Understanding the Root Cause

The actual OAuth tokens and connection status between a Salesforce Org and a Slack Workspace are sensitive and environment-specific. Salesforce does not expose active OAuth connections as a simple XML metadata file (.xml) that you can deploy, because a connection authorized in Sandbox cannot validly “transfer” to Production (it would point to the wrong org ID or lack the correct auth context).

This is why you cannot find the connection configuration in the MetaData API. Instead, you need to configure it manually in each environment.

Deployable Metadata

While the connection configuration is not deployable, there are other metadata objects that you can deploy via your DevOps pipeline. These include:

• Bot Definition: BotDefinition (The Agent itself, topics, instructions)
• Conversation Definition: ConversationDefinition (or similar new types for Agentforce configurations)
• Named Credentials / External Credentials: If your agent uses API calls, the definitions (NamedCredential) are deployable, but the Authentication Principals (passwords/tokens) often need re-authentication in the target org
• Connected App: If you created a custom Connected App for this, that is deployable

Manual Configuration Step

The step where you click “Connect” to link a specific Slack Workspace to your Agent is a runtime authorization. This step needs to be performed manually in each environment.

Post-Deployment Manual Step

Step 1: Deploy the Agent metadata via DevOps pipeline.
Step 2: Log into Staging/Production manually.
Step 3: Go to Setup > Agentforce > Slack (or the specific integration page).
Step 4: Click "Connect" and authorize the Slack Workspace.

Best Practices

Heads up: Document this as a “Post-Deployment Manual Step” in your release plan.

Checklist for Deploying Agentforce to Slack

• Deploy the Agent metadata via DevOps pipeline
• Log into Staging/Production manually
• Go to Setup > Agentforce > Slack (or the specific integration page)
• Click “Connect” and authorize the Slack Workspace
• Test the Agentforce configuration in each environment
• Document the manual configuration step in your release plan
• Review and update the Agentforce configuration as needed

Frequently Asked Questions

What is the difference between deployable metadata and environment-specific configuration?

Deployable metadata refers to the Agent logic and configuration that can be deployed via your DevOps pipeline, while environment-specific configuration refers to the connection configuration that needs to be performed manually in each environment.

Can I automate the manual configuration step?

No, the manual configuration step needs to be performed manually in each environment, as it requires runtime authorization and cannot be automated.

What happens if I try to deploy the connection configuration via MetaData API?

The connection configuration will likely break or not work, as it is environment-specific and cannot be deployed via MetaData API.

How do I test the Agentforce configuration in each environment?

You can test the Agentforce configuration by logging into each environment and verifying that the Agent is working as expected.

What are the best practices for deploying Agentforce to Slack?

The best practices include documenting the manual configuration step in your release plan, testing the Agentforce configuration in each environment, and reviewing and updating the Agentforce configuration as needed.