Thousands of robots collaborate with employees, using advanced computer vision and AI systems to automatically replace plastic materials with paper materials for packaging, reducing order processing time, and supporting more same day deliveries... Recently, a First Financial reporter visited Amazon's operations center (logistics center) in Nashville, Tennessee, in the southeastern United States. The combination of artificial intelligence and robots is redefining e-commerce.
&Amp; quot; We are undergoing a transformation in artificial intelligence (AI), and the innovation breadth of the Amazon team has never been so extensive& amp;quot;  On the same day, Doug Herrington, the Global CEO of Amazon Marketplace, was in attendance; quot; Deliver the Future 2024& quot; During the event, it was revealed that Amazon's new generation of AI driven operations centers have been put into use in Louisiana, and Amazon has also launched innovative solutions such as the conversational shopping assistant Rufus and the visual assisted parcel retrieval (VAPR) solution.
In fact, artificial intelligence is changing multiple aspects of the entire e-commerce process, including consumer purchasing, seller sales, and logistics delivery. Amazon's conversational AI shopping guides, AI sales assistants, etc. have already been implemented or released in beta versions. The combination of artificial intelligence and robots will increase the productivity of Amazon's next-generation operations centers by approximately 25%.
AI and robots support the new generation of logistics centers
Amazon's new operations center in Louisiana has five floors and covers an area of over 3 million square feet, equivalent to the size of 55 American football fields. It is one of Amazon's largest logistics facilities and will have 2500 employees when fully operational. It adopts a new multi-layer containerized inventory system; quot;Sequoia",  It can accommodate over 30 million items and expand seamless collaboration with robots, allowing employees to store and pick products faster and safer, helping to achieve faster delivery.
The Amazon Nashville operations center visited by the First Financial reporter is also one of Amazon's main operations centers, with the same robot team as the new operations center in Louisiana, but without the Sequoia system.
The reporter saw on site that after the customer selected the required goods for the order, the conveyor belt quickly sent the boxes containing the goods to the packaging station one by one. At the packaging station, staff operate automated equipment to package goods. They only need to pick up the goods from the box, scan the code in front of the computer, and put them into the automatic packaging machine. The packaging machine will automatically make paper packaging bags that match the size of the goods and seal them. The packaging material has changed from plastic to paper, and the filling material has also changed from plastic air pillows to paper materials. Amazon has retrofitted 120 packaging machines in 20 distribution centers, reducing the use of 130 million plastic bags this year.
Sorting stations are another key focus. The package containing the goods is transported by a conveyor belt to the sorting station. Robin, an AI driven robotic arm, sorts packages and places them one by one onto a blue drive (mobile robot) on top of a table tray. In a large enclosed warehouse, thousands of blue drives automatically move forward, turn, queue, and move on the ground, taking over goods and delivering them to the chute, allowing them to fall into the downstairs outbound dock, and then; quot; Empty body& quot; Come back and take on the work again, busy but not chaotic.
The outbound terminal is the final stop in the operation center process. Here, another AI driven robotic arm, Cardinal, automatically selects packages and categorizes them, placing them into carts corresponding to different communities or postal codes. The delivery truck from which community is ready to depart first selects this part of the package. Another mobile robot, Proteus, will move a cart filled with packages to a position near the dock door, in order to load them into trucks and deliver them to customers. Amazon uses big data to improve machine learning models. There will be robot rescue personnel here, and if any robot system malfunctions, rescue personnel will go to solve the problem.
Unlike the blue drive at the sorting station, which is restricted to a certain area and reads barcodes on the ground to move, Proteus is Amazon's first fully autonomous mobile robot that can effortlessly& quot; Drilling in& quot; The bottom of the cart containing the package, which is then transported to the outbound platform, can safely avoid employees walking in open spaces, looking like large green beetles.
The Amazon Nashville operations center processes hundreds of thousands of packages every day, and robots can sort nearly 95% of them. The integration of artificial intelligence and robots will increase the productivity of Amazon's next-generation operational facilities by approximately 25%& amp;quot; We will apply AI technology to robotic arms, such as helping Cardinal identify packages. Proteus uses AI to help determine the direction of movement, such as distinguishing between an employee and a cart in front of us, and it uses AI to avoid our feet& amp;quot;  Julie Mitchell, Director of Robot Sorting Technology at Amazon, told First Financial reporters.
