We regret to inform you that we will no longer be publishing Information Management. It has been an honor to provide you with the insights and connections to move your career forward. We wish you continued success on your professional journey and welcome you to explore our other titles at www.arizent.com/brands.

AI adoption at Walmart today and what’s coming next in 2020

As the 2019 holiday shopping season went into full swing over the past few weeks, keeping up with consumer demand for the most popular merchandise and grocery items while keeping prices low was incredibly important for Walmart.

With more than 1 million SKUs during the holiday season, and 4,700 U.S. stores carrying all this merchandise, that meant having to account for more than xxx product-store combinations and manage the logistics needed to get the right products to the right stores at the right time. And that didn't account for fulfilling e-commerce orders all over the world.

To do all that, we used artificial intelligence and machine learning to keep our operations running smoothly and our customers happy.

AI, in all its various forms, touches nearly every facet of Walmart’s business, including product forecasting, supply chain, logistics, marketing, merchandising, grocery, operations and so much more.

Because of our scale, it would be impossible for Walmart to operate without AI because our global supply chain processes orders at speeds that humans cannot match. As such, we continue to drive innovations in data science and technologies to improve our performance, such as using powerful GPUs to increase product forecasting frequency and accuracy and crunch more complex algorithms.

In 2020, we see continued advancement in our application of machine learning by leveraging one of the largest live datasets on the planet. By combining the insights from machine learning with expertise from our associates, we are able to leverage the power of AI to make optimal decisions.

To get a sense of how and why Walmart embraces AI at scale, it’s best to examine some of the ways we use AI, both behind the scenes, in stores and beyond.

Behind the scenes, we use machine learning for product forecasting, so we know what merchandise we need to buy, and when, to meet consumer buying trends and demands -- down to the store and shelf-level. We use it for optimizing supply chain and logistics -- how to get the merchandise from suppliers to our distribution centers to the right store at the right time, at the lowest possible cost. It helps us prioritize, organize and track what goes on our fleet of 60,000 trucks, custom-builds pallets for specific departments in specific stores to expedite off-loading, and plots the best routes to get products to stores quickly and cost-efficiently while minimizing environmental impact and driver fatigue.

Machine learning powers advanced technologies, such as computer vision to improve store flow, but it also helps us improve the quality and freshness of produce. We use it to customize what merchandise we display online and in stores, deliver greater ROI in our marketing efforts and optimize last-mile delivery to cater to individual customer preferences.

In stores, machine learning helps ensure optimal product availability on shelves and automates or accelerates certain time-consuming, data-intensive tasks, which frees up hours of time that associates can apply to delivering tangible value to our customers.

For instance, data from robotic shelf scanners along with a myriad of other data signals are fed into our machine learning models to help us track inventory in real-time and automatically alerts us and our suppliers when products are selling out, so that we can resupply our stores more quickly. It communicates with our Walmart Associate’s Downstock App, which tells them which specific products need to be unpacked from pallets and restocked first. It helps plan logical and efficient routes through the store for our merchandise and grocery “pickers,” which enables them to fulfill a dozen or more orders simultaneously.

At the same time, deep learning algorithms provides pickers with product substitution recommendations, based on the buying patterns of millions of other shoppers, if a customer’s preferred brand or quantity is not available.

Other applications for machine learning include voice-enabled shopping, where the technology has to understand both the content and context of the transaction, e.g., whether an interaction is part of the current query or the start of a new one; and be able to execute flawlessly.

Personalized shopping is another strong use-case for machine learning, which helps us make product recommendations and tailor the customer’s shopping experience and delivery preferences regardless of whether they shop in stores, online or both. We’re also using machine learning and computer vision to power our IRL concept store to better understand how consumers want to shop, and then using this data to improve our in-store, online and mobile experiences.

As you can see, data and machine learning are the lifeblood for keeping our operations running smoothly and costs low for ourselves and our customers. In the future, we will continue to adopt and develop innovative, advanced technologies and leverage more data to enable greater augmented and automated decisioning.

As amazing and powerful as these solutions are, we view AI and other tech innovations as enabling our associates to be more productive and empowering them to make better decisions. Because as the world’s largest retailer, our customers judge us not by the quality of our (mostly invisible) technology but rather by the positive shopping experience, quality merchandise and grocery items, and everyday low prices they’ve come to expect.

For reprint and licensing requests for this article, click here.