Question bank
Salesforce Certified Administrator
41 Salesforce Certified Administrator questions with the correct answer and a written explanation, organised across 8 exam domains.
- 41 questions
- 8 domains
- Foundation level
certilumQuestion index
Jump straight to any question. Answers and explanations are inline.
Agentforce
4 questions in this domain
- Question 1Single Choice
Why should an administrator configure an Agentforce agent to focus on a specific business domain such as Service or Sales?
- ATo prevent the agent from attempting tasks outside its intended boundariesCorrect answer
- BTo make the agent automatically access all Salesforce data
- CTo remove the need for agent-specific permission sets
- DTo allow the agent to work without any instructions
Why this is the answer
The agent role and purpose must be scoped to specific business domains so the agent does not attempt tasks outside its intended operational boundaries.
- Question 2Multiple Select
Which items must be explicitly configured when grounding an Agentforce agent to CRM data?
- ASpecific objectsCorrect answer
- BSpecific fieldsCorrect answer
- CKnowledge articlesCorrect answer
- DAutomatic access to all CRM data
Why this is the answer
Administrators must define the objects, fields, and Knowledge articles the agent can use. The agent does not automatically receive access to all data.
- Question 3Interactive
Review the configuration snippet in the code block. Which Agentforce component is being defined?
- AInstructions and GuardrailsCorrect answer
- BActions and Tooling
- CData Grounding
- DEinstein Trust Layer
Why this is the answer
The configuration contains high-level directives and boundaries defining how the agent should behave, which is the Instructions and Guardrails component.
- Question 4Single Choice
What is the purpose of masking sensitive PII before sending data to the LLM in the Einstein Trust Layer?
- ATo enforce data privacy and securityCorrect answer
- BTo make the agent respond faster
- CTo grant additional user permissions
- DTo bypass Salesforce sharing rules
Why this is the answer
Masking sensitive PII before sending to the LLM prevents raw personal data from leaving Salesforce, aligning with data privacy and security requirements.
Automation
6 questions in this domain
- Question 5Single Choice
Which configuration element defines the specific persona and scope of an agent?
- AAgent Role & PurposeCorrect answer
- BActions & Tooling
- CReasoning Engine
- DData Grounding
Why this is the answer
Agent Role & Purpose defines the persona and scope of the agent, preventing it from acting outside its intended business domain.
- Question 6Multiple Select
Which two components execute logic as Actions, rather than merely generate text?
- AApex classesCorrect answer
- BFlow Builder flowsCorrect answer
- CPrompt Templates
- DKnowledge Articles
Why this is the answer
Flow Builder flows and Apex classes execute logic, while Prompt Templates only generate text and Knowledge Articles are data sources.
- Question 7Interactive
The configuration snippet in the code block represents which key concept?
- AData GroundingCorrect answer
- BAgent Role & Purpose
- CActions & Tooling
- DEinstein Trust Layer
Why this is the answer
Data Grounding connects the agent to CRM objects, fields, and Knowledge articles so it can answer using real customer data.
- Question 8Single Choice
What is the purpose of the Einstein Trust Layer integration in an agent configuration?
- AMasking sensitive PII before sending to the LLMCorrect answer
- BReplacing the reasoning engine
- CGranting the agent access to all CRM data
- DCreating flowcharts for the agent
Why this is the answer
The Einstein Trust Layer masks PII before data is sent to the LLM and helps enforce the organization's data-sharing rules.
- Question 9Multiple Select
Which two mechanisms should an administrator use to ensure a human approves high-stakes refund requests?
- AHuman-in-the-LoopCorrect answer
- BGuardrails and ConstraintsCorrect answer
- CDynamic Context Injection
- DAgent Role & Purpose
Why this is the answer
Human-in-the-Loop requires human approval, and guardrails define boundaries that prevent unauthorized autonomous actions.
- Question 10Interactive
The code block shows a safety rule for an agent. Which concept is being illustrated?
- AHuman-in-the-LoopCorrect answer
- BDynamic Context Injection
- CData Grounding
- DAgent Role & Purpose
Why this is the answer
The agent pauses its workflow to request human approval, which is the Human-in-the-Loop pattern.
Configuration and setup
6 questions in this domain
- Question 11Single Choice
What is the primary purpose of defining an Agent's Role and Purpose during configuration?
- AIt defines the specific persona and operational boundaries of the agent.Correct answer
- BIt automatically grants the agent access to all Salesforce data.
- CIt replaces the need for instructions and guardrails.
