Accessibility Adjustments

Use these optional tools to adjust reading and display preferences. These tools cannot resolve every accessibility barrier. Please contact the website owner if you need assistance.

  • Text adjustments
  • Content scaling 100%
  • Font size 100%
  • Line height 100%
  • Letter spacing 100%
  • Colour adjustments
  • Orientation adjustments

QwenWork 2.0 adds team agents and event triggers

Listen to this article

QwenWork 2.0 arrived on November 9, 2026, adding shared spaces where teams work with an AI agent, tasks that react to business events, and ways to continue work across devices. The official desktop release notes identify version 2.0.0. This is an office product update, with no new Qwen model announced in those notes.

The useful question for a team is how a request moves from a conversation into a finished, reviewable result. QwenWork now puts more of that process in one place. Its permissions, execution limits and account requirements still determine what can actually run.

A shared space for people and an agent

Each collaboration space has a dedicated digital employee that works from its materials, discussions and confirmed decisions. The release notes describe project follow through and to do management, with the work and outputs retained in that space.

People can also open DingTalk documents, AI tables and Yida applications inside a task conversation. They can make a change themselves before asking the AI to continue, or select a particular passage or cell for revision. Materials Space collects uploaded and generated files and connects outputs to their source conversations.

That makes a specific pilot worthwhile: take one recurring team deliverable, keep its inputs in a shared space, and compare the final result with the agreed requirements. Count missed requirements and review time, rather than assuming that an agent completing a task means the result is correct.

Event triggers come with execution limits

The automatic tasks guide covers DingTalk messages, approvals and to dos. A task can watch up to ten events, then check conditions before running. Available events vary by account. Approval reassignment is not supported.

The free personal tier permits ten enabled automatic tasks in total. Triggered tasks are capped at 200 runs per hour, and excess events do not run during that hour. Each run creates a separate conversation and ordinarily consumes credits. Scheduled tasks support a manual test run; triggered tasks currently do not.

Execution location matters. A local task needs the computer awake and the client running. The guide says cloud execution does not require an open desktop client, but its troubleshooting section also calls out remote control and account authorization checks. Verify the intended setup before relying on an unattended workflow.

Start with a narrow condition and a reviewable output. For example, a message can become a proposed task with a source link before anyone lets the automation send an external reply or change a business record. This is a suggested pilot, not a test ByteForward has performed.

Remote access keeps its own permission boundary

The November 9 web notes say a phone can start or continue work on a designated computer after remote control is enabled, using the same account and organization. Permission requests can be handled remotely. Cloud tasks gain continuation across supported devices.

Enterprise remote control is denied by default when no policy is configured, and an administrator must confirm access. Ordinary files use Materials Space sharing controls. Online documents, tables and applications keep their existing permissions. Eligible remote Office previews are read only and require the relevant access.

A useful rollout checklist should therefore separate who can see a shared artifact from who can operate a computer. For another example of why these controls need separate review, see ByteForward’s Claude Enterprise security controls guide. The products have different features, but the access questions are relevant to both.

Access and pricing to check before adopting it

The getting started guide lists DingTalk, a web login with a DingTalk account, and the desktop client as entry points. The product homepage labels itself the China site. We have not verified a global rollout or a complete plan by plan entitlement list for every new feature.

At the time of review, the pricing page lists a free tier and recurring monthly personal plans at 78 yuan for Standard and 158 yuan for Advanced. Their base allowances are 2,000 and 4,000 credits per month. The free tier advertises promotional registration and monthly credits, alongside 1GB storage.

An unlimited standard mode promotion applies to eligible paid personal subscriptions. Heavy image, audio and video work still consumes credits. Teams should budget their actual task mix and confirm eligibility in their account before treating any advertised unlimited allowance as a general service guarantee.

QwenWork 2.0 provides a concrete new way to connect teamwork, events and deliverables. Its value will depend on whether a team can keep the correct context, permissions and review process attached to that work.

Jordan Reid
Jordan Reid

Jordan Reid is focused on AI tools, agents, developer products, and the way technology changes everyday work. Jordan approaches a launch from the user’s side of the screen. What can it actually help someone finish? The voice is practical, conversational, and skeptical of products that turn a simple job into five new settings. Coverage follows coding assistants, creative software, browser agents, and the workflows around them, with attention to pricing, permissions, setup, and the human work that remains.

Leave a Reply

Your email address will not be published. Required fields are marked *

Gravatar profile