JB Header
Just Pick up and Walk Out with Amazon Go | Tech-enabled futuristic retail stores have arrived
Technology as an enabler finally seems to be reaching a stage where its presence can make a real difference to our everyday lives! Amazon Go, the latest venture from Amazon, addresses a latent need of every shopper - that of not spending any time in the billing queue. If the introduction video is anything to go by then Amazon Go will cut in-store shopping times by a significant amount, as the time spent in the billing queue and the actual billing time will both become zero! Literally!

What is Amazon Go

Amazon Go stores are retail stores where a customer just picks up and walks out with whatever stuff they need.

Let us take a look at a typical customer's activities in a Amazon Go Store

STEP 1 - The customer swipes the Amazon Go App screen on the turnstile before entering the store.
STEP 2 - He\She browses through the items on display and picks up the items which they want.
STEP 3 - Customer walks out with their picked up stuff.


No wonder Amazon has named the technology behind the stores as Just Walk Out! Because a customer is literally just walking out after picking up what they need.

You might be wondering by now that how the payment works? We all know that there are no free lunches! The good news is that checkout for the items picked up is done behind the scenes, your card is billed, and you get an online receipt after exiting the store.

Smooth right? Yes. But then how does Amazon know which stuff you have picked up? The answer to this question is what stirred me into writing this blog post. Amazon uses a combination of next generation tech to register what you ave picked up.Let's have a quick look at technologies running the show at Amazon GO.

Technologies working behind the scenes

  • Sensor Fusion Sensor fusion technology takes the input from multiple sensors monitoring the stuff on display, and fuses them up. When you have multiple sensors to collaborate and corroborate data, then your ability to 'deduce' becomes more accurate. Data from multiple sensors helps enhance the correctness of Amazon Go's readings of what actually happened in the store with a particular customer.
  • Machine Learning Machine learning, or deep learning, is an area where the system improves, or learns, as it interacts with events and records them. Result is a system which can predict and analyze customers' actions in a store using past data and derived understanding. I am guessing that Amazon Go stores will be able to display products and offers on specialized displays when a customer walks past them based on that customer's 'big data' captured in the system.
  • Computer Vision Computer vision gives systems the ability to understand what they see in images and videos. So, if a customer picks up an item, smiles and then keeps it back - this could be signal of a potential purchase lost due to 'some' reason. Or, suppose a person opens a product which is NOT supposed to be opened unless it is consumed, then an alarm can be triggered if this event can be registered by the monitoring system aided by computer vision. These are, of course, plausible scenarios cited by the author based on where the science is heading, which may become true in near future.

Want to learn more about Amazon Go Stores

Have a look at the official Amazon Go Page to see how it all works. The page also has a explanatory video clip showing the actual Amazon Go experience in action!