- DIt generates Apex code for the agent.
Why this is the answer
Defining the role and purpose sets the agent's persona and scope, ensuring it focuses on intended business domains like Service or Sales.
- Question 12Multiple Select
Which of the following are considered Agentic Actions or Tools in Agentforce?
- AFlow Builder flowsCorrect answer
- BApex classesCorrect answer
- CPrompt TemplatesCorrect answer
- DKnowledge article records
Why this is the answer
Agentic actions include Flow Builder flows, Apex classes, and Prompt Templates. Knowledge articles are data used for grounding, not actions.
- Question 13Interactive
Based on the code snippet, what type of Agentic Action is being configured for the agent?
- AApex class
- BFlowCorrect answer
- CPrompt Template
- DData grounding
Why this is the answer
The implementationType is set to Flow and the flowApiName references a Flow Builder flow, so this is a Flow action.
- Question 14Single Choice
What is dynamic context injection in Agentforce?
- APulling relevant record data into the agent's reasoning processCorrect answer
- BGenerating text using Prompt Templates
- CMasking sensitive PII before sending data to the LLM
- DCreating a visual flowchart for the agent's conversation path
Why this is the answer
Dynamic context injection pulls record data like Case details into reasoning, allowing personalized, data-driven responses rather than generic replies.
- Question 15Multiple Select
Which configuration settings are part of Einstein Trust Layer Integration?
- AMasking PII before data is sent to the LLMCorrect answer
- BAdhering to the organization's data sharing rulesCorrect answer
- CGiving the agent access to every Salesforce object
- DRequiring a flowchart for every agent conversation
Why this is the answer
Einstein Trust Layer Integration masks sensitive PII and follows org data sharing rules. It does not grant universal data access.
- Question 16Interactive
The code snippet represents which Agentforce configuration concept?
- AInstructions & GuardrailsCorrect answer
- BData Grounding
- CReasoning Engine
- DDynamic Context Injection
Why this is the answer
The snippet defines autonomy limits, human approval, and allowed data objects, which are all guardrails and constraints.
Data and analytics management
7 questions in this domain
- Question 17Single Choice
In an Agentforce deployment, what is the main purpose of defining the agent’s role and persona?
- AKeeps agent within its intended operational boundariesCorrect answer
- BGuarantees a 100% task completion rate
- CAutomatically creates Flows and Apex classes
- DReplaces Salesforce security permissions entirely
Why this is the answer
Defining a persona and scope prevents the agent from attempting work outside its designated business domain, such as using Service actions for Sales tasks.
- Question 18Multiple Select
Which of the following components should be classified as agentic actions rather than as governance settings when configuring an Agentforce agent?
- AFlow Builder flowCorrect answer
- BApex classCorrect answer
- CPrompt TemplateCorrect answer
- DReasoning engine
Why this is the answer
Agentic actions are executable capabilities like Flows, Apex, or Prompt Templates. The reasoning engine makes decisions rather than being a tool the agent invokes.
- Question 19Interactive
Based on the configuration snippet provided, which action is intended to perform a business operation instead of generating text?
- AFlow named UpdateCaseStatusCorrect answer
- BPrompt Template named Summarize Case
- CData access for Case
- DAgent name
Why this is the answer
Prompt Templates generate text; a Flow is an action that executes logic in the agent. Data access is grounding, not an executable action.
- Question 20Single Choice
What does the process of data grounding accomplish when an administrator configures an Agentforce agent?
- AProvides contextually accurate responses using real-time dataCorrect answer
- BGenerates Apex classes for every object
- CCreates a flowchart for agent conversations
- DRemoves all security on the selected objects
Why this is the answer
Data grounding connects the agent to the right objects, fields, and Knowledge articles so answers are based on real-time CRM data and are contextually accurate.
- Question 21Multiple Select
Which of the following settings are examples of guardrails that control an Agentforce agent's behavior and data access?
- ARestricting objects and fields accessible to the agentCorrect answer
- BSetting explicit allowed actions for the agentCorrect answer
- CRequiring human approval for high-risk processesCorrect answer
- DGranting unrestricted access to all Salesforce data
Why this is the answer
Guardrails include boundaries on data access, action permissions, autonomy, and human-in-the-loop approval to prevent unauthorized or unsafe agent behavior.
- Question 22Interactive
Based on the security snippet provided, how should an administrator interpret the running-user context setting for data exposure?
