When naming actions for an AI agent, consistency is key to helping the Large Language Model (LLM) categorize and understand the available tools. Following a consistent naming convention, such as starting all retrieval actions with the verb "Get," allows the agent to more easily map user intent (e.g., "I need information about...") to the appropriate action. This reduces ambiguity during the reasoning phase of the agentic loop. Using inconsistent verbs like Find, Retrieve, or Identify (Option D) can confuse the model regarding which action is most appropriate for a given task.
Removing verbs entirely (Option A) makes it harder for the model to distinguish between an action (doing something) and a data object (a thing). Adding "Salesforce" (Option B) is generally redundant as the agent's context is already within the Salesforce environment.

In the Agentforce environment, deactivating an agent to make configuration changes (like adding a Topic or Action) does not immediately terminate existing sessions. To ensure a smooth user experience, Salesforce allows ongoing conversations to continue using the version of the agent that was active when the session started. The "deactivation" simply prevents new sessions from being initiated. Once the administrator reactivates the agent, new sessions will utilize the updated configuration, while current users finish their interactions without interruption or "system errors".
The window does not automatically close, and the conversation is not "paused" in a way that requires the user to wait for reactivation

To align the Salesforce mobile app with a company's visual identity, a Platform Administrator can customize specific branding elements through the Salesforce Mobile App Customization settings.
The two primary areas available for branding are:
Loading Page Logo: This allows the administrator to upload a company logo that appears when a user first opens the app.
Header Background Color: This allows the administrator to set the primary brand color for the top header area of the mobile interface.
These settings ensure that the user experience feels consistent with the company's internal branding from the moment they log in. Option A and B are incorrect because Salesforce does not currently allow for granular "App header" vs "Header background" color separation or the modification of the "Record background color," as the app maintains a standardized background for accessibility and readability. Customizing the loading screen and header is the standard way to implement a mobile rebrand.

To track specific stages of a support process and measure the time elapsed within those stages, Salesforce provides Entitlements and Milestones. Milestones represent required steps in a support process, such as "First Response Time" or "Resolution Time". When combined with Business Hours, Milestones allow the system to accurately calculate how long a case has been in a particular status or assigned to a specific team, excluding weekends or non-working hours. This provides leadership with granular reporting on "Team Performance" and SLA compliance.
Escalation Rules (Option A) and Assignment Rules (Option C) are primarily routing tools; they can move a case but do not natively provide the timestamp-based tracking and reporting necessary to see duration across multiple "handoffs". Record-Triggered Flows (Option B) could theoretically be used to stamp fields, but this would require significant custom development and would not offer the native, out-of-the-box reporting dashboards that come with the Milestones feature. Therefore, Entitlement Management is the standard solution for tracking time-based service metrics.

For an AI agent to provide "complete and accurate" support, it needs access to a unified, 360- degree view of the customer data. Data Cloud is the critical prerequisite because it ingests, harmonizes, and unifies data from multiple sources (Salesforce, external databases, legacy systems) into a single "Unified Profile." When Agentforce is "grounded" in Data Cloud, it can reference real-time customer interactions, purchase history, and cross-channel behavior that might not exist in a single Case or Contact record. Without Data Cloud, the agent's knowledge is limited to siloed data, which increases the risk of providing incomplete or irrelevant answers.
While Console layouts (Option B) and intake methods (Option A) are important for the UI, they do not provide the data foundation necessary for advanced AI reasoning.