Chinashop 2025 opened on Thursday at the Shenzhen World Exhibition & Convention Center, attracting thousands of attendees. As one of China’s most prominent retail trade shows, the three-day event brings together industry leaders, technology providers, and retailers to discuss the sector’s ongoing transformation.
Two of the most anticipated sessions, from hotpot chain Haidilao and supermarket group China Resources Vanguard, provided detailed insights into how AI is already reshaping customer engagement, internal operations, and strategic planning.
Dancing with data: Haidilao’s human-centric AI
Few brands have become as synonymous with obsessive service as Haidilao, the hotpot chain famed for free manicures and noodle dancers. At Chinashop, its deputy GM of information technology, Yang Xuanzhi, described how the company is now infusing that same service ethos into machine intelligence.
“Looking at a timeline, we can see that our first attempt to understand AI was when many people were unsure about how to use it, and we treated it more like a toy,” Xuanzhi told the attendees.
“Last year, we began using it for material recognition, combining large and small models, allowing us to deeply utilise AI to address some business pain points.”
Haidilao’s approach focuses squarely on the customer journey. The brand’s AI systems, still limited in backend applications, are heavily focused on what Yang calls the ‘C-end’ or customer-facing interactions. Here, AI is a service enhancement tool – recommending dishes based on dietary needs, anticipating bill disputes before they escalate, and even predicting menu innovation by mining user wishlists submitted via its WeChat mini-program.
“Good service is not robotic,” Yang said. “They provide delightful surprises. For instance, when booking flights, I once received a reminder that my passport was about to expire, which I hadn’t thought about.
“At Haidilao, similar personalised recommendations are possible. If a pregnant woman asks whether she can drink Angelica soup, or if someone into fitness wants vegetable dishes, or if a person with high uric acid needs dietary restrictions, these queries require detailed knowledge from servers. What we do is filter all dishes and tag them so that AI can intelligently recommend based on user conditions. Providing better service requires understanding user tags,” he added.
One of the most ambitious AI projects, revealed for the first time at Chinashop, involves memory-driven customer profiling, recognising if a guest prefers a window seat or requests halal soup, and auto-integrating those preferences into future bookings. The goal is not just loyalty, but intimacy.
Internally, Haidilao is also automating knowledge retrieval for its 2000 clerks, saving human bandwidth for more complex tasks. The company estimates that 20 per cent of its operational tasks are now AI-assisted, including the subtle but service-critical act of tracking how long it takes to clear a table post-payment.
And yet, Yang was clear-eyed about AI’s limits. “You must understand your business deeply,” he warned. “AI doesn’t replace your brain, it just enhances your choreography.”
The national grocer’s gamble: CR Vanguard’s digital reckoning
If Haidilao is pioneering AI in the realm of service intimacy, CR Vanguard, one of China’s largest state-owned retail chains, is taking on a more systemic transformation. Its deputy GM, Li Jingyao, framed digital transformation as a necessity in a retail landscape torn between consumption upgrades and downgrades.
Borrowing a metaphor from Liu Cixin’s The Three-Body Problem, Li remarked: “New technologies arrive as unpredictably as Shenzhen’s weather. And our job is to learn to live in that climate.”
For CR Vanguard, the challenge is less about novelty and more about coherence. Store managers complain of fragmented systems. Strategic initiatives often fail to trickle down to shop floors. In response, the retailer is constructing a unified digital infrastructure aimed at making “data speak.”
At the core of the overhaul is customer understanding, not via traditional loyalty programs, but through advanced tagging systems that match product attributes with increasingly nuanced consumer profiles. From optimising SKU assortments based on regional data to deploying AI vision systems for remote inspections, Li painted a portrait of a legacy enterprise slowly but deliberately re-architecting itself around data.
“The goal,” he said, “is not to be high-tech for its own sake, but to empower staff with better insights and free up time for meaningful work.”
His vision is that of a lightweight, cloud-based ecosystem where frontline workers are not burdened by data, but buoyed by it, a workplace where checkout clerks double as informed brand ambassadors.
A convergence of philosophies
Though vastly different in style and scale, both Haidilao and CR Vanguard illustrate the same essential truth: AI’s impact on retail is no longer theoretical. It is now operational, tactical, and personal.
As Chinashop 2025 continues, the question is no longer whether AI will redefine the industry. It’s how humans will feel.
Further reading: Can AI really replace a creative agency?
