Walmart embraces agentic AI with major machine learning platform upgrade and developer ‘super agent’ – News

Walmart Inc. is giving its Element machine learning platform a major agentic upgrade and adding a “super agent” for software development to its growing portfolio of semi-autonomous software robots.

The retail giant is using its Converge 2025 technology event today to reinforce the message that it is all in on agents, the artificial intelligence worker bees that take actions and operate with minimal supervision. Element, its cross-cloud machine learning platform, is being refactored to orchestrate armies of agents. Wibey is the new super agent, designed to serve up the optimal tools to developers, resolve pipeline problems and generally enhance productivity.

Reimagining retail

In an interview with News, Sravana Karnati, executive vice president of global technology platforms at Walmart, said agents are more than an upgrade. They are the key to reimagining how Walmart builds and scales up intelligent systems.

“We’re entering a pivotal moment in the evolution of AI,” he said. “The focus is shifting to building intelligent systems powered by agents that can operate independently with reasoning, memory,and context.”

Enhancements to Element support autonomous systems with memory, reasoning and the ability to interact with multiple tools and application programming interfaces. The platform has been given a stateful architecture that enables it to track what agents do, say, and intend. This means agents can retain context — not just for a single interaction, but across long-running workflows that span systems and services.

Karnati said that’s important for tasks like identifying and fixing code compliance gaps, validating fixes in test environments spun up on the fly and issuing the necessary pull requests.

Cross-agent coordination

A plugin and tool-calling ecosystem allows agents to natively interact with external systems. Element supports standardized communication protocols, including the new Agent-to-Agent and Model Context Protocols, for coordination across agents and applications. API orchestration extends coordination across distributed environments.

Walmart said these features combine to enable Element to function as a platform for managing the full lifecycle of intelligent agents — from development and testing to deployment and monitoring. There are also observability tools that let teams track decision paths, reasoning steps and performance metrics.

Karnati said the changes enable Element to treat agents the same way it has historically handled machine learning models. Developers can register agents in an agent catalog, evaluate their behavior with built-in guardrails, and reuse them across multiple projects.

Wibey is a centralized interface for interacting with Walmart’s internal systems. It’s the fifth in a growing portfolio of super agents targeted at customers, associates and partners. Super agents essentially coordinate the activities of task-specific agents, of which Walmart now has about 200, Karnati said.

Built on top of Element, Wibey interprets user intent and routes commands to the appropriate tools, APIs or agents. It provides a unified entry point for executing common tasks by software engineers, architects and product managers.

Taming tools

Walmart has assembled a large arsenal of development tools over the years and sees Wibey as a way to organize and apply them. “Wibey isn’t a dashboard or portal,” said Karnati. “It’s an invocation layer that understands developer goals and coordinates execution across Walmart’s technology ecosystem.”

For example, Walmart has developed agents that can scan codebases for accessibility compliance issues, identify gaps, apply recommended fixes and validate those changes in test environments. Another set of agents monitors third-party libraries for outdated or vulnerable versions and can propose upgrades, test compatibility and create pull requests automatically.

Wibey integrates with popular development tools such as command-line interfaces, Slack and Visual Studio. It also supports decentralized development by enabling teams across Walmart to register and expose their own agents through shared protocols and contracts.

Karnati said one of Wibey’s core uses strengths is to generate tailored starter kits for software projects. Developers no longer have to search through internal portals for templates or services. Instead, they can use Wibey to generate a customized project framework based on a prompt and preexisting context, cutting down on time spent searching.

“It used to be that developers would have to manually look up the right APIs or templates,” Karnati said. “Wibey can develop a starter kit that is very specific to their project. They don’t need to make a bunch of modifications. It is ready-made based on the questions they ask, and it knows the context in which they’re operating.”

Karnati said the shift toward agentic systems is laying the groundwork for a new phase of AI adoption at Walmart, one that supports not only greater automation, but more intelligent and adaptive systems across all parts of the business.

In software development, agents can handle tasks like debugging deployment pipelines or detecting and fixing infrastructure anomalies. Customer-facing agents are already being used to personalize interactions.

The impact of agents such as Wibey could exceed that of traditional generative AI tools, largely thanks to their reasoning capabilities and integration with core enterprise systems, Karnati said.

“This isn’t just about code generation,” he said. “It’s about building systems that can make decisions, carry out actions and improve over time.”

Photo: Wikimedia Commons