SF AF Spe -- SET 4 -- 02/2026

Universal Containers has created an Employee Agent. Which step should an Agentforce Specialist take to connect the agent with a Slack channel?. Create a connection between Salesforce and the Slack workspace. Create an embedded service deployment and connection between Salesforce and the Slack workspace. Create an Omni-Channel flow and connection between Salesforce and the Slack workspace.

Universal Containers (UC) wants to build an Agentforce Service Agent that provides the latest, active, and relevant policy and compliance information to customers. The agent must: • Semantically search HR policies, compliance guidelines, and company procedures. • Ensure responses are grounded on published knowledge. • Allow Knowledge updates to be reflected immediately without manual reconfiguration. What should UC do to ensure the agent retrieves the right information?. Enable the agent to search all internal records and past customer inquiries. Manually add policy responses into the AI model to prevent hallucinations. Set up an Agentforce Data Library to store and index policy documents for AI retrieval.

Cloud Kicks wants to integrate its agent with its custom website. The goal is for customers to interact with the custom agent chat interface. Which approach provides the framework for the custom web application to communicate with the agent?. Agent API. Model Context Protocol (MCP). Agent to Agent (A2A).

Universal Containers has PDF maintenance guides in an external folder, not yet in Salesforce. The team wants a standard, clicks- only setup for the Service Agent to use these documents. Which approach should the Agentforce Specialist implement?. Paste external PDF links into topic instructions and rely on the model to follow them, avoiding configuration of a retrieval source, index, or retriever action. Upload the PDFs as File source in the Agentforce Data Library which will build a Search Index, and create a retriever to ground responses from those documents. Configure Data Cloud to ingest file attachments and create custom index and retriever for product record and attachment dat.

Coral Cloud Resorts is about to start testing its concierge agent with guests. Which metrics should be captured to monitor the performance, correctness, and user experience?. Agent performance, token usage, and conversation duration. Response performance, tone, and CSATS. Response times, accuracy and relevance of answers, and resolution success.

Cloud Kicks (CK) recently finished the development of a new prompt template that uses its own large language model (LLM). CK is deploying a prompt template from a sandbox to a production org, and is receiving an error When trying to deploy the change set, CK is getting an error related to the LLM used in the prompt template. What is the cause of the error?. The prompt does not specify that it is a custom LLM. The name of the LLM does not match in sandbox and production. BYOLLM is not yet supported for in prompt templates in production.

Universal Containers (UC) wants to ensure its compliance team can retrieve exact matches of policy clause numbers from a structured legal document library. Which search type should UC implement?. Use hybrid search to blend keyword and semantic recall. Use keyword search for exact term matching on structured fields like clause numbers. Use semantic search to interpret synonyms of clauses dynamically.

A Service Agent at Universal Containers (UC) is designed to help customers resolve issues by searching against knowledge articles. Knowledge articles have PDF attachments that add critical details. UC reports that the agent provides excellent summaries of the knowledge articles, but seems completely unaware of the PDF attachments. How should an Agentforce Specialist configure the Data Cloud search index to include the content of these attached files?. Increase article chunk size and token limits for Knowledge indexing so larger contexts capture attachment references. Enable 'Include Related Attachments' for Knowledge kav and map the ContentDocumentLink unstructured data model object (UDMO). Use Data Cloud's 'Include Attachments' option and select the ContentDocumentVersion unstructured data model object (UDMO).

A service manager wants to use Salesforce Prompt Builder to help agents summarize customer case notes after a support call. The summary should: • Capture the customer's issue, troubleshooting steps taken, and next actions. • Be no longer than five sentences. • Use plain language (no technical jargon) • If no next action is identified, the summary should explicitly state "No next action required." Which prompt template follows Salesforce prompt design best practices? . Role: You are an experienced support agent. Task: Summarize the case notes. Context: Include customer issue, troubleshooting steps, and next actions. Constraints: Limit to 5 sentences, use plain language, and if no next action is found, state "No next action required." Format: Use numbered sentences for clarity. Role: You are a support agent writing a case summary. Task: Provide a professional summary of the issue and troubleshooting steps. Context: Include customer issue, steps taken, and next actions if available. Constraints: No strict sentence limit, but use plain language. If no next action is found, leave it out. Format:Use paragraphs for readability. Role: You are a case documentation assistant. Task: Write a summary of the support call. Context: Always describe the customer issue, troubleshooting, and resolutiondetails. Constraints: The summary should be comprehensive and professional, but there is no limit on length or language style. Format: Use complete sentences in a narrative style.

Universal Containers plans to enable Agentforce in Slack so teams can interact with agents directly in Slack channels. Which description represents the key steps required to enable Agentforce in Slack?. Enable the default Slack channel Agentforce, and assign Slack agent access to users. Configure the Slack workflow to invoke the Agentforce API, enabling users to interact with agents through predefined triggers and automated steps. Configure the Slack agent connection and, in Manage Agentforce, install the agent, then assign agent access to users.

