Pass the Salesforce AI Specialist Agentforce-Specialist Questions and answers with CertsForce

The correct answer is B because gating downstream actions requires a reliable state variable and an action availability rule. A mutable session variable can store whether patient verification has been completed, and available when can prevent scheduling actions from being available until that variable indicates verified status. This is the standard declarative Agent Script pattern for controlled action access. Option A is unnecessarily code-heavy and shifts simple state gating into Apex when Agentforce can manage it with variables and availability rules. Option C only changes descriptive text; it does not enforce whether the scheduling action can be used. In compliance workflows, the control must prevent the action from being callable, not merely tell the model what to do. Salesforce context engineering examples show variables and action filters being used to control action availability.

The correct answer is C because this requirement must be enforced deterministically at the action-availability layer, not trusted to natural language instructions. In Agentforce, filters can control access to subagents and actions so the agent can use them only when defined conditions are met. For refund processing, the “Issue Refund” action should be hidden from the reasoning engine unless the account tier condition evaluates to Platinum. Option A is weak because system instructions are interpreted by the model and can be ignored or misapplied. Option B is wrong because blocking a keyword is brittle and does not validate entitlement. Action filters are the correct control because they prevent the action from being available until the business rule is satisfied.

UC’s development team needs to identify a core component of a Custom Agent Action in Agent Builder. Let’s assess the options.

Option A: Action Triggers " Action Triggers " isn’t a term used in Agentforce Custom Agent Action configuration. Actions are invoked by topics or plans, not standalone triggers, making this incorrect.

Option B: InstructionsInstructions are a core component of a Custom Agent Action in Agentforce. Defined in Agent Builder, they guide the Atlas Reasoning Engine on how to execute the action (e.g., what to do with inputs, how to process data). Reviewing the instructions helps the team understand the action’s purpose and logic, making this the correct answer.

Option C: Output TypesWhile outputs are part of an action’s result, " Output Types " isn’t a distinct configuration element in Agent Builder. Outputs are determined by the action’s execution (e.g., Flow or Apex), not a separate setting, making this less core and incorrect.

Why Option B is Correct:

Instructions are a fundamental component of Custom Agent Actions, providing the AI’s execution directives, as per Salesforce documentation.

Dynamic Grounding in the Einstein Trust Layer architecture ensures that large language model (LLM) prompts are enriched with organization-specific data (e.g., Salesforce records, Knowledge articles) to generate accurate and relevant responses. By dynamically injecting contextual data into prompts, it reduces hallucinations and aligns outputs with trusted business data.

Prompt Defense (A) focuses on blocking malicious inputs or prompt injections but does not enhance responses with organizational data.

Data Masking (B) redacts sensitive information but does not contribute to grounding responses in business context.

When Universal Containers creates a new Sales Email prompt template using the " Save As " function, missing hyperparameters can result in different outputs. To ensure the new prompt produces comparable results to the standard Sales Email prompt, the Agentforce Specialist should manually add the necessary hyperparameters to the new template.

Hyperparameters like Temperature, Frequency Penalty, and Presence Penalty directly affect how the AI generates responses. Ensuring that these are consistent with the standard template will result in similar outputs.

Option A (Model Playground) is not necessary here, as it focuses on fine-tuning models, not adjusting templates directly.

Option C (Reverting to the standard template) does not solve the issue of customizing the prompt template.

For more information, refer to Prompt Builder documentation on configuring hyperparameters in custom templates.

Service Replies: This generative AI feature automates and assists in generating accurate, contextual, and efficient replies for customer service agents. It uses past interactions, case data, and the context of the conversation to provide draft responses, thereby enhancing productivity and reducing response times.

Case Summaries: Summarizes case information but does not assist directly in replying to customer chats.

Case Escalation: Refers to moving cases to higher-level support teams but does not address the need to improve chat response productivity.

Thus, Service Replies is the best feature for this requirement as it directly aligns with improving agent efficiency in replying to chats.

