Threaded Conversations in Wrike

TL;DR

Threaded conversations keep related replies together in Wrike’s item stream. You can reply directly to comments, track thread responses in Inbox, and use AI agents to manage conversations through the application programming interface (API) and Model Context Protocol (MCP). This experimental Wrike Labs feature may change or have limited availability.

• Overview
• How to Reply in a Thread
• Work With AI Agents
• What’s Next?

Overview

Threaded conversations let you reply directly to comments in the Wrike stream. They keep related messages together, show which responses need your input, and help you identify thread replies in Inbox.

How to Reply in a Thread

1. Open the relevant item and find the comment in its stream.
2. Click Reply 1.
3. Enter your response, and click Send 2.
4. Your response appears in a thread under the original comment.

Current limitations: Wrike automatically @mentions relevant thread participants. To notify someone new, add an @mention.

Inbox notifications show that the message is a reply rather than a new comment.

Work With AI Agents

AI agents can post and receive replies in threads as part of a human-in-the-loop workflow. Persistent conversation states and request IDs let agents retrieve responses later.

The application programming interface (API) and Model Context Protocol (MCP) support actions to create, read, update, respond to, and resolve comments and conversations.

Key Capabilities

• Thread Participation: AI agents can automatically post updates and receive user replies directly within existing comment threads.
• Grouped conversations: Agents can match responses to the right topic more accurately, so you can use multiple agents at once without context conflicts.
• Full Lifecycle Management: Through our Application Programming Interface (API) and Model Context Protocol (MCP), agents have the programmatic access required to:
  • Create new comments and conversation threads.
  • Read existing discussion history and context.
  • Update agent-generated messages as new data arrives.
  • Respond to specific user queries or feedback.
  • Resolve completed comments and close out resolved threads.

How It Works

1. Triggering an Agent: A user or system event initiates an action, prompting the AI agent to post in a thread with a unique request ID.

2. Human Interaction: Team members review the agent's input and reply directly within the thread to offer feedback, approvals, or additional context.
3. Ongoing conversations: Agents respond much like people in a human-to-human conversation. Conversations aren’t automatically marked as resolved.

What’s Next?

• Editing Comments
• Available Formatting in Comments