Bring your external agent into monday.com

 

Attention: Bring Your Own Agent is currently in Beta.

 

Who can use this feature
gray-cloud.svgUsers whose Admins have enabled AI capabilities, and the External agents permission in AI permissions

Bring Your Own Agent allows you to connect an agent built on an external platform into monday.com, where it becomes a full entity in the system. Once connected, the external agent can be chatted with, mentioned on items, assigned tasks, triggered, and monitored, just like an agent built inside the platform.


The agent's brain, including its instructions, skills, and tools, stays on the external platform and is managed there. monday.com acts as the governance layer, collaboration surface, and system of record for everything the agent does.

Use this article to learn how to connect an external agent, understand what it can do inside monday.com, and see how it compares to monday agents and MCP.

 

How to connect an external agent

To connect an external agent to monday.com:

1Open the AI Agents tab in the left navigation pane and click Bring Your Own Agent.

2Select your provider from the list.

3Provide the necessary details and connect your externally-built agent.

 

Once connected, the agent appears in monday.com as its own entity with a dedicated identity in the system. When the agent joins your workspace, it automatically receives workspace context, including board structure and team information, so no manual setup is needed to get it up to speed.

 

Supported providers

You can connect agents built on the following platforms. Additional providers are being added continuously.

See all supported providers
  • Anthropic (Claude Managed Agents)
  • Microsoft Copilot Studio
  • OpenAI
  • ElevenLabs
  • Cursor
  • Custom agents (your own stack)

Some providers require additional setup steps before the connection can be completed. The setup flow varies by provider. See below for provider-specific guidance.

Microsoft Copilot Studio

1Create a new agent on Copilot Studio

  • Log in to your Microsoft Copilot Studio account.
  • Navigate to the agent creation section.
  • Follow the prompts to create a new agent tailored to your needs.

2Publish the agent

  • After creating your agent, publish it. Publishing is required before the agent can be accessed externally.

3Retrieve the Direct Line secret

  • Go to your agent's Settings.
  • Select Security > Web channel security.
  • Copy one of the Direct Line secrets listed.

4Create a new Copilot Studio connection in monday.com

  • Log in to your monday.com account and navigate to the connections section.
  • Create a new Copilot Studio connection.
  • Use the Direct Line secret from step 3 as the API key.
  • After connecting, copy the monday token you receive. You will need it in the next steps.

5Add a new MCP tool in Copilot Studio

  • Return to Copilot Studio and open your agent.
  • Go to Tools > Add a tool > Add new MCP.
  • Select API Key authentication.
  • Set the header key to Authorization.
  • Paste the monday token from step 4.
  • Click Create to finalize the tool.

6Configure the server URL

  • Use the following server URL for the MCP connection: https://mcp.monday.com/mcp

7Create a new connection using the monday token

  • Using the token obtained from monday.com, create a new connection in Copilot Studio.

8Set credentials to Maker-provider credentials

  • In Copilot Studio, go to Tools and select the MCP connection you created.
  • Click Details > Additional details > Credentials to use.
  • Choose Maker-provider credentials.

9Disable authentication

  • Go to Settings > Security > Authentication.
  • Set authentication to No authentication.

10Publish the agent again

  • After completing all configurations, publish your agent again to apply the changes.

 

Customizing your external agent

Once connected, your external agent appears in monday.com with its own identity. You can customize its name and description directly in monday.com. These are display settings only and do not affect anything on the external platform. The agent's profile picture will feature the provider's logo. 

 

The agent builder includes the same tabs available for agents built inside monday.com: Brain, Triggers, Channels, and Activity. For a full breakdown of how these work, see the AI Agents on monday.com article.

For external agents, the Brain tab contains only two sections: monday Instructions and Knowledge and access. The agent's core instructions, skills, and tools stay on the external platform and are not surfaced here.

monday Instructions are monday-specific behavioral guidelines that define how your external agent interacts within monday.com. They can be set manually in the Brain tab, or they develop naturally through your chat with the agent. For example, through conversation, you might instruct your agent to create a new item on a specific board when it receives a new task, and post a summary Update on the relevant item when it completes its work.

Note: monday Instructions do not modify the agent's configuration or system prompt on the external platform.