- AAgent obeys the interacting user's permissionsCorrect answer
- BAgent uses a tenant-wide reusable context
- CAgent bypasses all sharing rules
- DAgent shares data with all users
Why this is the answer
With running-user context enabled, the agent respects the current user's permission set rather than using a broad system context, preventing data exposure.
- Question 23Single Choice
What is a primary security task the Einstein Trust Layer performs before sending data to an external LLM?
- AMasking PII before the LLM receives itCorrect answer
- BCreating new record owners
- CBuilding the agent's reasoning flow
- DDeleting unnecessary Knowledge articles
Why this is the answer
The Einstein Trust Layer masks sensitive PII before requests go to the LLM and enforces data sharing rules to maintain privacy.
Object Manager and Lightning App Builder
6 questions in this domain
- Question 24Single Choice
What is the primary purpose of defining an agent's role and purpose in Agentforce?
- AIt keeps the agent focused on specific business domains and prevents out-of-scope tasks.Correct answer
- BIt gives the agent automatic access to all Salesforce data.
- CIt creates a step-by-step flowchart for every interaction.
- DIt removes the need to configure actions and tooling.
Why this is the answer
Configuring the agent's role and purpose defines its persona and operational scope, preventing it from attempting tasks outside its intended business domain.
- Question 25Multiple Select
Which of the following must an administrator configure when setting up an agent? Select all that apply.
- AActions and ToolingCorrect answer
- BData GroundingCorrect answer
- CEinstein Trust Layer IntegrationCorrect answer
- DA visual flowchart for all user requests
Why this is the answer
Actions and Tooling, Data Grounding, and Einstein Trust Layer Integration are core configuration areas. A flowchart is not required because flows are tools the agent uses, not the map it follows.
- Question 26Interactive
The provided configuration snippet defines an agent with a specific domain and assigned actions. What concept is represented by the actions section?
- AAgentic ActionsCorrect answer
- BReasoning Engine
- CDynamic Context Injection
- DEinstein Trust Layer Integration
Why this is the answer
The action list contains specific assigned capabilities such as Flow and Apex, which are Agentic Actions.
- Question 27Single Choice
What is the main function of the reasoning engine in an agent?
- AIt interprets the input and decides which Salesforce actions are required.Correct answer
- BIt automatically approves all high-stakes actions without review.
- CIt masks all sensitive data before the LLM processes it.
- DIt generates a visual flowchart for the conversation.
Why this is the answer
The reasoning engine analyzes the user's goal, interprets natural language, and decides which Salesforce actions are needed using available metadata.
- Question 28Multiple Select
Which statements about agent data access are true? Select all that apply.
- AAgent data access must be explicitly configured through grounding and permissions.Correct answer
- BAn agent can access data only if it has been granted specific action permissions.Correct answer
- CAn agent operates within the context of the user's permissions and its own security settings.Correct answer
- DAn agent automatically inherits access to all objects and fields.
Why this is the answer
Administrators must explicitly configure an agent's data access through grounding and action permissions, and the agent operates under both user and agent-level security settings.
- Question 29Interactive
The code block pauses the agent's workflow to wait for approval. Which concept does this represent?
- AHuman-in-the-LoopCorrect answer
- BDynamic Context Injection
- CAgentic Actions
- DEinstein Trust Layer Integration
Why this is the answer
Human-in-the-Loop is a pattern where the agent pauses its autonomous workflow to request human approval, especially for high-stakes actions.
Productivity and collaboration
4 questions in this domain
- Question 30Single Choice
What is the primary purpose of configuring an Agent Role & Purpose in Salesforce Agentforce?
- ADefines the specific persona and scope of the agentCorrect answer
- BCreates a linear flow for the agent to follow
- CMasks sensitive PII before sending data to the LLM
- DAssigns the agent a permission set for all data
Why this is the answer
Agent Role & Purpose defines the specific persona and business scope, preventing the agent from attempting tasks outside its intended domain.
- Question 31Multiple Select
Which of the following are examples of Agentic Actions, or the tools an agent can use? Select all that apply.
- AFlow Builder flowsCorrect answer
- BApex classesCorrect answer
- CPrompt TemplatesCorrect answer
- DReasoning Engine context
Why this is the answer
Agentic Actions are the tools assigned to an agent: Flow Builder flows, Apex classes, and Prompt Templates. The Reasoning Engine is the logic layer, not an action.
- Question 32Interactive
Review the configuration excerpt below. Which type of agent capability does the component named UpdateCaseStatus represent?
