Cross Border AI Search, Meituan Tabbit Beta Test, Or a Two Birds With One Stone Strategy?

The AI search sector—a battleground fiercely contested by major tech companies—has gained a new entrant.

On March 2, 2026, Meituan’s Guangnian Zhewai team announced that its AI search product, Tabbit, has entered public beta testing. Tabbit integrates web browsing, searching, AI-powered conversations, and complex task execution into a single platform, aiming to transform browsers from passive information-display tools into proactive intelligent work partners.

Meituan stated that Tabbit is an AI-native browser aimed at a broader user base including office workers, students, and creators.

After companies such as Perplexity, OpenAI, and Microsoft launched AI search products, this sector has become a new focal point for competition among tech giants. Compared to other AI search products, Tabbit's differentiator is not in "searching," but in "what you can do after the search."

Meituan, as an e-commerce platform, has launched an AI-powered search tool, with the true strategic intent likely being to create a new entry point for its local services segment.

Zhuang Shuai, a retail e-commerce industry expert and founder of Bailian Consulting, analyzed to Times Weekly reporters that if AI search becomes a new super entry point for local life, its commercial value will not be limited to advertising and commissions, but will become the starting point of competition for a new generation of retail e-commerce infrastructure, and will reconstruct the entire industry chain.

Users can shop online with Tabbit

Meituan's public test this time is an AI "secretary" aimed at the general public.

Tabbit was developed by Guangnian Zhizhi, a team under Meituan. Guangnian Zhizhi was founded in 2018 by Wang Huiwen, co-founder of Meituan. In June 2023, Meituan acquired Guangnian Zhizhi for approximately RMB 2.065 billion. With Meituan’s support, Guangnian Zhizhi focuses on AI technology research, development, and application.

Currently, Tabbit has integrated Meituan's self-developed large language model, LongCat-Flash-Chat, and also connected to mainstream domestic large models such as DeepSeek, Doubao, Qwen, and Kimi. In addition to offering the information gathering and question-answering capabilities found in most AI search products on the market, Tabbit can also perform complex web operation tasks, simulating manual workflows to automatically open web pages, extract information, fill out forms, integrate data across platforms, and ultimately deliver complete results to users.

From a UI design perspective, Tabbit adopts a vertical tab bar layout, which not only accommodates more tabs but also features intelligent organization—automatically categorizing tabs by topic—and supports one-click expansion or collapse.

Beyond organizing information and creating tables, Tabbit's consumer-facing applications have matured. For example, when a user types into the chatbox, "Open Taobao, buy cat food, and check for discounts," Tabbit can navigate to the Taobao webpage, automatically perform product searches, analyze discounts, compare prices, and add the desired item to the shopping cart, ready for checkout. Although branded as an AI search tool, Tabbit already possesses the functional attributes of an AI agent.

Cui Lili, deputy director of the Institute of Digital Economy at Shanghai University of Finance and Economics, told Times Weekly that for Meituan, the popularization of Tabbit needs to leverage its strengths and avoid weaknesses. Meituan can take advantage of the traffic benefits of its existing APP matrix and set up a Tabbit entry in a prominent position within the APP; at the same time, it should emphasize the authenticity and reliability of data, thereby changing the public's established impression of Meituan merchants asking for positive reviews. In addition, Meituan also needs to strengthen the user's personalized adaptation experience with AI.

Future may become the entry point of instant retail

The launch of Tabbit not only helps Meituan expand into the internet technology sector but may also serve a dual purpose: this product could potentially be integrated into Meituan's local services business segment in the future.

From a strategic perspective, Wang Xing stated in the earnings call that Meituan is deploying AI across three dimensions: AI infrastructure, AI in the workplace, and AI in products.

The first two correspond to Meituan's self-developed LongCat large model and the no-code programming tool NoCode.

On the product side, Meituan has launched its AI lifestyle assistant "Xiao Mei," which supports natural language interaction and helps users place food delivery orders, get restaurant recommendations, make reservations, navigate, plan itineraries, and access other local lifestyle services.

In addition, on February 12th, the Meituan app launched an AI butler named "Ask Xiaotuan." This product deeply integrates local lifestyle information. Users only need to ask questions related to local food, drinks, entertainment, and leisure activities. "Ask Xiaotuan" can then, through in-depth analysis, collect accurate and up-to-date merchant and service information from the Meituan platform, perform a second verification, and, combining real user reviews, provide users with choices that match their needs and can be directly transacted.

The aforementioned two AI products focus on dynamic itinerary planning for local lifestyle services, while Tabbit is poised to become an intelligent assistant capable of directly completing booking or ordering actions, competing with similar products like Alibaba's Qwen for entry-point traffic in the local lifestyle segment.

As of the time of writing, Meituan has not yet responded to this.

Zhuang Shuai analyzed to the Times Weekly reporter that if Tabbit is integrated with local life services, the most likely initial application scenario would be cross-store automatic ordering and other instant retail scenarios. Compared to dynamic itinerary planning, the decision-making threshold for instant retail goods (meals, fresh produce, daily necessities) is lower, with a higher repurchase rate, and the success rate of AI intervention is also very high. Cross-store ordering is a complex manifestation of such high-frequency, essential scenarios, and the value of AI lies in freeing users from the cumbersome operation of "constantly switching between multiple apps," which better embodies the core value of a "personal life butler" than itinerary planning.

According to a report by the China Ministry of Commerce Research Institute, the scale of China's instant retail market is expected to reach 971.4 billion yuan in 2025, and exceed 1 trillion yuan in 2026.

This means that if Tabbit integrates with the instant retail business, it will face an incredibly vast market.

Cui Lili analyzed for Time Weekly reporter that the role of AI in the local living sector is reflected in two dimensions: on one hand, it can further enhance user loyalty from the user experience level, achieving traffic acquisition and customer retention; on the other hand, it can effectively improve the overall operational efficiency of the industry.

"According to Zhuang Shuai, with Meituan's current technical capabilities, breaking down the technical barriers between Tabbit and the local life sector is not difficult, but cultivating user habits will still take some time."