In late 2018, Amazon changed the grocery store market forever with their launch of Amazon Go, a convenience store without cashiers. Shoppers simply enter the store, pick their items off the shelf, and are automatically charged as they walk out. The technology is even able to sense when customers pick up an item and leave it somewhere else in the store without being charged. Amazon still employs a few attendants to restock shelves and help customers, but there are no cashiers. While the overall concept is simple, the implementation is quite advanced. Equipping a store with the necessary cameras, computers, and any other hardware costs about $1 million on average. Reviews of the first stores have been glowing as customers appreciate the increased convenience of avoiding checkout lines and the detail of the impressive features the stores are offering. While Amazon Go is a promising first step into AI stores, it is far from a finished product.
One of the biggest draws of Amazon Go is the ability to avoid cashiers and checkout lines. To enter the store, customers scan their app, but they don’t get billed until a few hours after they’ve left to allow time for any potential returns. While this system is functional, it certainly leaves room for improvement, specifically with facial recognition technology. Facial recognition could allow a competitor to bill people who enter their store without an app. One possible way of achieving this would be by setting up a one-time facial registration that is akin to the way the iPhone configures face ID. Once registered and linked to payment information, a customer would be able to come and go as they please without the hassle of scanning their app or fear of their phone running out of battery before they can enter. While this could increase the potential for shoplifting, stores could enact policies for handling people that aren’t recognized by the facial recognition technology. Stores will have video footage of a shoplifter’s face and could easily report them to the local authorities or have an in-house security team deal with them. While facial recognition could slightly increase the risk of theft, the benefits of expanding the ability to access a store and bring in new customers are worth it.
Google Assistant, Siri, and Alexa took the world by storm when they were released, but their creation is only the start of unlocking AI’s potential. One of the ways of improving upon Amazon’s foundation is by adding an AI program, such as an improved version of Alexa, with the ability to understand how different people talk. Right now, the three big name AI services can take commands and share Google results when asked a question, but they can’t fully understand the intricacies of different dialects or regional terms. This limits these services from implementation in automated grocery stores, the restaurant industry and other similar markets. For example, if AI technology were to be used in a Chinese restaurant for fielding delivery and takeout orders, it would need to be able to understand multiple pronunciations of “General Tso’s Chicken”. Even if people pronounce the dish the same way, if they have different accents, their voices could still confuse the AI. The technology would also need to know that some dishes have multiple names like potstickers and fried dumplings. In addition to learning how to handle orders, the AI needs to be able to adjust to people who change their minds. Many customers make spur of the moment decisions during calls or mess up their order and change it. While a human can understand this error and adjust appropriately, AI is not as forgiving. Once AI improves its ability to understand how humans operate and has an interactivity capability, it could be incorporated in stores as an assistant to ask about item location or price. The addition of AI assistants and facial recognition technology would allow stores to vastly improve the shopping experience for their customers and ensure they don’t get left behind as the entire in-person shopping market innovates.
