Free Salesforce Certified Agentforce Specialist Practice Questions

Sample Practice Questions

Try these 10 sample questions from the Salesforce Certified Agentforce Specialist practice bank. Each includes the answer and explanation. Continue with the interactive practice above to review the full question set.

1. 1A team needs reusable instructions that combine CRM data with an LLM to generate personalized content. Which Salesforce capability addresses this requirement?

   1. A. A validation rule evaluated when records are saved
   2. B. A report formula evaluated across grouped records
   3. C. A prompt template maintained in Prompt Builder
   4. D. A duplicate rule comparing incoming records

   Answer: C

   Explanation: Salesforce states, "With Prompt Builder, you create, test, revise, customize, and manage prompt templates that can make use of your CRM data." This supports reusable, CRM-grounded content generation; the other capabilities evaluate rules or report calculations.

   Reference: Get to Know Prompt Builder — Meet Prompt Builder

2. 2Which platform capability must be set up to use Agentforce Data Libraries?

   1. A. Marketing Cloud Engagement
   2. B. Data 360
   3. C. Tableau Server
   4. D. Heroku Connect

   Answer: B

   Explanation: Salesforce states, "To use Agentforce Data Libraries, you must have Data 360 set up." The other named platforms do not replace that prerequisite.

   Reference: Setting Up Data Libraries — Setting Up Data Libraries; prerequisites

3. 3Which Agent Script block contains global instructions and required welcome and error messages?

   1. A. config
   2. B. variables
   3. C. system
   4. D. language

   Answer: C

   Explanation: The block reference assigns "Global instructions and required messages (welcome, error)" to system. Configuration, variable declarations, and locales have separate blocks.

   Reference: Read the Language Behind the Canvas — The File and Its Blocks; system row

4. 4Why should an administrator run action-executing Agentforce Testing Center tests in a sandbox?

   1. A. Because actions remain read-only in testing but become writable in production
   2. B. Because Testing Center automatically rolls back each action’s data changes
   3. C. Because agent tests can execute actions that modify CRM data
   4. D. Because action writes are simulated and never reach the connected CRM

   Answer: C

   Explanation: Trailhead warns, "Testing agents can modify CRM data." Testing does not make actions intrinsically read-only, automatically roll back writes, or guarantee simulation. Running in a sandbox isolates these possible changes from production data.

   Reference: Set Up Testing Criteria — Run Your Tests; Important

5. 5Where does Agentforce Session Tracing store detailed agent-session data for monitoring?

   1. A. Only in Salesforce debug-log files for Apex execution
   2. B. In Data 360 as detailed agent-session records
   3. C. Only in CRM field-history records for changed fields
   4. D. Only in Event Monitoring log files instead of Data 360

   Answer: B

   Explanation: Salesforce states, "Turn on Agentforce Session Tracing to collect and store, in Data 360, detailed data about agent sessions." Its detailed session records are stored in Data 360, rather than being replaced by Apex debug logs, field-history records or Event Monitoring files.

   Reference: Set Up Agentforce Session Tracing — Set Up Agentforce Session Tracing; purpose

6. 6Several business domains need independently replaceable and testable agent capabilities. Which architectural choice directly supports this modularity requirement?

   1. A. Place all domain behavior in one agent that must be replaced as a unit
   2. B. Create domain-specific agents but require them to share one inseparable release
   3. C. Use specialized agents with independently replaceable and testable lifecycles
   4. D. Create separate front ends while keeping domain behavior in one coupled agent

   Answer: C

   Explanation: Salesforce states, "Individual agents can be added, replaced, modified, and tested with greater ease". Independent specialized agent lifecycles satisfy the stated modularity requirement. A single coupled agent, mandatory joint release or front-end-only separation does not permit the requested independent domain changes.

   Reference: Enterprise Agentic Architecture and Design Patterns — Multi-agent architecture benefits; Modularity & Extensibility

7. 7Sales representatives need AI-written outreach that reflects each customer's relationship history. What role can Prompt Builder play?

   1. A. Generate personalized email drafts from customer context
   2. B. Generate email drafts without resolving the recipient’s CRM context
   3. C. Generate email drafts using one customer’s details for all recipients
   4. D. Generate email drafts by inserting field names instead of stored values

   Answer: A

   Explanation: The email use case asks an LLM to write emails "while using your customer relationship history and insights". Recipient-specific CRM context supports personalization; omitting it, reusing another customer’s details, or inserting labels instead of values fails that requirement.

   Reference: Get to Know Prompt Builder — Email Generation

8. 8Which permissions does Salesforce specify for setting up Agentforce Data Libraries?

   1. A. Data Cloud admin and System Administrator permissions
   2. B. Prompt Template User and Run Flows permissions
   3. C. Einstein Sales Emails and Manage Flow permissions
   4. D. Knowledge User and View Reports permissions

   Answer: A

   Explanation: The permissions table specifies "Data Cloud admin AND System Administrator permissions." Access to other AI, flow, or reporting features does not satisfy this setup requirement.

   Reference: Setting Up Data Libraries — User Permissions Needed

9. 9Where does Agent Script define the agent's developer name, agent user, and agent type?

   1. A. The config block
   2. B. The language block
   3. C. The variables block
   4. D. The system message text

   Answer: A

   Explanation: Trailhead assigns "The agent's identity, such as developer name, agent user, and type" to config. Locale settings, state declarations, and global message text do not define this identity.

   Reference: Read the Language Behind the Canvas — The File and Its Blocks; config row

10. 10A suite tests only correctly formatted customer requests. Which addition best follows Salesforce's guidance for more robust coverage?

    1. A. Add negative, unexpected, and invalid requests
    2. B. Duplicate each successful request without changing its conditions
    3. C. Add more ordinary valid requests along the same successful path
    4. D. Vary customer data while keeping every request valid and expected

    Answer: A

    Explanation: Trailhead directs, "Then test for negative, unexpected, or even invalid results." These exercise boundaries and failure handling. Duplicated or additional valid happy-path cases do not add the missing negative/unexpected coverage.

    Reference: Set Up Testing Criteria — What to Test; positive and negative testing