- You can now bring your AI agents into Wrike, with a real identity. Assign work to it or @mention it as you would a colleague.
- Every Workmate action is permission-scoped, logged, and traceable, and nothing moves without a human confirmation.
- Internal Workmates are coming later this year, starting with Project Manager, Campaign Performance and ROI, SoW and Scoping, and Portfolio Change and History.
At Collaborate 2026, I had the pleasure of announcing our latest AI development: Workmates.
Workmates are AI agents with real identities in Wrike that you can now assign to execute work and interact with other users, just as you would with any other colleague.
Perhaps you’ve already set up AI agents in Wrike. These agents are huge time savers, helping to automate within Wrike — from intake to triage to risk management reporting. Workmates are different. They can work across platforms and spaces. You can connect your Workmate to the other places where your work lives: your document store (such as Google Drive), your design tool (such as Figma), and your CRM, as long as the platform allows it.
A Workmate connected to Google Drive can read the brief sitting there and build the project in Wrike from it, in one job. This is what makes it a Workmate rather than an agent that only knows Wrike.
Think of it as a partnership: the Workmate performs tasks and handles the heavy lifting, always under human guidance and judgment. No work begins without your confirmation, and remains fully customizable to your team’s structure and governance.
Now, let’s take a deeper look at how Wrike Workmates actually work.
What makes Workmates different?
Picture Carlos, a project manager who needs to allocate his team’s resources for an upcoming high-stakes deployment. Every team member has their individual skills, expertise, and capacity schedule in Wrike. Instead of doing the manual admin at the first stages, Carlos @mentions his Project Manager Workmate with a simple ask: propose a way to optimize resource management across the team.
This is when the magic happens: the Workmate reasons across what Wrike already knows about the team, who has done similar work, who is overloaded, how past projects were staffed, and then builds a whiteboard with two visual options for the manager to review.

Note that the Workmate does not decide on its own. It waits for Carlos’ input and judgment to move on to the next steps, as the human always remains the final approver.
To make the best decision, Carlos wants to set up a kickoff meeting with his team. Again, he can simply assign his Workmate to find a time that works, invite the team, attach the whiteboard, and frame the whiteboard with the two options as the agenda. Once the team has discussed and decided on an option, the Workmate will record the outcome on the Wrike task where the decision was made.
That’s what makes Workmates different. They don’t just transfer data and context between systems, which Wrike’s integrations could already cover. They operate with a goal in mind, reason across the full context of work connected via Wrike, and take actions according to their defined role. Workmates hold their own identities inside your workspace and receive work through the same channels and permissions as your team members. Assign a task to a Workmate or @mention it, and it instantly starts the work.

How to set up a Workmate-human collaboration
What I truly love about Workmates is that they give the AI agents you’ve already built a real seat at work. They develop their own identities, receive work as colleagues would, and leave a traceable record of what they did. Remember, this is a partnership. You set the boundaries, you approve the judgment calls, and the agent carries the rest.
Here are a few tips for effectively incorporating your Workmates into your team’s daily routine:
- Know your Workmate’s role: Treat it as you would a new hire with a clear set of responsibilities: project management, following up on marketing campaigns, retrieving project history, and so on. A defined role tells your team when to turn to it and tells the Workmate when to let a human coworker step in.
- Only connect it to the data it needs: Scope your Workmate’s access to internal systems and Wrike data to the role you assigned it. Since each Workmate has its own identity in Wrike, you can configure its access to spaces and projects in the same way you do today for users.
- Align Workmate’s operating instructions with your configured workflows: For example, outline when a Workmate can update a status on its own, and when it needs to wait for human approval before making specific changes to projects and tasks it works on or oversees in Wrike.
Note that you can even connect your custom AI as a Wrike Workmate to allow your AI assistant to access your project data through Wrike MCP.
Internal Workmates: What’s coming
I am excited to announce that Wrike is building a fleet of internal AI Workmates, which will be available later this year. Each one is designed to accelerate your specific workflows and help your team work smarter.
Let’s take a look at a few of them:
Project Manager Workmate
We’ve already met this Workmate in the real-life example above. Its role is to suggest project and task owners based on their skills and capacity, and to assess what threatens project delivery by showing you the evidence behind each risk. It can also turn meetings and requests into project updates in Wrike, while checking with you before making any changes along the way.
Campaign Performance and ROI Workmate
Ideal for marketing teams, this Workmate checks every campaign’s spend and lead count against the budget, the target it was meant to hit, and its current status. It refreshes the figures, flags when a campaign starts to trail its goal, and suggests a new status for you to confirm. Moreover, it logs each campaign’s performance over time, so the trend is always there to look back on.
SoW and Scoping Workmate
This Workmate takes a deal the moment it’s closed-won and assembles a review-ready draft: a statement of work (SoW), a work breakdown structure, a milestone plan, and the risks worth watching. It pulls from the signed contract, your existing blueprints, and comparable projects you’ve delivered before, and it waits for a human to approve before starting the work.
Portfolio Change and History Workmate
While Wrike tells you the state of your portfolio right now, this Workmate keeps a record of what it looked like before. Over a defined period, it saves portfolio values to a history workbook, detects when dates, statuses, or calculated fields change, and surfaces the three changes most worth your attention.
Bring Workmates into your workflows
One thing many of our customers agree on across their projects is that their teams do a lot of preliminary work before making their real contribution. This includes creating project structures and assigning people, building insight reports, organizing meetings, and so on. Workmates can handle that preparation and provide your team with a clear starting point, allowing them to focus on what truly requires human judgment, such as weighing trade-offs, understanding clients, and setting project direction.
What would be the first role you’d assign a Workmate? Book a demo to see how Workmates can help your team. If you’d like to chat more about the future of agentic AI, feel free to get in touch with me on LinkedIn.
FAQ
What are Workmates, and where can I learn how to use them?
Workmates are AI agents capable of performing complex tasks. Today, you create Workmates by bringing your own existing agents into Wrike. Soon, you will be able to build Workmates in Wrike directly or use one of the Workmates built by Wrike.
Visit the Workmates Library & Inspiration community forum for examples, setup guides, and discussions. Workmates are available in Wrike Labs for Business plans and above during testing.
What is the difference between Wrike Workmates and AI agents?
Agents automate work in Wrike, whereas Workmates perform various complex tasks across Wrike and other connected systems. You can assign a Workmate work or @mention it in Wrike like a colleague. It understands the full context of the underlying work and performs various actions needed to complete the task. It comes back to clarify what is unclear and to ask for your approval where required.
Can a Workmate act without my approval?
It depends on how a Workmate was instructed. You can instruct a Workmate so that no changes are applied without human confirmation. When a step calls for judgment, the Workmate asks for your team’s input or decision. You can also set it to send an approval request in Wrike.
Which access permissions are granted to Workmates?
Only the ones you define for it. You can give a Workmate the same access level as your team, or narrow it to a specific perimeter so it only reaches the work items, projects, and connected systems that its role requires. A Workmate never operates beyond the permissions you set.
How are Workmate actions kept accountable?
Since each Workmate has its own identity, every action is attributed to the agent that performed it and logged in the audit trail. You can see afterward what the agent did and what a person did.