advantage of this information to maximise profits. For a business to efficiently manipulate this data to produce useful information requires an addition Management Information System (MIS), to create reports for managers and allow daily decision making. A Decision Support Systems (DSS) is also required in the managing side of Amazon GO to allow support for such decisions by creating data models and statistical projections.
Despite using state of the art technology and multiple well-designed information system architectures, Amazon GO consists of flaws not normally found in traditional stores. The main issue is the requirement of the app to enter, and this inherently demands phone compatibility and battery charge, but it is stated that there are power banks located at the entrance of the store and once scanned through, it is not required for the phone to be turned on. It is possible to bring people with you inside by scanning your code for each person, however caution needs to be taken that whatever they take is billed to your Amazon account. Issues can also arise determining peoples actions if there are too many people inside the shop. This is however mitigated by having staff oversee the AI behind the scenes of the shop, similar to how supervised machine learning works.
Regardless of the few weaknesses of the new type of shopping, it also generates a lot of advantages and opportunities. The shop is designed towards being a convenience store as opposed to a supermarket and is perfect for people in a hurry. The advanced AI also has minimal errors, accurately determining items that have been picked up and returned on a shelf there has been mention of difficulty in shoplifting: All my efforts at shifting items around didnt fool the system at all. My career as a shoplifter ended just as swiftly as it started. This technology application is also at its beginning, creating the foundation for bigger shops in the near future that provide a larger variety of consumer selections.
Concluding the analysis of Amazon GO, the combination of the impressive synergy between the technology and the information system support systems applied has created a new pavement for the future of shopping and customer experience. At its current state, shelves are being restocked manually by staff, however in the future this could be replaced by an automated conveyor belt system and raises the questions of whether jobs could be as a natural consequence of automation, but it could also be argued that it creates the urgency of a more technical workforce to maintain. The future of entering could also involve a biometric system as facial recognition or fingerprint scanning to enter without the app as they gather more data, but this opens up concerns of privacy and it is clear that a distinct line is drawn to ensure the safety of the public is not compromised.