The soon-to-open Amazon Go store in Seattle will be an interesting experiment in some technologies that, before now, haven’t been put into combined operations.
As Chris Preimesberger points out, sensor fusion, machine learning, computer vision, artificial intelligence and other technologies are getting wrapped up into a big enterprise retail application for the first time. Before Amazon Go, according to the company, the closest endeavor using so much varied technology was self-driving cars.
The first Amazon Go store currently is open to Amazon employees only, and the company says it will open to the public in early 2017. The store, which will look a lot like a specialized convenience store for tech hipsters, won’t be a full-service grocery store. Whether it can compete with even the typical 7/11-style convenience store is still unclear.
But what it won’t do is displace a large number of employees, even if it does eventually scale into a real grocery store or, eventually, some other type of general merchandise store. Despite the apocalyptic headlines in the New York Post, the reality is that the biggest change for an Amazon Go-like store will be more about the customer experience than about cutting back on jobs.
There are some important reasons. First, those products won’t get on the shelves by themselves. Somebody has to stock the shelves. While Amazon can use vendors to do some of the shelf-stocking, as other stores already do, this doesn’t take care of the majority of products. Perhaps robots can do some of this work in the future, but they’re not available yet.
Second, nothing about the Amazon Go plan as currently revealed will do much to stop theft. While shoplifting by registered customers probably won’t work very easily, that’s not the only form of theft that impacts convenience stores—or grocery stores, for that matter.
Here in Washington, we see an ongoing problem with teams of thieves that simply invade a store and steal everything they can grab before leaving. While both a cashier and an automated monitoring system can call the police if this happens, it’s not clear whether Amazon has found a way to stop this, short of hiring security guards.
There are other kinds of theft, including replacing a full container on the shelf with an identical empty one. Will Amazon’s tracking system be able to detect this? Or would there be need for another security guard?
The need for human intervention goes beyond crime. Customers sometimes will change their mind after picking up a product and then place in on a random shelf, rather than on the shelf where it belongs.
Worse, they’ll place a refrigerated product on a non-refrigerated shelf, which means a store employee will need to be available to return it to its proper place before it gets warm.