- AAgentic ActionCorrect answer
- BPrompt Template
- CData Grounding
- DGuardrail
Why this is the answer
The excerpt describes an Apex class invocation, which is an Agentic Action. Prompt Templates generate text, while grounding and guardrails serve different purposes.
- Question 33Single Choice
What does Data Grounding ensure in an Agentforce agent configuration?
- AConnects the agent to the correct Salesforce objects, fields, and Knowledge articlesCorrect answer
- BPlaces the agent in a linear flow to resolve cases
- CMasks PII before the agent processes a conversation
- DPauses the agent for human approval on high-stakes actions
Why this is the answer
Data Grounding connects the agent to CRM data, including objects, fields, and Knowledge, so it can use real-time data in responses.
Sales and marketing applications
4 questions in this domain
- Question 34Single Choice
What is the main purpose of defining an agent role in Agentforce configurations for a Sales team?
- ARestrict to specific business domains.Correct answer
- BGrant access to all Salesforce data.
- CSet chat window branding.
- DProvide a fixed decision flowchart.
Why this is the answer
An agent role defines its persona and scope, preventing it from handling tasks outside its intended business domain.
- Question 35Multiple Select
Which two Agentic Actions execute business logic rather than simply generate text for an Agentforce agent?
- AFlow Builder flowsCorrect answer
- BApex classesCorrect answer
- CPrompt Templates
- DKnowledge article layouts
Why this is the answer
Flows and Apex classes execute business logic, while Prompt Templates only generate text and do not perform system changes.
- Question 36Interactive
An administrator configures the access settings shown in the code block below. Which key Agentforce concept is being configured?
- AData GroundingCorrect answer
- BHuman-in-the-Loop
- CEinstein Trust Layer
- DReasoning Engine
Why this is the answer
Data Grounding connects the agent to specific Salesforce objects, fields, and Knowledge articles for accurate responses.
- Question 37Single Choice
What is the main function of the Einstein Trust Layer in an Agentforce deployment for a Sales organization?
- AMask sensitive PII before sending to LLM.Correct answer
- BGenerate response text automatically.
- CDefine the agent's business domain.
- DCreate Apex triggers for the agent.
Why this is the answer
The Einstein Trust Layer masks sensitive personally identifiable information before data is sent to the external LLM.
Service and support applications
4 questions in this domain
- Question 38Single Choice
Which configuration element in an Agentforce agent defines the specific persona and operational scope of the agent?
- AAgent Role & PurposeCorrect answer
- BActions & Tooling
- CData Grounding
- DEinstein Trust Layer Integration
Why this is the answer
Agent Role & Purpose is the configuration element that defines the agent's persona and scope. This keeps the agent focused on specific business domains and prevents tasks outside its intended boundaries.
- Question 39Multiple Select
Which two statements correctly describe the role of Actions and Tooling when an administrator configures an Agentforce agent?
- AThey execute specific capabilities such as updating Cases or querying Knowledge.Correct answer
- BThey are the reasoning layer that interprets user intent.
- CThey are the same as Prompt Templates in the agent runtime.
- DThey are configured to call Flows or Apex classes.Correct answer
Why this is the answer
Actions execute logic, not text generation. They are configured by mapping natural language instructions to Apex classes or Flows and provide capabilities like updating a Case or querying Knowledge. Prompt Templates are not Actions.
- Question 40Interactive
Examine the JSON snippet from an Agentforce agent configuration. Based on the value shown for the action type, which statement correctly describes the capability being added?
- AIt calls a Flow to update a Case status.Correct answer
- BIt uses a Prompt Template to generate a response instead of executing logic.
- CIt prevents the agent from updating any record.
- DIt masks sensitive data before LLM processing.
Why this is the answer
The JSON action type is 'Flow' with a case status update path. Therefore the agent capability is implemented by invoking a Flow rather than by Prompt Template, restrictions, or Trust Layer.
- Question 41Single Choice
What is the main purpose of the Einstein Trust Layer Integration when configuring an Agentforce agent?
- AIt defines the agent's persona and business scope.
- BIt masks sensitive PII before data is sent to the LLM and applies data sharing rules.Correct answer
- CIt provides the instructions used by the reasoning engine.
- DIt connects the agent to Knowledge articles for context.
Why this is the answer
Einstein Trust Layer Integration handles privacy and security by masking PII before sending data to the LLM and by enforcing the organization's data sharing rules.
Frequently asked questions
Everything about the Salesforce Certified Administrator question bank and how to use it.