Qianwen Office 2.0 adds "Collaboration" so teams and Agents work in one shared space

Qianwen Office (千问办公) released version 2.0 on November 9, with a new "Collaboration" (协作) feature that lets multiple people and Agents work together in a shared space. The company calls it the industry's first feature of its kind. The release is the largest upgrade since the product launched three months ago, and all features are now live for all users in the Qianwen Office client.

The details come from a Chinese-language report by QbitAI (量子位), which republished material provided by Qianwen, and from the company's blog post "Qianwen Office 2.0: Proactive and Collaborative, Toward the Team Agent Era."

What shipped

Collaboration. Users can set up a workspace where a team and "digital employees" work together. Inside it, the digital employee can link in real time to multiple DingTalk group chats, documents, AI tables and AI listening notes (AI听记). It supports parallel sessions plus task management and assignment. When a member assigns a task, the digital employee uses the workspace context to review progress, summarize consensus, and create and follow up on to-dos.

Auto Tasks (自动任务). Agents move from responding on request to starting work on their own. Users subscribe to events and configure rules, and the Agent begins processing when conditions are met. The first supported triggers are messages from a specified person, messages in a specified group that @-mention the user, and approval tasks. In one example, a regional manager sets the rule "if someone in the group @s me about a stockout, read the inventory table and start a transfer request." The Agent then reads inventory and fills the transfer details into a table.

Human-AI Co-creation (人机共创). Conversation, content preview and artifact editing are combined. Users can view a document preview while giving revision requests in the chat box, or select paragraphs, cells, slides or app regions, mark multiple comments and submit them at once. The AI then edits the original document directly. For online documents, AI tables and Yida (宜搭) apps, multiple people and multiple Agents can co-create simultaneously.

Materials Space (资料空间). This collects AI-generated and user-uploaded images, documents, tables, PPTs, web pages, apps and local files, with viewing, search and sharing. Selecting one or more files can start a new conversation or continue editing through co-creation. The company's example is selecting several industry research files to generate a market plan.

Three-end sync (三端互联). Tasks, conversations and artifacts sync in real time across phone, computer and web. Tasks run in the cloud and continue after the computer is shut down. With user authorization, a phone can remotely operate the computer.

Customer example

According to the company, the card-and-coupon operations team at Intime Department Store (银泰百货) used Collaboration to improve efficiency. Previously, its 5 operators collected coupon submissions from more than 60 store group chats one by one, checked them, and entered them into a spreadsheet. The team now links those store groups and one AI table to a single workspace. Communication habits are unchanged, and the Agent aggregates information across groups and updates the table automatically.

Context

Qianwen Office had previously connected to IM tools including DingTalk, Feishu and WeCom. The company says it is now evolving into an enterprise Team Agent as atomic capabilities from IM tools, native Office and SaaS software are integrated into the Agent client. Its blog argues that mainstream office Agents still mainly serve individuals and that the next generation must enter the places where teams work. The "industry first" claims are the company's own.

What this means

The 2.0 release shifts Qianwen Office from a single-user assistant toward a shared team tool. Collaboration and Auto Tasks together move the Agent into group chats and event-driven workflows. The Intime example shows the intended use case: pulling scattered group-chat information into structured tables without changing how staff communicate. Cloud execution and phone-to-computer control extend that work beyond an open desktop session. Whether these features perform as described at scale has not been independently tested, and the efficiency results come from the company's own account.