Automation (4)

Amazon Go - Automatic Grocery Store

In January of 2018 Amazon announced its first Amazon Go location in Seattle, Washington, would be open for public use. Previously, the store was only open to Amazon employees to act as "beta testers" for its technology. The main attraction of this store is the fact that it lacks a checkout line of any kind. Instead, customers walk in, collect their groceries, and walk out. Currently, 4 of these stores exist in the world. The most recent was opened only a few days ago on September 17, 2018 in Chicago, Illinois. The other 3 are all in Washington where Amazon's headquarters resides. 

How Does it Work?

Customers must have a smartphone with the Amazon Go app installed to shop at any of the locations. Upon entering the store, they scan the app to let the store know they are currently shopping. Using a combination of hundreds of cameras and weight sensors, the store can keep track of what item is being picked up by whom. Upon exiting the store, customers are simply charged electronically for whatever items they have with them.

The technology to make this work involves computer vision and machine learning. Computer vision aims to replicate tasks that can be performed by the human visual system. For example, analyzing and understanding components within images or videos. Machine learning is, in a way, a sort of automation itself. Essentially, machine learning writes a set of instructions (that would previously be written by a programmer) for itself by gathering data from thousands of examples. For example, Amazon Go's cameras and sensors used computer vision and machine learning to perfect itself over time. That period in which Amazon Go remained employee only was essentially a beta test to make sure the technology learns and functions properly. The longer the technology is used, the more efficient it will become at correctly determining who is picking up what.

In the event the system cannot identify what item someone has picked up, it resorts to other means. One method uses a customer's past shopping history and current shopping inventory to determine what the item is. The takeaway here is that the store relies on a myriad of technology to provide an accurate and efficient means of shopping without having to wait in a checkout line. 

Will Automation Replace Jobs?

Unfortunately, the answer is most likely yes. Fundamentally, automation means replacing the need for human assistance. It remains difficult for anyone to determine how big of an impact this technology will have on the retail job market. The technology is still new and limited to specific customers (people with smartphones) and locations. Amazon has announced its plans to have up to 3,000 cashier-less stores by 2021. Only time will tell whether or not this will really happen.