Qianwen connects Alibaba's entire ecosystem, seizing the global AI super gateway.
```
Author | Zhou Zhiyu
Editor | Zhang Xiaoling
The most genuine transformation in the commercialization of AI has reached a critical point.
On January 15, Alibaba’s AI assistant Qwen was no longer content to act as a mere chat companion. After announcing that its monthly active C-end users surpassed 100 million, Alibaba revealed a long-prepared trump card: teaching AI to “get things done.”
At an event that day, Alibaba Group Vice President Wu Jia announced that Qwen is ushering in the era of AI task assistance.
In the latest version, Qwen is no longer just a simple dialog box—it connects directly to Ele.me, Fliggy, and Taobao’s foundational systems. With just a single sentence, AI can penetrate the barriers between apps and complete the cycle of ordering food delivery, booking flights, or purchasing goods.
According to people close to Alibaba, these are only part of the new features; more functions will be released at today’s launch event.
Alibaba’s move shows that large models cannot stay only at the level of showing off technical prowess—they must also have the ability to complete commercial cycles.
In the past, AI Agents were often seen by tech giants as distant technical concepts, and sometimes viewed by the market as “PPT buzzwords” useful only for storytelling in the secondary market. But Alibaba’s actions clearly intend to tell the market: AI that helps you get things done has arrived.
01 Farewell to “Empty Hype”
For the past two years, the large model track has been bustling. From the hundred-model wars to price reduction sprees, all players faced an embarrassing “empty hype”: no matter how high the parameters climbed, if no transactions occurred, it was all meaningless.
Although the 2025 large model hype excited the industry chain for a whole year, by the beginning of 2026, primary market investors started to sober up. They realized that after burning billions of computation power, all they got were user requests like “write me an acrostic poem” or “generate a picture.”
This type of interaction cannot support trillion-level valuations.
The K-line chart of the capital market is the most honest signal. Large model manufacturers urgently need to find a scenario that can convert huge traffic into real GMV (gross merchandise value). This sense of urgency instantly awakened the “muscle memory” of internet giants: the iron rule from the PC era to the mobile era—whoever controls the transaction entrance is king.
Alibaba is well aware of the pain of unmonetizable traffic. So they chose to unveil a killer move this year.
Unlike most models on the market that can only “give suggestions” or “look up tips,” Alibaba’s logic is to link up its app ecosystem internally. Wallstreetcn found in their actual test that Qwen can already call, in conversation, instant ordering from Taobao, ticket booking via Fliggy, and price comparison of Taobao products.
It’s as if the large model has been given hands and feet. In a single dialog box, Qwen’s large model is responsible for deep understanding, while Alibaba’s vast service ecosystem performs execution.
This “front shop, back factory” model lets Alibaba bypass complex app jumps and directly enter the core transaction flow. When a user says, “I want to book a ticket to Beijing for tomorrow,” he no longer needs to open Fliggy, input dates, and filter flights. AI directly pushes the final payment card in front of him.
02 The Battle of Strategies
Of course, in the fight for this trillion-level mega-entry point, Alibaba has no shortage of rivals.
Zhang Yiming’s ByteDance is the “challenger” that keeps everyone up at night.
Doubao has already surpassed everyone in daily active users, becoming the de facto “national-level AI.” But when it comes to how to make AI “get things done,” Doubao and Alibaba have taken two completely different directions.
Because Alibaba owns Ele.me, Fliggy, and Gaode—its “own progeny”—it controls the service side. It doesn’t have to depend on others, and can directly connect internal data interfaces. The advantages of this approach are stability, high order success rate, real-time data sync, and almost no delay.
In contrast, Doubao takes a more radical, more geeky “Auto-UI (Automated User Interface)” approach.
Through the Doubao mobile assistant, ByteDance is trying to give AI “eyes.” It uses vision models to identify every pixel and button on a phone screen, then simulates human fingers to tap and swipe.
In Doubao’s logic, AI doesn’t need app developers to open interfaces; it takes control of your phone directly at the system layer. You say “call a cab,” it opens Didi; you say “order food,” it opens Meituan.
But the challenge for Doubao Mobile Assistant is that it needs to overcome the smartphone system’s huge barrier to cross-app operations. It's betting that AI will become a "super operating system" above all apps.
Whoever cultivates user habits first will get the entry ticket for the next decade.
03 Global Consensus
This leap from “chitchat” to “getting things done” is not unique to China. Across the ocean, a similar commercial alliance is emerging.
The partnership between Google Gemini and retail giant Walmart is basically a global mirror of Alibaba’s model.
Google has the most powerful brain, but still fears Amazon’s e-commerce dominance. So, it teamed up with Walmart, which has the strongest offline network. Under their partnership framework, based on the Universal Commercial Protocol (UCP), Google Gemini can directly access Walmart’s real-time inventory data.
When a US user asks “what should I buy for a barbecue party,” AI can not only provide a menu, but also directly identify items on Walmart’s shelves, use Google Pay to pay, and even arrange curbside pickup.
The logic behind this is the same as Alibaba’s: using the certainty of service to offset the uncertainty of technological iteration.
This also explains why OpenAI is hurriedly preparing a browser agent codenamed “Operator.” Silicon Valley elites have discovered that pure SaaS subscription models are hitting a ceiling—only by entering the real-world transaction flow can AI tell a bigger story than the internet itself.
04 The Twilight of the App Era
As C-end competition in large models enters deep waters, the market is returning to reason. Alibaba chooses this moment to feed the “family bucket” (all-in-one apps) to AI, hoping to be the first to establish a commercial model amid today’s traffic anxiety.
But to truly bring smart assistants into daily life, it will still take time.
UBS China Internet analyst Xiong Wei said to Wallstreetcn on January 14 that it will still take time for AI agents to be rolled out at scale, popularized, and monetized. Not only does technology need to ensure accuracy and stability, users must also accept and adapt, the industry chain must connect upstream and downstream, cooperative relationships must be built, economic interests redistributed, and regulatory concerns considered.
In Xiong Wei’s view, having an AI agent inside a large company’s ecosystem that integrates internal resources is an important evolutionary stage. The next stage is an agent that connects operations across all platforms—providing more comprehensive user assistance, decision support, and even collaboration between agents, which is something for a future phase.
Letting AI get things done is really about stabilizing the boom-bust cycle of large models “burning money but not making money”—seeking long-term stability.
The universality of service helps resolve the homogeneity of technology, and the reality of transactions helps fill the emptiness of mere dialogue. This is what genuine AI application should look like.
Risk Warning and DisclaimerThe market has risks, investment requires caution. This article does not constitute personal investment advice, nor does it take into account individual users’ special investment goals, financial circumstances or needs. Users should consider whether any opinions, viewpoints or conclusions in this article suit their particular situation. Investing accordingly is at your own risk.
```