Amazon may have led the AI race in retail for years, but Walmart is catching up fast. The American retail giant has refocused its AI strategy around four powerful “super agents” built to make life easier for both customers and employees. A real-world case that could spark ideas for how you bring AI to life in your own e-commerce universe.
In Denmark and Europe, Walmart doesn’t make many headlines. But in the U.S., it’s the hypermarket of hypermarkets. Across North America, much of Africa, China, and India, chances are you’ll stumble upon one of their 10,000+ stores.
Simply put: Walmart is enormous. Bigger than Amazon, even. At least in terms of total revenue. Traditionally, Walmart has owned physical retail, while Amazon has dominated online. That balance may be about to change.
Walmart has just overhauled its AI strategy and introduced four new “super agents” to supercharge its e-commerce operations. The ambition? To double down on digital and make online sales half of its total revenue within five years - up from just 17.8% in 2024.
Meet Walmart’s new superheroes
Walmart’s four new “super agents” are the result of a broader strategy to consolidate its AI efforts and make things simpler. In early 2025, the company had built so many AI agents that it started to confuse users.
“It quickly became clear that we could simplify dramatically,” says Suresh Kumar, Walmart’s CTO and CDO. “If I have one agent helping with payroll and another helping with merchandising trends, I shouldn’t have to remember which one to talk to.”
So, Walmart is now streamlining everything into four super agents, each with a clear area of focus:
Sparky – for customers
Integrated into the Walmart app. Helps customers with everything from reordering groceries to summarizing reviews and suggesting recipes based on what’s in the fridge. It can even fill your shopping cart with everything you need for your next themed party - so no more excuses for skipping that Traitors-inspired 40th birthday.
Marty – for suppliers and sellers
Automates the onboarding of new vendors, sellers, and advertisers on Walmart’s platform. Streamlines order flows and introduces automated campaign and advertising tools.
Associate – for store employees
Provides centralized access to calendars, sales data, and HR tools all in one interface to save time and make internal processes smoother.
Developer – for the dev teams
Brings together testing, deployment, and innovation in a unified environment, enabling faster scaling of new features and AI tools.
All four agents are built to connect through the MCP standard, allowing them to call on smaller agents and access internal apps and data sources when needed.
Why Walmart’s approach stands out
Walmart’s super agents are built to remove friction. They take complex, messy processes and turn them into something effortless. Whether it’s a customer shopping online, an employee managing shifts, or a developer pushing new code.
But what’s truly remarkable is the discipline behind it. For years, the mantra across the industry has been: “Just get started with AI.” Every experiment, every prototype, every half-finished agent was seen as progress.
Walmart’s move signals a shift in mindset. Rather than spreading efforts thin across endless AI projects, they’ve condensed their focus to four crystal-clear use cases. It’s a sign that AI maturity is evolving. From curiosity and chaos to clarity and convergence.
In Double Diamond terms, Walmart has moved from exploring the possibilities to refining the solution - while most others are still wandering through the early stages of discovery.
Walmart’s four AI super agents represent an ambitious attempt to rethink the entire e-commerce value chain. Each one anchors AI in a real use context: The customer, the employee, the partner, and the developer. For every group, Walmart aims to build meaningful and relevant AI layers that act as true engines of efficiency and value creation.
Are they really that “super”?
It’s a fair question: Just how “super” are Walmart’s new agents, really? In the end, it all comes down to experience. Sparky can be technically brilliant, but if customers find it clunky or irrelevant, the magic disappears. And Associate can gather all the data in the world, but if it doesn’t make sense to store employees in practice, the value fades fast. These agents only become super when users feel that the tools genuinely make their day easier.
Regardless, Walmart’s case is a strong example of how to move AI from PowerPoint slides to real-world ecommerce impact. Instead of chasing grand, abstract transformations, they build one layer at a time - one agent at a time - all connected within a clear ecosystem and purpose.
For Danish (and European) ecommerce businesses, there’s inspiration to take from this: think in specialized AI layers that solve specific problems. Make them feel human. Ensure they help the user, not hinder them. And tie everything together in a structure that can be measured, implemented, and scaled.
Then maybe you’ll have your own “super agents” soon - though something slightly less ambitious might do just fine, too.
.webp)
.webp)