ChinaTravelNews, Ritesh Gupta – Meituan is shifting the frontier of local commerce and the travel segment by pairing a focus on operational efficiency with a heavy architectural push into foundational artificial intelligence (AI), turning open-weight models like LongCat-2.0 into an ecosystem moat.
As local commerce and quick commerce mature, market priorities are also shifting.
As Meituan’s CEO Xing Wang noted during the recent earnings call, “In the second quarter (Q2), the quick commerce industry shifted its focus to improving marketing and operating efficiencies.”
At the same time, Meituan is placing greater emphasis on personalised engagement. Wang also highlighted this shift.
“We saw a clear shift in local services consumption from standardised spending towards more personalised, experiential, and emotion-driven experiences. Authentic experiences and deep engagement are playing an increasingly important role in consumers’ decision-making. We believe that in the AI era, authentic experiences, authentic reviews, and authentic trust will continue to be the most vital infrastructure for the local services industry.”
For a purchase as complex as travel, reviews are nothing new, but they continue to play a vital role in the travel booking funnel. In China, authentic reviews remain an important influence on traveller intent and booking decisions.
Meituan is anchoring its AI travel strategy in authentic user feedback.
Wang mentioned that the company released the latest edition of the Must-Eat List (“必吃榜”), leveraging 1.5 billion authentic user reviews in Q2. It now covers 264 cities and regions, with 120 newly added cities.
The Must-Visit List (“必玩榜”), too, has grown, covering a lot more interest-based consumption scenarios.
This massive repository of user feedback also gives Meituan a rich data foundation for understanding consumer preferences and improving recommendations.
Also, as consumers demand more intuitive planning tools, conversational interfaces are stepping in to replace rigid directory searches. Inside the Meituan app, this is anchored by Xiaotuan, an AI-powered conversational search assistant.
Highlighting the rapid adoption of this tool, Wang stated, “For complex local services decisions, consumers increasingly turn to Xiaotuan, our AI assistant, built in Meituan App. As adoption grows, we are seeing a meaningful lift in user engagement on Meituan.”
For travellers planning complex trips (such as coordinating multi-stop itineraries, boutique hotel stays, and dining reservations), Xiaotuan moves past simple chat. It acts as an execution engine, translating multi-constraint natural language inputs into single, seamless bookings. Furthermore, through merchant-facing solutions like the CatPaw platform, which now features specialised AI agents for hotel companies, Meituan is shifting from a simple digital listing channel to an active “AI business partner.” The platform acts as a digital copilot, streamlining daily property operations and inventory management for hoteliers.
LongCat-2.0 and the architecture of local intelligence
Beneath these consumer-facing touchpoints lies Meituan’s heavy infrastructure play in LongCat-2.0.
Addressing the company’s broader technological roadmap, Wang explained, “I will begin with LongCat-2.0… an important foundation for our AI strategy. What we are really focusing on is evolving our AI to bridge the digital and physical worlds, and doing so deepening Meituan’s competitive moat in local services over the long term.”
“For us, AI is less about competing on the models. It’s more about leveraging AI to reshape our organisation, product, and workflow.”
Meituan’s strategy relies on three core pillars: building large language models, deploying AI at work, and embedding AI in products. LongCat-2.0 represents a major milestone as an open-weight, trillion-parameter foundation model trained entirely on Chinese infrastructure.
“LongCat-2.0 has made significant progress in core agentic capabilities, particularly in coding, reasoning, and tool calling, and complex task execution. It has gained positive feedback across the global developer community. We believe AI creates a lasting value where it’s deeply embedded in real workflow, and where it can solve real problems as reliable infrastructure. Our full-stack domestic infrastructure for training and inference gives us a structural advantage in cost infrastructure control over the long run,” said Wang.
From discovery to execution
This technological depth directly dictates the product-level evolution of the superapp. On the product side, Meituan continues to scale Xiaotuan, moving from understanding complex multi-constraint queries to full agentic task execution. Ultimately, the goal is to deliver a seamless closed-loop experience that takes users from travel discovery and decision-making all the way through to transaction and fulfilment.
Organisationally, this AI adoption is deepening internal productivity, with a rapidly climbing share of code and workflows generated by AI tools. For hoteliers and travel brands watching Meituan’s trajectory, the message is clear: the future belongs to ecosystems where open-weight intelligence bridges the gap between digital intent and the physical realities of travel and hospitality.
