If the main theme of Chinese retail in 2024-2025 is the stock adjustment of "learning from Pang Donglai," then across the ocean, the world's largest retailer, Walmart, is undergoing a more intense self-revolution—a generational revolution driven by AI agents.
The past year has been lively for Chinese retail.
From Yonghui to Bubugao, from Hunan to Henan, we've seen a frenzy of store overhauls. We're busy adjusting shelves based on traffic flow, compressing payment terms from 90 days to 45 days, learning how to give employees "grievance awards," and introducing craft beer and viral mooncakes.
This is certainly a good thing—it's catching up on the old debts of product and service capabilities owed over the past two decades.
But when we zoom out and look at CES 2025 and Walmart's latest earnings call across the ocean, a strong sense of dislocation arises.
While our adjustments remain at the level of display, product selection, culture, and mechanisms, Walmart has quietly completed a more profound transformation—reconstructing the entire operating system from consumer search, supply chain flow, to employee execution through GenAI.
Walmart is no longer just a retailer; it is becoming a tech company disguised as a supermarket.
Today, we will not only list what Walmart has done in the past two years but also peel back the surface to explore how AI permeates every business link of Walmart's search, supply chain, store operations, employee management, and commercial monetization, driving a new growth flywheel.
Top-Level Strategic Adjustment:
From Omnichannel Strategy to Adaptive Retail Strategy
To understand Walmart's AI strategy, one must first understand the shift in its strategic definition.
Over the past decade, Walmart's core narrative has been "Omni-channel," which means integrating online and offline. Now, Walmart has proposed a new concept: "Adaptive Retail."
In this new strategy, AI is no longer a nice-to-have tool but a true infrastructure.
Unlike many companies that pursue AI for its own sake with flashy but useless consumer-facing applications, Walmart's AI strategy is extremely pragmatic, closely centered around the retail impossible triangle—cost, efficiency, and experience.
We break down its AI strategy into four core areas:
- Consumer side: Reconstructing search and purchase decisions;
- Supply side: Building a transparent and predictive supply chain;
- Store side: Digitization and automation of physical space;
- Organization side: Empowering 2.1 million employees with a super assistant.
Consumer-Side Revolution:
Generative AI Reconstructs the Search Box
The first stop in retail is discovery and capturing consumer interest.
In traditional e-commerce logic, search is keyword-based. Users type "chips" and get different brands of chips; type "T-shirt" and get different styles. This experience is fragmented and SKU-based.
But in real life, consumer needs are often scenario-based. For example, "I'm hosting a Super Bowl watch party this weekend; what do I need?"
In the past, you would need to search separately in the Walmart app: large-screen TV, chicken wings, chips, disposable plates, beer, decorations... This required dozens of clicks and searches.
Now, Walmart has completely changed this with GenAI.
Walmart partnered with Microsoft to use large language models to reconstruct its app's search function.
Now, users can simply type "Help me plan a dinosaur-themed birthday party for a 6-year-old," and the AI not only understands the complex semantics but also generates a cross-category shopping list: dinosaur cake toppers, green balloons, kid-friendly snacks, disposable tablecloths, etc.
The business logic behind this is highly impactful:
- From single-product search to solutions: This directly increases average order value. Traditional search paths are prone to missed purchases, while AI-generated lists are structured, significantly increasing cross-selling;
- Shortening decision paths: Reduces time spent jumping between pages, lowering bounce rates;
- Intent data acquisition: This conversational search gives Walmart more precise user profile data than keywords. It knows why you buy, not just what you buy.
Additionally, Walmart launched InHome Replenishment. This is not just delivery to the door; AI predicts when you'll run out of milk or eggs based on your historical consumption data and usage speed, automatically adding items to your cart or even delivering before you order.
Walmart is trying to eliminate the habit of European and American consumers shopping with a shopping list. When AI knows your fridge better than you do, it builds the highest business barrier.
Supply-Side Reshaping:
Symbotic and Inventory Forecasting
The essence of retail is the supply chain. If front-end AI is the face, then back-end AI is the real substance.
In Walmart's supply chain AI landscape, there is a core keyword: precise forecasting and automated flow.
- Inventory Forecasting
Traditional inventory management is based on simple linear extrapolation of historical sales. Walmart's current inventory AI model incorporates thousands of variables: weather forecasts, community events, social media trends, and even local flu outbreak data.
For example, if AI predicts an impending snowstorm in a region, it automatically increases inventory dispatch instructions for snow shovels, de-icing salt, and emergency food—all without human intervention.
- Symbotic and Automated Warehousing
Walmart isn't doing it all alone; it cleverly binds the industry's most advanced player through investment and deep cooperation—Symbotic.
In Walmart's distribution centers (DCs), Symbotic's AI robot systems are taking over. These robots not only sort at high speed but also have impressive algorithms.
The AI algorithm calculates how to stack products of different shapes and weights on pallets to maximize space utilization (reducing logistics costs) and optimize stability (reducing damage).
Furthermore, AI even decides the stacking order on pallets based on the store shelf layout after truck unloading—that is, reverse unloading logic.
This means that when a truck arrives at a Walmart store, the order in which employees unload goods not only matches the aisle order but also the shelving order.
This detail is mind-boggling: Walmart uses AI to extend its supply chain reach to the last meter of store shelves, greatly reducing backroom sorting pressure and labor costs.