Since entering the robotics field in 2012, Amazon has developed and applied over 750000 robots in its operational network. The industry estimates that the number of Amazon robots will exceed the number of employees by 2030.
Julie Mitchell stated that; quot; We deployed 750000 robots today and have over 1 million employees worldwide. Robots come in different sizes and shapes, and different robots work together to perform a certain task, while our employees in the same building make the execution of those tasks more perfect. So they are executing different parts of the same function, in this sense they are collaborative. Our robots that perform critical tasks cannot do without the help of our employees& amp;quot; The next generation operation center equipped with advanced robots will require 30% more employees in reliability, system maintenance, and engineering positions than before.
The application of AI and robotics technology is extending to; quot; The Last Mile& quot; Delivery in progress. With the help of& quot; Visual assisted package retrieval& quot;  The VAPR (Vision Assisted Package Retrieval) solution eliminates the need for delivery drivers to spend time reading labels or verifying customer names, addresses, and other information every time they stop for delivery. They only need to look for the green light emitted by the VAPR, grab the package, and leave. The ceiling inside the truck is equipped with cameras and LED projectors, which can recognize packages delivered at this station and project green light on them. VAPR will be deployed on 1000 Amazon electric delivery vehicles by early 2025. This technology was first used in the operation center.
Generative AI assisted shopping, selling, and services
When it comes to how to use general artificial intelligence to improve the user and customer experience, Doug Herrington said; quot; We are undergoing an extraordinary technological transformation; quot; 。 Amazon Marketplace has also undergone some changes in the past, with a significant shift towards mobile and an interesting shift towards social media. However, none of them have been as big or fast as the shift towards artificial intelligence& amp;quot; The toolkit provided by artificial intelligence for all our teams is amazing, from which we can achieve lower costs and better service quality; quot;。
Rajiv Mehta, Vice President of Search and Conversational Shopping at Amazon, told First Financial reporters that they are using AI to help users achieve a better shopping experience. Firstly, there is the Amazon AI shopping assistant Rufus, which is built using large language models (LLMs) developed by Amazon's internal scientists, using generative AI to assist users in making purchasing decisions. For example, when a user asks& quot; Gifts for Mom; Rufus will search and provide recommendations in the Amazon product catalog. Rufus also raises follow-up questions to clients, such as& quot; Tell me what your mother's interests are; quot;。 Once the user provides relevant information, such as their mother's love for traveling, Rufus can provide more useful recommendations, such as recommending a book about travel.
Secondly, Amazon has launched an app and mobile website in the United States; AI Shopping Guides& quot;， Help users who are not very familiar with a certain product, provide key information on the functions of different product types, as well as popular brands and features within that type. For example, if a user wants to buy a television, the shopping guide will provide& quot; This is different TV technology such as OLED or QLED; quot;、" What does outdoor TV or indoor TV focus on; quot; Waiting for information makes it easier for customers to find the products they need& AI Shopping Guide; quot; At present, it covers over 100 product types, and in the future, it will cover even more product types.
In addition, the function of Review Highlights is also driven by AI. It integrates and summarizes customer reviews of products purchased on Amazon in a simple and understandable way, helping users understand the positive and negative reviews of other users.
For sellers, Amazon also uses generative AI to provide marketing and sales assistance. Sellers only need to provide a product image and a short piece of text and other key information, and Amazon can help them use generative AI to create product detail pages, saving sellers time, energy, and costs. For example, helping a telescope seller generate an AI image of looking at the stars with a telescope at the edge of a cliff. So far, over 500000 sellers have used Amazon's generative AI listing tool. Amazon's AI tools can also help sellers create product videos and personalized content for different users. For what sellers sell and how to sell, Amazon has also released a beta version of its AI tool Project Amelia for sellers. AI sales assistants can provide suggestions and information, such as T-shirt pattern themes and inventory status. Amazon's supply chain optimization team also uses artificial intelligence to upgrade their predictive models, increasing prediction accuracy by 10% -20%.
&Amp; quot; The combination of artificial intelligence and collaborative robots will redefine e-commerce& amp;quot; Woody Merton, Vice President of Global Operations at Amazon, said.