 

What external agents can do in monday.com

Once connected, an external agent can:

  • Chat with you directly: conversations are mirrored between monday.com and the external platform, so a chat started in monday.com is also visible on the external platform and vice versa.
  • Respond to @mentions on items and act on item assignments: the agent can be mentioned on items and assigned to items to trigger it to take action.
  • Run based on triggers: the agent can be triggered by monday.com events such as status changes or item creation, on a set schedule, and more.
  • Participate in AI Workflows: external agents can be placed into multi-agent flows managed through the AI workflow builder, with monday.com coordinating the triggers and handoffs between agents.
  • Appear in the activity log as its own identity: every action the external agent takes is attributed to the agent itself, not to the user who connected it.
  • Be managed in the Agents directory tab, under the Administration section: external agents appear alongside monday AI agents in the Agent directory tab, with a clear label distinguishing them. Admins can deactivate any external agent at any time.

 

What stays on the external platform

The following are managed on the external platform and are not accessible from within monday.com:

  • Instructions: the agent's core prompt and behavioral logic
  • Skills: capabilities defined and managed on the external platform
  • Tools: integrations and external connections configured outside monday.com

This is intentional. monday.com does not duplicate what the external platform manages. Instead, it provides the governance, observability, and collaboration layer on top of the agent's existing setup.

 

Permissions and admin controls

External agents follow an allowlist permission model. By default, a connected agent has no access to any content in your account, including public boards and monday docs. Access must be explicitly granted for each asset the agent needs. For example, for boards, you can grant access by subscribing the agent as a board member, by adding the board in the Knowledge and access section of the Brain tab, or by granting the permission via chat. Nothing is inherited from any user's permissions.

To enable Bring Your Own Agent, an admin must grant the External agents permission in AI permissions for the relevant roles.

Admins can deactivate any external agent at any time directly in the Agent directory, under the Administration section. All agents, both monday agents and external agents, appear in the same view within the Administration section, with external agents clearly labeled and filterable. Every action an external agent takes is logged under the agent's own name in the activity log, so it is always clear what the agent did versus what a person did.

 

How Bring Your Own Agent compares to monday AI agents and MCP

There are a few ways to bring AI capabilities into your monday.com workflow. The right approach depends on where your agents are built and how you want them to operate.
 

monday AI agents

These agents are built directly within the monday.com UI using the monday AI agent builder. No code is required, and agents are configured, governed, and monitored entirely within the platform. Permissions are admin-controlled and granular, you define exactly which boards the agent can access and what it can do within each. Every action the agent takes is logged in the activity log under the agent's own name.


Bring Your Own Agent

This is for teams that have already built agents on an external platform and want to bring them into monday.com, and continue using them in this context as well. The agent's brain stays on the external platform, while monday.com provides the governance, collaboration, and observability layer on top. External agents follow an allowlist model, they need to be explicitly granted access to monday assets, like boards. Like monday AI agents, every action is logged under the agent's own name, making it clear what the agent did versus what a person did.


MCP

MCP connects an external LLM to monday.com so it can read and act on your data. It is initiated by a person in a chat window and runs with that person's full permissions. Access cannot be scoped down below what the user already has access to. Actions are logged under the user who initiated them, not as a separate entity. It is best suited for technical users performing ad hoc queries.

 

Usage and credits

Credit consumption for an external agent depends on the type of action performed:

Action type What it consumes
Chatting with the external agent Third-party provider credits (for example, Anthropic tokens), not monday AI credits
Standard monday.com actions (adding items, updating boards, and similar) API calls against the account's API quota
AI-specific monday.com actions (for example, generating an AI column) monday AI credits

To learn more, see the AI Credits article.

 

FAQs

Do I need to rebuild my existing agents to use them in monday.com?

No. Bring Your Own Agent lets you connect your existing agents as they are. You continue managing the agent's instructions and settings on the external platform, and monday.com provides the governance and collaboration layer on top.

Can external agents access all my boards automatically?

No. External agents use an allowlist model and have no access by default, including public boards. Access must be explicitly granted for each asset you want the agent to use.

Does using Bring Your Own Agent consume AI credits?

Chatting with an external agent uses your third-party provider's credits, not monday AI credits. monday AI credits are only consumed when the agent performs AI-specific monday.com actions, such as generating an AI column.

If you have any questions, please reach out to our team right here. We’re available 24/7 and happy to help.

Last modified on

Still have questions?
We can help.

Chat support

Get in touch with our support team through our chat for any questions, concerns, or inquiries.

Community forum

Learn, share ideas and connect with other monday.com customers.

Expert help

Hire a monday.com expert to optimize your workflows.

Is this article helpful?
Help us improve our articles.