Store-Side Evolution:
Digital Transformation of Physical Stores
For Walmart, with over 4,600 stores, physical stores are both an asset and a burden. How to make this massive offline store system intelligent?
- Electronic Shelf Labels (DSL): Not Just Price Changes
Walmart is deploying digital shelf labels (DSL) in stores across the U.S. Many think this is just for easy price changes. Wrong, very wrong.
In Walmart's AI system, DSL is a key node for inventory visualization.
- For employees stocking: When employees need to stock items, shelf labels flash to guide them to the right location. For large stores with tens of thousands of SKUs, this is an order-of-magnitude efficiency improvement;
- For employees picking: Walmart stores also serve as e-commerce fulfillment centers. When pickers pick for online orders, labels flash, significantly reducing search time;
- For customers: Through the app, customers can navigate precisely to a product's shelf location.
- Sam's Club Checkout Counters Are Disappearing
At Walmart's Sam's Club, AI is eliminating queues.
Sam's Club has introduced "exit gate technology" based on computer vision.
Previously, members using "Scan & Go" needed manual receipt checks at the exit. Now, when customers push carts through the exit gantry, high-speed cameras and AI vision algorithms identify all items in the cart within milliseconds and compare with payment records.
If matched, the gate opens automatically; if not, manual intervention is required.
This not only solves the retail industry's most painful "queuing" issue, improving experience, but also frees up exit verification manpower.
Organization-Side Empowerment:
Providing an "External Brain" for 2.1 Million Employees
This is the most overlooked part of Walmart's AI strategy, but perhaps the most humanistic and efficiency-enhancing.
Walmart has the largest workforce globally—about 2.1 million people. If each employee's efficiency improves by 1%, the value is in the hundreds of millions.
Walmart launched a super app called "Me@Walmart" with a built-in GenAI assistant called "My Assistant."
What problems does this assistant solve?
- Rapid onboarding and training: In the past, new employees with questions (e.g., "How to handle expired bakery products?") had to flip through hundreds of pages of employee manuals or ask supervisors. Now, they can ask My Assistant directly on their phones, and the AI provides precise operational guidance based on the internal knowledge base;
- Complex shift management: Employees can converse with AI, "I want to swap my shift next Tuesday to Wednesday," and AI automatically checks availability, matches policies, and processes requests;
- Even benefits inquiries: "What does my dental insurance cover?"
Many retail companies focus only on customers when digitizing, ignoring employees. Walmart's logic is "People-led, Tech-powered."
AI is not meant to replace employees but to eliminate tedious, repetitive, low-value tasks, freeing them to do what only humans can do—serve customers, tidy displays, and solve unexpected problems.
Business Model Loop:
Using AI to Make Money, Then Reinvesting in AI
AI is expensive. Computing power, talent, and hardware investments are astronomical. How does Walmart balance the books?
The answer: Walmart is becoming an advertising company.
This brings us to Walmart Connect (Walmart's advertising business).
Walmart has massive, real transaction data. Unlike Google or Facebook, which have interest data, Walmart has actual purchase data.
Through AI algorithms, Walmart can recommend sponsored products with extreme precision. When you search for "grill," AI knows you'll likely need "charcoal" and "steak" next, so it precisely pushes ads for related brands.
In fiscal 2024, Walmart's global advertising revenue reached $3.4 billion, up 28% year-over-year. This business has extremely high margins (typically 70-80%), far exceeding retail margins.
Thus, the flywheel effect emerges: AI optimizes user experience, users use the Walmart app more frequently, generating more data, making ad targeting more precise, increasing ad revenue, and using high ad profits to continue investing in AI technology and low-price strategies, further optimizing experience and reducing costs.
This is a perfect positive cycle. Walmart is using the internet's profit model to disrupt traditional retailers.
Summary:
Time Is Running Out for Chinese Retail
Writing this, my mood is complex.
We used to pride ourselves on China's retail leadership in mobile payments, live-streaming e-commerce, and private domain traffic. Indeed, China leads the world in marketing-side digitization.
But after seeing Walmart's AI layout, we must clearly recognize: in industrial-side intelligence, using AI to reconstruct supply chains, improve organizational efficiency, and hardcore automation technology, Walmart has already taken the lead.
Walmart's AI strategy offers us three profound insights:
AI is not a marketing gimmick; it's survival infrastructure. Stop treating AI as material for press releases. Walmart shows how AI penetrates the most mundane but core areas like unloading, scheduling, price changes, and inventory forecasting. Efficiency gains are the only antidote for retail to counter cycles.
Only by mastering scenarios and data can technology have value. Walmart succeeded not because its algorithms are stronger than OpenAI's, but because it has the most complete retail scenarios (4,600+ stores) and the most authentic data (decades of transaction data). Technology must grow on business.
People remain core. Walmart isn't trying to create unmanned stores with AI; it's arming employees with AI. In the service industry, warmth is still irreplaceable. AI handles efficiency; humans handle warmth—that's the ultimate form of future retail.
Walmart, this elephant, has not only danced but performed an extremely precise and powerful modern dance. For domestic retail peers, understanding Walmart's AI transformation may have more long-term strategic significance than studying how to renovate a few more stores.
The future is already here, and Walmart is clearly on that side of the future.