Alibaba arms Qwen AI to order food, pay and book travel inside chat

Alibaba has upgraded its consumer-facing Qwen artificial intelligence app so the system can perform real-world tasks end to end inside a single chat interface. The company embedded core services from its ecosystem — Taobao and Taobao Instant Commerce, payment platform Alipay, travel service Fliggy and mapping service Amap — to let users move from intent to completion without switching apps.

The company framed the upgrade as a shift away from systems that merely understand user requests toward systems that can execute them. Wu Jia, vice president of Alibaba Group, said, "AI is evolving from intelligence to agency." In public test builds now available in China, Qwen can complete full ordering flows, make in-chat payments, plan and book travel, call restaurants and manage multi-step tasks within the chat UI.

In a live demonstration of the new features, Qwen ordered 40 cups of bubble tea from a local store using Taobao Instant Commerce, applied discounts automatically, completed payment inside the chat and scheduled delivery, all without leaving the conversation. The upgrade also includes a task assistant feature that is available to select users in the initial test, allowing the AI to chain actions across commerce, logistics and urban services.

Alibaba presents the move as a strategic pivot toward consumer-facing AI. The company has invested heavily in enterprise AI through its cloud business in recent years, and this upgrade signals a renewed push to capture consumer engagement in areas where rivals have been active. The expansion into agentic capabilities follows a broader industry trend in which technology firms are linking large language models to real-world services so the systems can act on users' behalf.

Market reaction to the announcement was mixed. Alibaba’s Hong Kong-listed shares, cited under tickers 9988 and BABA, showed a decline in early market snapshots, with one data point listing 9988 down about 2.96 percent at the time of publication. Analysts see the upgrade as both a defensive and offensive move: it brings Alibaba’s consumer apps and payments closer to an integrated AI interface while testing how reliably an agentic system can operate in commerce and travel workflows.

Significant questions remain. Alibaba has not disclosed a global rollout timetable, specific user counts for the public test, technical model details such as model size or training data, or the guardrails that will govern automated actions. There is also no public account of regulatory approvals for expanded payment or booking functions mediated by AI.

The capabilities announced raise familiar policy and safety concerns. In-chat payments and autonomous ordering amplify the need for robust fraud prevention, consent controls and transparent error recovery. They also expand the surface for privacy and data protection issues as a single chat can connect purchase history, payment credentials and location services.

For consumers, the promise is convenience: fewer app switches and faster task completion. For regulators and product managers, the challenge will be ensuring reliability, safety and clear accountability when an AI moves beyond suggestions and into actions that have financial and logistical consequences.