Setup Connections in Agentforce Builder: The Basic Deployment Workflow

Connecting an Agentforce agent to a real channel is the moment your configuration becomes a live user experience. Until you set up a connection, the agent exists as a design in the builder. Once a connection is configured, the agent can actually receive messages in a specific interface and respond using the correct rules and formats. The exam typically expects you to understand where connection setup happens, what is included in connection settings, and why testing matters before rolling out to real users.

Connection setup is done inside Agentforce Builder in the Connections tab or panel. This is the central place where you link the agent to channels like Salesforce UI experiences, messaging, email, or collaboration platforms. The workflow begins by selecting Add connection. This step creates a dedicated connection configuration that binds the agent to the chosen channel and defines how the interaction should work in that environment.

Connection settings are more than an on or off switch. They include the agent surface and the response formats that the channel should support. The surface defines channel-specific instructions and guidance, while formats define how the agent can present information in that interface. In other words, the connection is where you decide not only that the agent is connected, but also how it should behave and what adaptive response formats it is allowed to use, such as text, buttons, links, images, or video where supported. This is why connection settings are the practical control point for channel-ready behavior.

For Lightning Experience and Mobile, a common scenario is deploying an Employee Agent to internal users. In this case, once the agent is activated and the connection is configured, the agent becomes available through the Agentforce panel. This is the visible user-facing panel inside the Salesforce interface where employees can interact with the agent. The key deployment idea here is that activation matters. The agent can be fully configured but not available until it is activated, and the connection ties that activated agent to the Lightning and Mobile experience.

Testing is a critical step that should never be skipped, especially because channels can differ in how they render output. Message formatting that looks correct in one interface may appear broken or less readable in another. That is why using a test channel is so important. A test channel lets you validate that the connection preserves the intended formatting and that adaptive response elements display correctly. It also lets you confirm that the agent surface instructions are being applied and that the agent behaves as expected in the real interface, not only in design mode. This protects your production rollout, because formatting problems are one of the fastest ways to reduce user trust even when the underlying logic is correct.

Once you have a stable deployment in a test channel and you are satisfied with formatting and behavior, you can confidently move toward broader rollout. At this point, you have completed the basic deployment workflow: you configured the connection in the Connections tab, defined the surface and formats in connection settings, activated the agent, and validated the real channel experience through testing.

Next, we move into a more specific and commonly tested channel: Slack. The next topic will focus on Slack connection setup, key considerations, and how Slack changes the way you design surfaces, formats, and routing for employee interactions.