Free Agentforce-Specialist Practice Test Questions 2026

Explanation:

To generate account action plans using Salesforce’s generative AI features, Universal Containers (UC) should use:

REST API

Why? The Einstein Generative API (part of Salesforce’s AI services) is REST-based and allows:

. Programmatic access to prompt templates and LLM-generated content.
. Integration with legacy systems via HTTP requests.
. Example: Call the API to generate an action plan from a prompt template grounded in Account data.

Why Not the Other Options?

B. "Metadata API":
Used for configuration management (e.g., deploying objects), not generative AI.

C. "SOAP API":
While functional, it’s less efficient for AI integrations compared to REST.

Implementation Steps:

Enable Einstein Generative API in Setup.

Use REST endpoints like:
POST /services/data/vXX.X/einstein/generativeai/prompts

Pass Account data as grounding inputs.

Reference:

Salesforce Help - Einstein Generative API

Explanation

In this scenario, the Einstein Copilot capability that best helps the agent is its ability to execute tasks based on available actions and answer questions using data from Knowledge articles.
Einstein Copilot can assist the service agent by providing relevant Knowledge articles on canceling and rebooking flights, ensuring that the agent has access to the correct steps and procedures directly within the workflow.

This feature leverages the agent’s existing context (the travel itinerary) and provides actionable insights or next steps from the relevant Knowledge articles to help the agent quickly resolve the customer’s needs.

The other options are incorrect:

B refers to invoking a flow to create a Knowledge article, which is unrelated to the task of retrieving existing Knowledge articles.

C focuses on generating Knowledge articles, which is not the immediate need for this situation where the agent requires guidance on existing procedures.

Explanation

Universal Containers (UC) wants to improve sales operations by analyzing data using AI-powered predictions in Einstein Studio. Einstein Studio is Salesforce’s platform for building, deploying, and managing predictive AI models using CRM and external data — typically for structured predictions such as scores, classifications, or numerical forecasts.

🔍 Why Option B is correct:

B. Predict customer lifetime value (CLV) of an account is a classic predictive analytics use case and fits well into Einstein Studio because:

It uses historical sales data, customer behavior, engagement, etc.
It results in a numeric prediction (e.g., total value over X months).

Explanation

Comprehensive and Detailed In-Depth Explanation: The Agentforce Specialist needs a comprehensive view of user interactions, errors, and action issues for troubleshooting. Let’s evaluate the options.

Option A: Plan Canvas Plan Canvas in Agent Builder visualizes an agent’s execution plan for a single interaction, useful for design but not for aggregated troubleshooting data like errors or all interactions, making it incorrect.

Option B: Agent Settings Agent Settings configure the agent (e.g., topics, channels), not provide interaction logs or error details. This is for setup, not analysis, making it incorrect.

Option C: Event Logs Event Logs in Agentforce (accessible via Setup or Agent Analytics) record all user interactions, including errors, incorrectly triggered actions, and incomplete plans. They provide detailed telemetry (e.g., timestamps, action outcomes) for troubleshooting performance issues, making this the correct answer.

Why Option C is Correct: Event Logs offer the full scope of interaction data needed for troubleshooting, as per Salesforce documentation.

Explanation:

In this scenario:

The Prompt Builder test returns the correct grounded response — which means the prompt template and data grounding are working correctly.

But in Agent Builder and the Experience Cloud site, the agent does not return grounded Knowledge article data, suggesting a permissions or access configuration issue with the live agent execution context.

Why B is Correct:
The Agentforce Service Agent (or user running the agent) must have access to the prompt template used for grounding.
If the prompt template is not shared or accessible to that user or role, the AI agent won’t be able to run it during interaction, which would prevent grounding from working in the Experience Cloud or Agent Builder context.

A. Create a new permission set that assigns "Manage Knowledge" and assign it to the Agentforce Service Agent User
❌ Incorrect – “Manage Knowledge” is for creating and publishing Knowledge articles, which is not required for grounding responses in Knowledge. The agent only needs read access.

C. Ensure the Data Cloud User permission set has been assigned to the Agentforce Service Agent User
❌ Incorrect – This permission set relates to Data Cloud features (e.g., segments, unified profiles). It is not required for grounding responses in Knowledge Articles via Agentforce Data Library.

Explanation:

The Einstein Trust Layer in Salesforce is a security framework designed to ensure safe and secure use of generative AI while protecting sensitive data. Among its features, Prompt Defense specifically addresses the risks of jailbreaking and prompt injection attacks, which occur when malicious users manipulate AI inputs to bypass safeguards or elicit unintended responses.

Option A: Secure Data Retrieval and Grounding
This feature ensures AI responses are based on secure, authorized data from Salesforce, enhancing accuracy and context. However, it does not directly address jailbreaking or prompt injection, which involve manipulating prompt inputs rather than data retrieval.

Option B: Data Masking
Data Masking protects sensitive information (e.g., PII) by obscuring it during AI processing, but it does not prevent malicious prompt manipulation. It focuses on data privacy, not prompt security.

Option C: Prompt Defense
Prompt Defense is designed to minimize risks from jailbreaking and prompt injection attacks by validating and sanitizing incoming prompts. It analyzes prompts to detect and block malicious instructions, ensuring the AI behaves as intended and preventing unauthorized outputs. This is explicitly supported by Salesforce’s Einstein Trust Layer Technical Guide and Trailhead: Optimizing AI with Salesforce’s Einstein Trust Layer.

Why It Matters:

Jailbreaking and prompt injection attacks can compromise AI integrity, leading to data leaks or harmful responses. Prompt Defense ensures prompts are processed securely, maintaining system reliability. After enabling Prompt Defense, the Agentforce Specialist should test prompts in a sandbox to verify effectiveness, as recommended in Salesforce Help: Secure AI Operations.

