Personalization and Recommendations

Available in: Lightning Experience in Enterprise and Unlimited editions for an additional cost. To purchase, contact your Salesforce account executive.

Available in: Aura Experience Cloud sites that use Build Your Own Template

Available in: LWR Experience Cloud sites that use Build Your Own Template

Personalization is the ability of the Self-Service agent to tailor its interactions based on a user’s unique identity, history, and real-time behavior. Instead of providing a generic experience, the agent uses Data Cloud DataGraphs to ground its greetings and recommendations in actual customer data. This ensures that every interaction, from the initial greeting to the suggested next steps, is relevant to the user’s current needs.

The system references your configured Personalization Points to show suggestions in two primary locations:

• Interactive Buttons: Small buttons appear directly below the Unified Prompt Bar. These buttons represent specific Suggestion Themes. When a user clicks a button, the system displays an expanded list of relevant actions, such as starting an intake form from the Unified Catalog.

• Concierge Sidebar: Grouped suggestion themes appear in the collapsible sidebar. This allows users to browse categories of recommendations and persistent conversation history in one place.

To turn on personalization, the Agentforce Orchestrator references several foundational layers:

• Data Cloud DataGraph: This serves as the source of truth for user context. It provides the unified profile data, such as case history and preferences, that the AI needs to make accurate recommendations.

• Prompt Builder Templates: These templates define the logic for how the AI speaks. They reference the DataGraph to inject user-specific details into greetings and suggestion summaries.

• Personalization Points: Personalization Points are configuration records in Setup that define the placement. They act as logical anchors that tell the orchestrator where a specific greeting or suggestion is allowed to appear on your Experience Cloud site. They bridge the gap between your backend data and where the content actually appears for the user.

The personalization engine uses two complementary approaches to generate recommendations. Deterministic rules follow strict administrative logic, while probabilistic models use AI to predict user needs based on patterns. When the personalization service is unavailable or a user has no engagement history, the system shows the Static Greeting Message defined in your Agentforce Orchestrator configuration. This ensures the UI remains functional and welcoming even without a personalized reference.

Personalization and Recommendation Core Elements

Element	Role in Personalization
DataGraph	Provides the data the AI knows about the user.
Prompt Template	Determines how the AI formats the greeting or suggestion.
Personalization Point	Defines where the content appears on the Experience Cloud site.