Walmart, in another move to capitalize on artificial intelligence technology, is looking to advance agentic AI functions into its operations to give consumers new ways to search, shop and purchase while optimizing workforce efficiencies.
Already Walmart integrates agentic capabilities into existing workflows across the business. Hari Vasudev, Walmart U.S. chief technology officer, pointed out in a blog post that early testing demonstrated Walmart agents work best when deployed for highly specific tasks to produce outputs that address complex activities. Walmart is developing agentic AI in house to deal with its specific systems and operations, instead of engaging third-party providers that are formulating applications to suit more general-use cases.
Walmart reported it is defining and refining core agentic AI capabilities that are foundational to how it intends to craft cohesive, best-in-class tools and experiences. In one example, Vasudev identified Walmart’s use of its retail-specific large language model, or LLM, in fashioning agents that can take on tasks such as item comparison, deep personalization and shopping journey completion to enhance the company’s GenAI-powered shopping assistant. Walmart is training its agentic AI on the company’s own data. As such, Walmart can tie its model to other LLMs to create responses and complete tasks based on specific contexts that can satisfy specific shopper needs.
Agentic AI, which can make determinations in applying data analysis to generate more specifically tailored results, is a next step in Walmart’s strategy in baking artificial intelligence into its business practices. Indeed, Vasudev said, many of Walmart’s GenAI-powered co-pilot tools are on their way to components of autonomous agents. Walmart said it is updating merchant tools that automate time-intensive tasks and trend-to-product analysis, shortening production timelines by as much as 18 weeks.
Walmart is exploring agentic systems with the intent of further optimizing endeavors across its technology ecosystem including in-store employee tasks, online shopping and merchandize planning. Walmart customer support assistant agents already are routing, resolving and acting autonomously to automate mundane duties and free associates to focus on more complex tasks. With the companies advancing artificial intelligence initiatives, Walmart’s GenAI-powered shopping assistant already is using multi-agent orchestration, fallback handling and evolving voice/camera capabilities to support customers from discovery to purchase.
Personal shopping, a goal of advancing artificial intelligence application, is an evolving AI-driven function that, Vasudev noted, will require a collaborative approach between retailers, providers and customers. He maintains that it’s not just about how Walmart refines personal agents. It’s also about how the company engineers infrastructure with providers and how shoppers learn to utilize the tools provided effectively. As it is now, a shopping agent acts in effect as a sophisticated web search with automated purchasing capabilities. To unlock agentic AI potential, two key pieces need to fall into place. First, customers need to train their agents effectively by providing specific query parameters, such as budgetary limits, brand preferences, sizes, colors and preferred store locations. Given feedback over time, an advanced agent will learn and adapt to individual needs. Second, retailers and providers need to construct bridges that support personal shopping agents by effectively communicating the data the businesses compile. With that, Walmart can validate whether it is meeting the needs of a customer’s personal agent and help facilitate purchases.
A trained agent can research, make decisions and take actions for consumers, Vasudev indicated. Walmart believes such capabilities will be particularly useful in shopping for essentials, given the regular cadence of purchasing everyday items. As the technology matures and users better understand how to maximize its potential, agents will be able to accomplish more complex shopping-related tasks, according to Walmart.
Agentic AI agents might search and shop somewhat differently and more pragmatically than the people they work working to help. As such, emotionally charged imagery might not have the same impact. The effects could go deep and even require fresh approaches to SEO, Vasudev observed.
Agentic AI, in Walmart’s case, is subject to careful evaluation as to what actions are best suited for autonomous agent execution and where human oversight and approval remain essential, the retailer said. Walmart said it is approaching agentic AI in a way that ensures the company is framing systems that are not only operationally efficient and trustworthy but also keep human experience central. The potential over time is for the establishment of complex personal shoppers that understand nuanced preferences, dynamic store environments that adapt to customer needs in real-time and self-optimizing logistics networks that do a better job getting products to the right place at the right time.
As it embraces artificial intelligence tech, Vasudev said, Walmart is addressing AI-related advancement with precision and intention, fostering a collaborative ecosystem and spearheading the next major evolution of retail.
