Loading
PLATFORM
CAPABILITIES
ASANA AI
PLANS
Company Type
Teams
Industries
Use Cases
LEARN
SUPPORT
SERVICES
Help Center
search-icon

Topics

Understanding access control for AI Teammates

In This Article

Maintaining security and privacy is just as important as the work itself, and AI Teammates are designed to follow the exact same rules as your colleagues. They operate within the boundaries of your existing permissions, ensuring they only see, search, or summarize the tasks and projects they’ve been explicitly invited to. By mirroring Asana’s standard access levels and using memories that respect your team's privacy, these AI Teammates help scale your productivity without ever overstepping your organization's data guardrails.

Related articles

Access control for AI Teammates

When first created, AI Teammates can see public tasks and projects in the domain – just like new users. AI Teammates must be explicitly added as members or collaborators on private work graph objects (e.g., tasks, projects, teams) to see and search them. 

AI Teammates respect Asana's standard access levels:

Access level

What the Teammate can do

Editor

Create, edit, and complete tasks; post comments; update custom fields; manage subtasks

Commenter

Post comments on tasks; cannot create or edit tasks

Viewer

Read task information only; cannot comment or edit

 

Note iconNote

If an AI Teammate is added to a team, it will inherit the team’s access levels.

An AI Teammate cannot elevate its own permissions.

When you assign a task or ask a question to an AI Teammate, the AI Teammate will only respond using information from tasks, projects, and other objects that both you and the AI Teammate can access. AI Teammates can’t reveal information that you, as the “triggering” user, don’t already have permission to view.

For example, if you ask your AI Teammate to answer questions about a particular Asana project, they can only access and respond to you about that project if both of you have access to that project and/or its tasks. In other words, both you and your AI Teammate need to be authorized project members or task collaborators. 

Note iconNote

Keep in mind that AI Teammate comments are visible to everyone with access to that task. Before triggering a request, consider whether the audience in that context is appropriate for the content you're asking for.

Details about access control

  • AI Teammates see public Asana data by default, like any new user in your domain (e.g., tasks in public projects; projects in public teams).
  • Requests involving or accessing information from third-party tools, such as Google Drive, use the triggering user’s access tokens.
  • When creating or editing an AI Teammate, authorized members may upload key resources (e.g., playbooks, policies, instructions) for the AI Teammate's ongoing reference and access.

In practice, this means an AI Teammate cannot be used to look up information you do not already have access to. It can only work with the same tasks, projects, teams, and third-party files that you and the AI Teammate can both see.

If you want an AI Teammate to help on a private project or team, make sure it has been added as a member or collaborator, just as you would add a new colleague.

Memories

To stay helpful over time, AI Teammates can remember important details about the work they have done. This section explains how those memories work and how they respect the same access controls as your tasks and projects.

After an interaction, an AI Teammate may store a memory for future reference. Memories are most commonly created during task executions, but can also be created when an AI Teammate reads key resources like documents. Think of memories as private notes that can only be viewed by users who can both see the associated object and trigger the AI Teammate.

Details about memories

  • Memories are stored on tasks as point-in-time recollections or summaries of interactions related to the task.
  • A user who loses access to a task will also lose access to its memories.
  • When an AI Teammate is removed from a project or task, the AI Teammate will forget the related memories.
  • Memories can be viewed by authorized members of the AI Teammate on the AI Teammate’s profile page.
  • A memory is only visible to users who already have permission to view the source task. For example, if a Teammate creates a memory from a private task, a colleague who cannot see that task will not be able to see the memory, even if they have full access to the Teammate.

Put simply, memories help an AI Teammate stay consistent with your preferences and past decisions, but they never widen access. Only people who can already see the underlying task and work with the AI Teammate can see its memories.

Over time, this lets teams build shared context with an AI Teammate so it can pick up work where it left off and avoid repeating the same questions, while still following your existing permissions.

Integrations

In addition to reading your Asana workspace, AI Teammates can connect to external tools to extend their capabilities.

Note iconNote

Authorizing an integration for one Teammate grants access to that integration for all of your Teammates.

Google Drive

AI Teammates can read, search, edit, and create content in Google Drive, including:

  • Google Docs: Read existing documents; create new documents; edit documents with content changes
  • Google Sheets: Read spreadsheet data; create new spreadsheets; edit cell values
  • Google Slides: Read slide content

The AI Teammate's actions are performed under the permissions of the user who initiates the trigger, and these actions will appear as though the user performed them. Consequently, the AI Teammate's access is restricted by the user's permissions; for example, it cannot access a Google Doc in a task if the user lacks access to it. It will not share the credentials of other collaborators on the task.

Microsoft Office 365 (beta)

  • Read, edit, and create new documents for SharePoint/OneDrive docx, pptx, and xlsx files.

To use SharePoint/OneDrive with an AI Teammate, the Teammate uses the credentials of collaborating users on the task. The collaborating user must have access to the relevant files.

Web access

AI Teammates can read public web pages when given a URL. This is useful for:

  • Researching topics or competitors
  • Reading public documentation
  • Pulling information from public data sources

Related articles

Loading
AI Teammate Access Control: Permissions, Privacy, and Memories | Asana Help Center