Explanation:

The scenario involves building a Sales Email using a custom prompt template with grounding on:

. Opportunity
. Products
. Events near the customer
. Tone and voice examples

This requires combining multiple objects and possibly external context (like events or tone examples).

🔍 Why Option A is correct:

Prompt-initiated Flows allow:

Dynamically fetching related records across objects (standard or custom).
Bringing in external data or calculated values (e.g., events near customer).
Grounding the prompt template with the real-time, relevant context needed by the AI model.

✅ This is the best practice when:

Grounding needs cross-object logic
Involves non-standard relationships
Requires data transformation before it can be used in a prompt

💡 Example:
A flow can:

Query opportunity line items (products)
Query nearby event records based on location
Pull in tone/voice examples from a config object
Package all this into a grounded input for the template

❌ Why the other options are incorrect:

B. Create a flex template that takes the records in question as inputs
Flex templates support some input flexibility, but do not fetch related records dynamically.
They require inputs to be passed explicitly, which doesn’t scale for multiple, related datasets.

C. Utilize a standard email template and manually insert data fields
Manual insertion defeats the purpose of automation and dynamic AI generation.
Lacks flexibility and scalability — not viable in production AI use cases.

🔗 References
Salesforce Docs — Prompt Builder & Prompt-Initiated Flows
Salesforce Developer Guide — Grounding Custom Prompt Templates

Explanation

The User Utterances dashboard (Option A) is the correct functionality for analyzing user inputs, requests, and queries to identify patterns and trends. This dashboard aggregates and categorizes the natural language inputs (utterances) from users, enabling the Agentforce Specialist to:

Identify Common Queries: Surface frequently asked questions or recurring issues.

Detect Intent Patterns: Understand how users phrase requests, which helps refine intent detection models.

Improve Bot Training: Highlight gaps in training data or misclassified utterances that require adjustment.

Why Other Options Are Incorrect:

B. Agent Event Logs dashboard: Focuses on agent activity (e.g., response times, resolved cases) rather than user input analysis.

C. AI Audit & Feedback Data dashboard: Tracks AI model performance, audit trails, and user feedback scores but does not directly analyze raw user utterances or queries.

Explanation

Einstein Studio (Model Builder) allows organizations to connect and utilize external foundational models while fine-tuning them with company-specific data. This capability is particularly suited to businesses like Universal Containers (UC) that require customization of foundational models to better align with their unique data and use cases.

Option A: Adjusting model temperature is a parameter-level setting for controlling randomness in AI- generated responses but does not necessitate connecting to an external foundational model.

Option B: This is the correct answer because Einstein Studio supports fine-tuning external models with proprietary company data, enabling a tailored and more accurate AI solution for UC.

Option C: Changing frequency penalties is another parameter-level adjustment and does not require external foundational models or Einstein Studio.

Explanation:

To validate that the correct fields are being masked by Einstein Trust Layer's Data Masking, the Agentforce Specialist needs visibility into what data is actually being sent to the LLM. This is only possible if audit data is enabled and stored.

✅ Why Option C is Correct:

Einstein Generative AI Audit Data logs:

. The prompts, responses, and masking actions that occurred.
. Whether fields were masked properly before being sent to the LLM.

To view this, you must first:

Enable collection and storage of Einstein Generative AI Audit Data via the Einstein Feedback Setup page.

This is the required first step to later review, validate, and troubleshoot masked fields.

Once enabled, you can:
Export audit logs.
Review field-level data sent to the LLM.
Confirm whether sensitive data (like PII) was correctly masked or anonymized.

❌ Why the Other Options Are Incorrect:

A. Use a Flow-based resource to debug merge fields
Merge field debugging doesn't reveal whether masking occurred — it only shows runtime field values.
Doesn’t validate Trust Layer masking behavior.

B. Request audit data from the Security section
There's no manual "request" option for audit data there.
You must first enable collection in the Feedback settings (Option C) before anything is available to view.

🔗 References:
Salesforce Help — Enable Einstein Generative AI Audit Data
Salesforce Trust Layer Overview

Explanation:

To audit, track, and view masked data with the Einstein Trust Layer, organizations should:

Use the Audit Trail in Setup

The Trust Layer automatically logs:
. All prompts submitted to the LLM (including timestamps and users).
. Masked sensitive data (e.g., PII replaced with placeholders like [EMAIL]).
Admins can export these logs for compliance or troubleshooting.

Why Not the Other Options?

A. "Audit trail in Data Cloud":
Incorrect. The Trust Layer’s audit logs are native to Salesforce, not stored in Data Cloud.

B. "Use Prompt Builder to request masked data":
Prompt Builder cannot retroactively reveal masked data—this would violate the Trust Layer’s security model.

Steps to Access Logs:

1. Go to Setup → Audit Trail.
2. Filter for "Einstein Generative AI" events.
3. Export logs to analyze:
. Raw prompts (with masked fields).
. User/IP details.
. Response metadata.

This ensures transparency without compromising data security.

Explanation:

When grounding a prompt template using related lists in Salesforce Prompt Builder, not all related lists are supported. Specifically:

The User related list is not supported for grounding in prompt templates.
This is due to security model constraints and data structure limitations in how user data is accessed and exposed to AI models via the Einstein Trust Layer.

❌ Why the other options are incorrect:

A. The User related list should have View All access
Even if full access is granted, this related list still isn't supported for grounding in prompt templates. So this access setting has no effect.

B. The User related list needs to be included on the record page
Including a related list on the record page is often necessary for visibility, but it has no impact on grounding availability. And again, the User related list is unsupported, so its presence on the layout doesn’t matter.

🔗 Reference:
Salesforce Help — Prompt Builder Related List Grounding
Prompt Template Limitations — Salesforce Docs