New AI-Powered Walmart Store Will Help Define the Future of Retail
By Mike Hanrahan
May 20, 2019
For years, companies such as Google and Microsoft have been leading the way in the development of artificial intelligence software and platforms, while the technology itself remained largely untested in real-world retail settings. In fact, just 23 percent of retailers are using AI in store operations. With the launch of Walmart’s Intelligent Retail Lab (IRL) a few weeks ago, we’ve accelerated the timeline on practical AI applications in the retail space.
IRL is now implementing, learning from, and improving upon operational AI on a scale that is unmatched in retail. We’ve implemented cameras, sensors and state-of-the-art hardware throughout a 50,000-square-foot Walmart store. The scale of our AI lab inside a full grocery and retail location with 30k+ items in stock and more than 100 associates will empower us to make data-driven assessments and quick improvements to the various applications that IRL has spearheaded.
We have built an AI factory that will be producing and scaling solutions, providing insights that we can leverage to improve stores. And we believe that this ambitious project will position Walmart as a world leader in applied AI technologies.
Let’s take a deeper dive into how our team at IRL reached this milestone, what we’re working to achieve next, and how these innovations will lead to a smarter, faster, better shopping experience.
Fostering Innovation in AI
Advancements in sensor technology, computing, and software have led us to a pivotal juncture: AI’s time has come, and it is poised to transform the in-store retail experience. Teams throughout Walmart have been working on integrating innovative AI into our business, and IRL is no exception.
IRL’s position as a portfolio company of Store Nº8, Walmart’s incubation arm, has enabled us to iterate with speed while utilizing the deep operational expertise of Walmart. We’ve been able to experiment, fail, and learn while working with some of the latest technologies available. In the process, we’re positioning Walmart to succeed, as more and more innovative technologies transform the retail sector.
In the process of building IRL, we worked with groups throughout Walmart to identify some 300 use cases for AI in our stores. We narrowed that list down to not only what was technically feasible, but to what would provide the most value for Walmart and its customers. We needed to be both innovative and practical. Store Nº8 operates in many ways like a startup, able to take more risks and be more agile, but at the same time accountable to Walmart to produce real returns.
Transforming the Customer Experience
Unlike other AI-powered retail labs, IRL is operating inside a preexisting store — a Walmart Neighborhood Market in Levittown, New York — rather than within a custom store built to fit our needs. That means we’ve taken on a much bigger challenge. The IRL will help Walmart learn how to operationalize and productionize AI models in a real environment so we can build solutions that work in the real world and at scale.
What kinds of returns can retailers expect from AI, and how can it help their customers? The possibilities, in the long term, are endless. But even in the short term, IRL has been able to identify and develop some meaningful applications.
Within IRL, we’re working on a suite of products. Some are designed to track and restock items, while others detect the ripeness of our produce. Others still may expedite the shopping experience by alerting associates to open more checkout lanes or ensuring that online orders are ready for pickup. All of these products work together to ensure that we have what our customers want, when they want it.
Along the way, we’ll be deploying and testing a whole range of different sensor types, including everything from RGB cameras to depth-sensing cameras to load cells. The AI-powered technologies and techniques we’re using include computer vision, machine learning, and sensor fusion. Our capacity to use AI for real-time detection and response on a large scale sets IRL apart from other retail labs, and will have a direct and immediate impact on the customer experience.
Helping to Build the Future of Retail
IRL’s task now will be to determine how well each of these products functions and how scalable they are to other locations. IRL chose to launch in Levittown because it’s home to one of the busiest Neighborhood Markets in the United States. We wanted a store that would challenge us to test and develop the best AI possible — technology that, if it succeeded at this high-volume site, could augment any of Walmart’s 4,700+ stores.
IRL’s goal remains, as always, to learn and improve, and in Levittown, we’ll be able to do so with the help of our customers. As they interact with the store, we’ll discover what truly improves their experience. Meanwhile, we’ll be adapting and improving our products to better respond to their needs.
At IRL, we believe that AI has immense potential for good and the ability to unlock benefits to serve customers, empower associates, and improve our business. That’s why Walmart, a company that’s been people-led and tech-empowered, is laying the groundwork for a new kind of retail.