The technology behind Amazon Go: retail IT taken to the next level
Having worked in the IoT industry for many years, I’ve somewhat become immune to the sparkle of new product or solution announcements. However, the announcement of Amazon Go recently simply stopped me in my tracks. Amazon Go is a new kind of store with no checkout required. No lines, no chip and pin machine. Customers simply walk in, pick up their products and walk out. However, I’m convinced that when it gets to the point that the first store opens and I’m walking out of the door with my goods underarm, I’ll be looking over my shoulder waiting to hear “’Ello ‘ello ‘ello”.
Amazon Go was due to open to the public by the end of this month, after launching in beta mode to employees in December. Reports suggest that the technology generally functions flawlessly if there are fewer than 20 customers present, or when their movements are slow. However, due to slight kinks in the technology used to automatically charge customers as they leave, and the difficulty of keeping tabs on an item if it has been removed from its specific spot on the shelf, it is now unclear when the store will open.
The delay proves we can’t take connectivity for granted. Amazon Go will potentially handle millions of transactions per day, and this includes the personal data of each individual customer that walks through their doors. This tiny glitch emphasises the need for a connectivity solution that is robust, secure and most likely cloud-connected. Distributed and mobile enterprises across many industries are embracing cloud-connected technologies to increase business agility, empower distributed workforces, and gain operational insights.
Innovative user experiences like Amazon Go are transforming how we connect and distribute data. As a result, an increasing amount of enterprise network traffic is moving off private IP networks and onto the public Internet. Technologies are being developed now for this new “Interprise” era, and companies are deploying private cloud networks over wired and wireless broadband Internet services.
The technology in the Amazon Go store is far from simple. It is supported by a complex system of computer vision, deep machine learning and sensor fusion that works seamlessly to keep communications running between the physical store, the customer’s smartphone and the free Amazon Go app. When a customer arrives, they scan their smartphone to enter, and begin shopping. Digital images of the stock and multiple sensors are used to detect which item is being selected for purchase. If there is any confusion, the deep machine learning aspect of the system kicks in, looking at previous purchases to ascertain which item it is most likely to be. This data is then registered with the customer’s Amazon account and a receipt is sent directly to their smartphone upon leaving the store. The artificial intelligence system comes together to create a retail experience with the convenience of online shopping and the control of walking into a traditional brick-and-mortar store.
By walking out of the store, the items are billed in real time to the customer’s online Amazon account so they can save time usually spent waiting in line. Considering that some surveys suggest the average UK consumer spends as long as 18 days a year in the shops, a concept like this – that saves time and merges the convenience of online shopping with the familiarity of the traditional retail experience – could means a significant reduction in the time it takes to do the weekly shop!
This exciting merge is a ‘digital first’ and going forward, IoT applications like this have the potential to fundamentally change how we connect and interact with each other. The norm is beginning to shift in the retail sector with the opening of this store, and many other sectors will soon follow suit. We’ll have to make sure the connectivity keeps up too!
By Hubert Da Costa, VP EMEA, Cradlepoint