Salesforce · Als-Con-201
Core Concept
Agentforce Life Sciences for Customer Engagement leverages generative AI capabilities to produce intelligent provider summaries that consolidate critical information about Healthcare Providers (HCPs) into digestible narratives. These summaries draw upon data stored across multiple related objects within the Salesforce platform, including provider demographics, interaction history, affiliations, and preferences. The Actionable Relationship Center (ARC) graph serves as the underlying data model that defines which objects and fields are included in the summarization process.
Cumulus Pharma wants to customize the provider summary to include additional information stored in Agentforce Life Sciences for Customer Engagement. An Agentforce Life Sciences Consultant updated the Actionable Relationship Center (ARC) graph used in provider summarization to include additional objects and fields.
How should the consultant ensure that Provider Summary Generation uses the new information to meet business goals?
Provider Summarization and Prompt TemplatesAgentforce Life Sciences for Customer Engagement leverages generative AI capabilities to produce intelligent provider summaries that consolidate critical information about Healthcare Providers (HCPs) into digestible narratives. These summaries draw upon data stored across multiple related objects within the Salesforce platform, including provider demographics, interaction history, affiliations, and preferences. The Actionable Relationship Center (ARC) graph serves as the underlying data model that defines which objects and fields are included in the summarization process.
Provider summarization in Agentforce relies on prompt templates that contain instructions for how the AI should interpret and synthesize the input data. These templates define the tone, structure, and content focus of the generated summaries. When the ARC graph is updated to include additional objects or fields, the prompt template must be correspondingly updated to ensure the AI model receives and processes this new information appropriately.
The prompt template versioning system allows consultants to iterate on summarization logic without disrupting production processes. Each version can specify different input datasets, modify the instructions given to the generative model, and refine the output format. Activating a new prompt template version makes it the current template used by the Provider Summary Generation process, ensuring that newly expanded data sources are incorporated into future summaries.
Detailed Explanation
When an Agentforce Life Sciences Consultant modifies the ARC graph to include additional objects and fields, this change expands the data available for provider summarization but does not automatically integrate that data into the summarization output. The Provider Summary Generation feature relies on prompt templates to determine what data to pull, how to interpret it, and how to structure the narrative summary.
Prompt templates in Agentforce contain two critical components: the input dataset specification and the summarization instructions. The input dataset defines which objects and fields from the ARC graph should be fed to the AI model. The summarization instructions provide guidance to the generative AI on how to process this data — what aspects to emphasize, how to structure the output, and what tone to adopt.
To ensure that newly added objects and fields contribute to provider summaries, the consultant must create a new version of the existing prompt template. Within this new version, they update the input dataset to reference the additional objects and fields now present in the ARC graph. They also refine the summarization instructions to guide the AI on how to incorporate this new information meaningfully — for example, instructing the model to highlight recent affiliations or to incorporate specific preference data.
After creating and configuring the new prompt template version, the consultant must activate it. Activation designates this version as the current template that the Provider Summary Generation process will use. Without activation, the new template remains in draft status and does not affect the summaries generated for field sales representatives.
Option A correctly identifies the configuration pathway required to incorporate new ARC graph data into provider summaries. Creating a new version of the prompt template allows the consultant to modify both the input dataset and the summarization instructions in a controlled, version-managed manner. Updating the input dataset ensures that the newly added objects and fields from the ARC graph are included in the data fed to the generative AI model. Updating the summarization instructions ensures the AI knows how to interpret and present this new information.
Activating the new prompt template version is the essential final step that makes the updated configuration operational. This action tells the Provider Summary Generation system to use the enhanced prompt template for all future summarization tasks, thereby meeting Cumulus Pharma's business goal of including additional information in their provider summaries.
This approach aligns with Salesforce's best practice of using versioned templates to manage AI-driven content generation, providing traceability, rollback capability, and clear governance over what data and instructions drive automated summaries.
Option B
Customizing the Generate Provider Account Territory Summary flow is not the correct approach because the flow is responsible for orchestrating the summarization process, not defining what data is included or how it is interpreted. While flows can trigger summarization jobs and handle pre- or post-processing tasks, they do not control the input dataset or the AI instructions that determine summary content. The prompt template, not the flow, governs which objects and fields are used and how the AI synthesizes them into narrative form. Modifying the flow would not ensure that the newly added ARC graph data appears in the generated summaries.
Option C
Updating the Batch Summarization job configuration does not address the core requirement. The batch job is responsible for scheduling and executing summarization tasks at scale, processing multiple provider records according to defined criteria. However, it does not define which objects and fields are included in the summarization or how the AI interprets that data. The batch job operates based on the active prompt template. Without updating the prompt template itself to reference the new ARC graph data, the batch job will continue generating summaries using the previous, incomplete dataset. This option confuses job execution configuration with content definition.
A life sciences organization uses Agentforce to help sales representatives prepare for Healthcare Provider meetings. Initially, provider summaries include basic information such as specialty, practice location, and recent interaction history. The organization later implements a custom object to track HCP research interests and another to capture preferred communication channels.
The Agentforce Life Sciences Consultant updates the ARC graph to include these new objects and their relevant fields. To ensure this information appears in provider summaries, the consultant navigates to the prompt template management interface and creates a new version of the existing "Provider Summary" template.
In the new version, the consultant adds the Research Interest and Communication Preference objects to the input dataset specification. They also update the summarization instructions to include a directive such as: "When research interests are available, highlight the provider's top three areas of focus. Include preferred communication channels to help the representative tailor their outreach strategy."
After saving the new version, the consultant activates it. The next time the batch summarization job runs, field sales representatives receive enhanced provider summaries that include research interests and communication preferences, enabling more personalized and effective engagement strategies.