Caocao Mobility integrates Doubao Taxi for gray testing; AI assistant is exploring more scenarios.

```

News on June 22: ByteDance’s Doubao App has officially launched a grey-scale test of its instant travel service.

Some qualified users can directly state their needs in the Doubao chat window using natural language. After the system automatically recognizes the departure and destination locations and user preferences, it matches Caocao Mobility’s transportation capacity via API interface. Users can confirm the information and dispatch an order with one click.

This move is not only an upgrade in interactive experience, but also reveals the core ambition of current AI To C application evolution: general AI assistants are trying to break the boundaries of pure content interaction by integrating real-life services as plugins into their platforms, evolving towards being a super gateway that connects everything.

Around the intersecting area of “AI + Mobility”, the industry has now formed several forces with different logics, presenting a two-way penetration between universal large model platforms and native mobility giants.

For integrating AI platforms with ecological transportation resources, a representative example is Qianwen and AutoNavi, which is the Alibaba system’s loop that was completed in March this year.

Tongyi Qianwen App launched its AI ride-hailing function, relying on the large model’s natural language understanding on the front end, and directly calling AutoNavi Map’s massive aggregated ride-hailing pool on the back end. This represents giants with full ecosystem business matrices using AI gateways to deeply couple their internal traffic bases with mature underlying services.

Another representative is the native platform’s endogenous defense, like Didi’s launch of Xiaodi.

Facing the trend of AI assistants attempting to intercept ride-hailing needs from the front end, mobility giants have not passively waited. Didi also opened its built-in travel Agent Xiaodi to all users in March this year. Didi’s logic is to defend its main territory, leveraging its closed ecosystem and massive transportation resources, using large models as underlying tools to optimize its order planning and dispatch efficiency, thus strengthening the moat of its independent travel gateway.

This constitutes a clear industry game: universal AI giants want to turn mobility into a functional component on their platforms; while travel giants want AI to become a technological engine to optimize their own service systems.

Returning to Doubao's strategy this time, it is mainly about competing for the gateway. Before integrating the ride-hailing service, Doubao has already densely tested business in local life and e-commerce sectors.

In March this year, Doubao started grey-scale testing of its AI e-commerce feature, focusing on one-sentence shopping. When users input “Budget around 100 yuan, recommend a good office mouse” in the chat window, Doubao no longer just gives text suggestions, but directly matches products, generates selection advice, and attaches product links. Users don’t need to switch to the Douyin App, they can complete ordering and payment directly within Doubao. Soon after, Doubao further extended its reach to local group-buying and offline redemption scenarios.

From e-commerce shopping, offline group-buying, to now introducing Caocao Mobility’s transportation, Doubao’s business logic is very clear: use a universal large model as a central processor, make the chat window a unified user interface, and gradually embed high-frequency application scenarios like food, clothing, housing, and travel as functional plugins into its platform.

In the past two years, the main narrative for large model C-end applications has been information retrieval and text generation, but the pure tool attribute is difficult for long-term user retention. The continued integration of real-world delivery services aims to cultivate user habits of substituting traditional graphical interfaces with natural language interaction, thereby diverting traffic from various niche apps.

For the general AI assistant to connect everything in the chat window, it does deliver a dimensionality-reducing upgrade in experience at the information matching layer. But when it comes to landing in local life verticals, business logic still needs to return to the offline common sense of heavy assets.

No matter how smart Doubao or Qianwen’s intentions are, final delivery of ride-hailing services still heavily depends on the density of offline transportation resources. During rush hours or supply-demand imbalance in extreme weather, if there are no available cars in real life, even the most powerful AI gateway cannot fulfill the order. However light the front-end gateway is, it can’t cover up the heaviness of the back-end transportation capacity.

Doubao is currently only conducting grey-scale testing, with the core purpose of calibrating the spatial geographic mapping accuracy of its large AI model through real business order trials.

Doubao's grey-test ride-hailing service is an important signal of the AI assistant’s leap from a single tool to a platform-like super gateway. The strategic intent of major AI platforms to take over the offline world within the chat window is now evident.

However, in traditional tracks like mobility with long chains and heavy operations, when interaction efficiency is leveled by technology, the moats that decide final victory remain the accuracy of underlying location services and the ability to dispatch offline transportation resources. The battle for “who is whose plugin” in gateway competition has only just begun.

Risk Warning and DisclaimerThe market carries risks, and investment should be done cautiously. This article does not constitute personal investment advice and does not consider the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. Investment based on this is at your own risk. ```