From Digital Operations to AI-Native Retail: How Dmall and Wumart Won the Smartest Store in the World 2026

Dmall and Wumart were named the global winner of the Smartest Store in the World 2026 at the awards ceremony in Poland, recognised for an AI-powered store transformation that demonstrates how AI can be embedded into real retail operations.

The winning case goes beyond a showcase of smart technologies. It demonstrates how AI can be embedded into the daily operating model of physical retail, connecting decisions, execution and feedback across merchandising, store operations, supply chain and customer engagement.

The award is jointly organised by NACS and Insight Research, with Accenture as the official knowledge partner. It focuses on how deeply technology is integrated into real retail operations, and whether it can deliver measurable business value.

In 2026, the competition brought together retailers from 12 countries across North America, Europe and Asia-Pacific, covering formats such as convenience stores, community supermarkets, airport retail, energy retail and unmanned stores. The final ranking was determined through a combination of professional judging and global voting, with retail experts assessing cases based on technology innovation, scenario-based execution and business impact.

Against this international field, the AI-powered store co-created by Dmall and Wumart took the top honour, competing alongside leading retail brands such as 7-Eleven Philippines, FairPrice Finest Singapore and FamilyMart China.

Why Physical Retail Needed a New Operating Model

Over the past decade, retail digitalisation has helped stores bring core workflows online. Orders, inventory, membership, payments and store operations have become more visible, connected and traceable.

Yet for many physical retailers, digitalisation has not fully solved the deeper operational challenge. Stores still often rely on experience-led decisions and reactive execution.

In merchandising, retailers may have large volumes of transaction, inventory and customer data, but still struggle to turn that data into timely commercial decisions. Questions such as what to stock, when to replenish, how to price, when to mark down and which products to remove from shelves are still too often handled through manual judgement.

On the shop floor, operational tasks such as shelf checks, planogram compliance, replenishment, expiry control, loss prevention and energy management have traditionally depended on manual inspection and fragmented coordination. This creates delays, inconsistency and a ceiling on labour productivity.

The challenge, in other words, is not whether retailers have data. It is whether they can turn data into decisions, decisions into tasks, and tasks into measurable results.

How Dmall Embedded AI into Daily Retail Operations

Dmall and Wumart set out to close the gap between decision-making and frontline execution.

At the centre of the transformation is an integrated AI operating loop built around decision, execution and feedback. Instead of placing AI on top of existing systems as a separate analytical layer, Dmall embedded AI into the workflows that determine how the store is run every day.

In merchandising, Dmall’s AI Merchandise Agent supports high-frequency decisions across assortment, pricing, replenishment, sales forecasting and product lifecycle management. It helps convert buyer experience, historical data and market signals into more systematic and executable strategies, improving both commercial accuracy and response speed.

In store operations, Dmall’s AI Store Agent serves as an intelligent coordination layer. It helps monitor operational status, identify exceptions and convert issues into executable tasks for store teams. Scenarios such as shelf display, inspections, loss prevention and energy use can be sensed, tracked and adjusted more dynamically, improving the consistency and timeliness of frontline execution.

The deeper shift is that AI no longer sits only in dashboards or reports. It becomes part of the operating process. Strategies are generated, tasks are dispatched, execution results are captured, and feedback flows back into the system to refine future decisions.

This is what makes the model fundamentally different from a conventional technology rollout. It is not just about adding tools. It is about rebuilding the operating logic of the store.

A Co-Creation Model Built for Real Retail Conditions

The project was developed through close collaboration between Wumart and Dmall.

Wumart provided real store scenarios, business data, operating standards and frontline feedback. Store teams tested AI-generated recommendations in daily operations and helped validate what worked, what needed adjustment and how tasks should be executed on the ground.

Dmall provided the digital platform, AI capabilities, model training, strategy configuration and algorithm iteration. Drawing on its long-term retail digitalisation experience and broad customer ecosystem, Dmall translated AI capabilities into practical workflows that could operate in real store conditions.

This co-creation model was critical. AI cannot scale in retail simply by installing a new system. It needs to be embedded into routines, roles and operating standards. Store teams need to understand not only what the system recommends, but why it matters and how to act on it consistently.

Through ongoing validation and iteration, AI gradually became part of the store’s daily operating rhythm.

Measurable Business Impact

The results show why the case stood out globally.

After the transformation, average daily in-store sales increased by 258.7%. In-store transaction volume rose by 144.7%, while average basket value increased by 46.6%. Product sell-through remained consistently above 90%, and inventory turnover improved to around 20 days.

Customer experience also improved. Customer satisfaction reached more than 95%, and active members contributed around 85% of sales, showing stronger loyalty and engagement.

Operational efficiency gains were equally significant. Intelligent energy management reduced overall energy consumption by 25.9%. Labour costs decreased by approximately 30%, while AI-enabled loss prevention reduced self-checkout missed scans and shrinkage by around 85%.

These results point to the real value of AI in retail. It is not only about better analysis. Its value is realised when AI is embedded into operational routines, connected to frontline execution and measured through business outcomes.

From a Winning Store to Scalable AI Retail Infrastructure

The winning case was not designed to remain a one-store experiment. Using the AI-powered store as a reference model, Wumart has already replicated the validated capabilities across 77 stores, covering different formats, regions and operating conditions.

This scalability matters. A successful pilot proves potential; a repeatable rollout proves operational maturity.

For Dmall, the case reflects a broader strategic shift. The company is evolving from a retail digital solutions provider into a retail AI agent provider. Its focus is to build AI retail infrastructure that is practical, reusable and continuously improving, helping retailers move from digital operations to AI-led growth.

The next stage of retail AI will not be defined by isolated algorithms or standalone automation. It will be defined by how effectively AI agents can work with physical AI capabilities, including AIoT-connected devices, intelligent hardware and emerging embodied AI applications, to connect business decisions with real-world execution.

That is what this case demonstrates. AI can become part of the operating model of retail. It can help stores make better decisions, execute faster, learn continuously and scale proven practices across formats and regions.

As the global winner of the Smartest Store in the World 2026, the Dmall and Wumart case shows that China’s AI retail practices are not only gaining international recognition, but also contributing a practical, measurable and scalable model for the future of retail.