The correct answer is A because the observed failure is caused by ambiguous subagent matching, not by missing content or insufficient agent count. In Agentforce, subagent classification depends heavily on the subagent’s name, classification description, scope, instructions, and actions. If two subagents semantically overlap, the reasoning engine can route the same user intent to the wrong place. Refining the competing classification descriptions and narrowing scope removes ambiguity and improves deterministic routing. Option B can make the architecture worse by adding more subagents before fixing the overlap. Option C is wrong because knowledge articles help answer questions after routing; they do not solve subagent selection errors. Salesforce troubleshooting guidance specifically recommends checking the subagent name and classification description when expected utterances are misclassified.

The task requires analyzing user inputs, requests, and queries to identify patterns and trends in Agentforce interactions. Let’s assess the options based on Agentforce’s analytics capabilities.

Option A: Agent Event Logs dashboard.Agent Event Logs capture detailed technical events (e.g., API calls, errors, or system-level actions) related to agent operations. While useful for troubleshooting or monitoring system performance, they are not designed to analyze user inputs or conversational trends. This option does not meet the requirement and is incorrect.

Option B: AI Audit and Feedback Data dashboard.There’s no specific " AI Audit and Feedback Data dashboard " in Agentforce documentation. Feedback mechanisms exist (e.g., user feedback on responses), and audit trails may track changes, but no single dashboard combines these for analyzing user queries and trends. This option appears to be a misnomer and is incorrect.

Option C: User Utterances dashboard.The User Utterances dashboard in Agentforce Analytics is specifically designed to analyze user inputs, requests, and queries. It aggregates and visualizes what users are asking the agent, identifying patterns (e.g., common topics) and trends (e.g., rising query types). Specialists can use this to refine agent instructions or topics, making it the perfect tool for this task. This is the correct answer per Salesforce documentation.

Why Option C is Correct:

The User Utterances dashboard is tailored for conversational analysis, offering insights into user interactions that align with the specialist’s goal of identifying patterns and trends. It’s a documented feature of Agentforce Analytics for post-deployment optimization.

In Salesforce Agentforce, a Data Library is a feature that allows organizations to upload files (e.g., PDFs, documents) to be used as grounding data for AI-driven agents. Once the Data Library is created and configured, the uploaded files are indexed to make their content searchable and usable by the AI (e.g., for retrieval-augmented generation or prompt enhancement). The key question is where this indexing occurs. Salesforce Agentforce integrates tightly with Data Cloud, a unified data platform that includes a vector database optimized for storing and indexing unstructured data like uploaded files. When a Data Library is set up, the files are ingested and indexed into Data Cloud’s vector database, enabling the AI to efficiently retrieve relevant information from them during conversations or actions.

Option A: Indexing files in a " location specified by the user " is not a feature of Agentforce Data Libraries. The indexing process is managed by Salesforce infrastructure, not a user-defined location.

Option B: This is correct. Data Cloud handles the indexing of uploaded files, storing them in its vector database to support AI capabilities like semantic search and content retrieval.

Option C: Salesforce File Storage (e.g., where ContentVersion records are stored) is used for general file storage, but it does not inherently index files for AI use. Agentforce relies on Data Cloud for indexing, not basic file storage.

Thus, Option B accurately reflects the process after a Data Library is created and configured in Agentforce.

Salesforce Agentforce Documentation: " Set Up a Data Library " (Salesforce Help: https://help.salesforce.com/s/articleView?id=sf.agentforce_data_library.htm & type=5)

Salesforce Data Cloud Documentation: " Vector Database for AI " (https://help.salesforce.com/s/articleView?id=sf.data_cloud_vector_database.htm & type=5)

Einstein Trust Layer’s Toxicity Scoring categorizes content on a scale of 0 to 1, where 1 indicates the highest level of toxicity (e.g., harmful, biased, or inappropriate language). This scoring helps organizations filter unsafe AI-generated content. A score of 1 triggers mitigation actions, such as blocking the response or alerting administrators.

A score of 0 would indicate no toxicity (B is incorrect).

The scoring system does not use " least toxic " as a category (A is misleading).