Amazon Go, Artificial Intelligence, and Humanoid Robots

amazon-go-artificial-intelligence-and-humanoid-robots

Amazon Go is a revolutionary new pilot project from the e-commerce giant, introducing an artificial intelligence (AI)-based, brick and mortar convenience store that aims to revolutionize retail. Using computer vision and deep learning techniques, Amazon will allow customers to walk into the store using mobile IDs, pick up their groceries, and walk out. While the details of the technology have yet to be released, Amazon will likely have cameras placed in the store, along with sensors, which will keep track of items being picked up or left on the shelf. This eliminates lengthy checkout lines and the need for human staff at checkout stations.

Tractica has spent some time analyzing the innovation gap that exists in the retail market. Given Amazon’s extensive supply chain footprint and close relationships with manufacturers, it is not surprising that the company wants to shake up retail using the latest innovations. In fact, more than 5,000 brick and mortar businesses closed in the United States during 2015, showing how traditional retail is struggling with the challenges presented by e-commerce and customer engagement innovation. Tractica’s white paper, Utilizing Humanoid Robots for Customer Engagement, recently published in collaboration with SoftBank Robotics, argues that:

Humanoid robots provide a playbook for customer engagement, and can be viewed as an evolutionary step from self-service kiosks to conversational commerce, combining the two in smart and unique ways.”

Bridging Online and Physical Worlds

Like Amazon Go, humanoid robots also use computer vision and AI techniques, allowing retailers to bridge the online and physical worlds. Humanoid robots are a natural evolution of customer engagement channels. They can add new dimensions to in-store analytics, provide emotion recognition and multilingual capabilities, and easily automate mundane and repetitive tasks. While Amazon Go targets a specific pain point in convenience store retail, which is the lengthy checkout line, humanoid robots provide retailers with a complete playbook in innovating around the customer experience. As outlined in our white paper, humanoid robots can have multiple use cases, including receptionist, product hero, e-commerce, customer survey, entertainment, and counter clerk.

While retailers have adopted self-service kiosks and introduced handheld bar scanners to speed up checkout lines, Amazon Go challenges the status quo and circumvents the payment task completely. Amazon Go uses the mobile app as the primary user interface. While details of the mobile app have not been released yet, it is assumed that the app will provide access to the list of items picked along with in-store search capabilities. A number of questions still remain, such as whether or not the app will perform facial recognition and tag the customer with a particular item, what happens if one mistakenly drops an item after picking it up, how returns work, etc.

Applying Computer Vision and Deep Learning

AI techniques like deep learning are really driving computer vision use cases, from self-driving cars to recognizing cancer cells in medical images. In fact, vision accounts for a large number of the top 15 AI use cases identified by Tractica. Amazon has already made major advancements in natural language processing (NLP) and speech recognition using Amazon Echo, which continues to improve over time. Echo-powered assistants in Amazon Go stores could potentially perform customer service tasks. Amazon Go is an innovative application for computer vision and deep learning, but it is only a sliver of what is possible when utilizing AI in a brick and mortar store environment.

In the end, the customer is at the center of innovation within the retail sector. Both humanoid robots and Amazon Go showcase the power of AI and its real-world applications, improving the customer experience and empowering disruptors like Amazon and other innovative retailers.

Comments are closed.