Coral Cloud Resorts (CCR) uses Agentforce to assist customers with booking and service issues. CCR wants to implement a triage process so that: • High severity requests must be escalated to a human service rep. • Lower severity requests should result in creating a support case for the guest. • The requirement is to achieve the highest reliability and determinism in the response from the agent. Which approach should an Agentforce Specialist recommend?. Write the triage and routing logic in Topic Instructions using an IF, THEN, ELSE pattern: "Escalate to human service rep if the request is considered severe, otherwise create support case". Use absolute keywords like "Always" and "Never" in Topic Instructions to enforce logic, such as "Always escalate when severity is high" and "Never create a support case when severity is high". Create a custom variable severityLevel populated by a Triage action. Add filters so the "Escalate to human service rep" action only runs when severityLevel = 'High', and the "Create Support Case" action runs only when severityLevel != 'High'.

Coral Cloud Resorts wants visibility into credit usage associated with testing. Which feature supports this?. Agentforce Analytics. Digital Wallet. Testing Center.

An Agentforce Specialist builds a new Service Agent that uses a custom action built on a flow. The agent has been tested in a sandbox and is now ready to deploy; What is a key consideration regarding the activation status of the agent in the production environment?. The agent will be activated automatically only if the flow is also active. The agent must be manually activated in production, regardless of its status in the sandbox. The agent will automatically be activated upon successful deployment.

Universal Containers has set up a Service Agent to allow customers to look up their order status. The topic setup includes: Name: Order Inquiry Classification Description: Handles user requests to look up order status, including tracking details and delivery estimates for orders placed within the last 90 days. Scope: Your job is only to assist authenticated users in looking up the status of their orders placed within the last 90 days. If the order is pending delivery, provide the tracking number and estimated delivery date. Do not handle inquiries for orders older than 90 days. Which information will be used by the Agentforce reasoning engine to choose this topic?. Topic Name and Classification Description. Topic Name and Scope. Classification Description and Scope.

The Agentforce Specialist for Coral Cloud Resorts wants to create an agent that will automate the resolution of a large portion of guest complaints related to their vacation experiences. The agent will be able to offer upgrades, hotel credit, and other complimentary options. The agent will also be in charge of escalating the case to a human when a guest has suffered a major disruption (such as cancellation). Following Salesforce best practices, which type of agent should the Agentforce Specialist create?. Sales Agent with a Flex prompt template. Custom Agent with a Flex prompt template. Service Agent with a Flex prompt template.

What is a key benefit of the Agent-to-Agent (A2A) protocol?. Provides a standardized framework for cross-vendor agent discovery and communication. Allows auto-onboard third-party agents without additional contracts, trust scores, or shared identity controls. Provides a standardized runtime engine for internal agent discovery and communication.

Universal Containers is indexing millions of product manuals where users may ask both structured queries (model numbers) and natural language Questions (for example, "How do I reset my device?"). Which retrieval approach should the company use?. Use keyword search only, since model numbers dominate queries. Use semantic search only, as natural language is always preferred. Use hybrid search to combine keyword precision with semantic flexibility.

When using a prompt template, what should an Agentforce Specialist consider with their grounding data and chosen model?. Review the token limit in the Einstein Trust Layer. Ensure queries used for grounding employ offset so the token limits of models are not exceeded. Review the model limitation in Prompt Builder versus the grounding data size.

Universal Containers wants to test agents while preserving real data and isolating from production. Which environment should the company use with Testing Center?. Use personal developer orgs unrepresentative of production dat. Use production org directly with test assertions. Use sandbox environments replicated from production for safe testing.

Coral Cloud Resorts wants to handle frequent customer misspellings of package names in queries. Which approach should the Agentforce Specialist implement?. Hybrid search. Vector search. Keyword search.

Universal Containers (UC) has configured a data library and wants to restrict indexing of knowledge articles to articles which are. only publicly available in their knowledge base. UC also wants the agent to link sources that the large language model (LLM) grounded its response on. Which settings should help UC with this?. In the data library setting window, under Knowledge Settings, enable Use Public Knowledge Article and select Show sources. In the data library setting window, under Knowledge Settings, enable Use Public Knowledge Article. It is not possible to display articles that the LLM grounded its response in. Use Data Categories to categorize publicly available articles to index. Sources are automatically displayed when knowledge articles are categorized as Publi.

Universal Containers (UC) has a library of custom-built personalized investment portfolio APIs, and is planning to extend it to agents. Which method should UC's agent choose to dynamically use the best API service?. Agent-to-Agent (A2A) protocol support. Model Context Protocol (MCP) server support. MuleSoft connector for custom hosted processes.

Coral Cloud Resorts (CCR) wants to configure its agent so that booking actions are only available when a customer's membership tier is "Premium" or "Elite". This business rule must be enforced deterministically. What should CCR implement?. Set up custom validation rules on the underlying booking objects to prevent non-eligible customers from completing bookings. Configure topic instructions that clearly state booking actions should only be used for Premium or Elite customers and include examples. Create a context variable mapped to the customer's membership tier field, then add a conditional filter on MembershipTier.

Universal Containers is building a digital shopping assistant that needs to dynamically generate product recommendations using information from the company's external product recommendation predictive model through APIs. Which Agentforce capability should make it easier for the agent to consume the external product recommendation tool?. Model Context Protocol (MCP). Hugging Face. Agent-to-Agent (A2A) protocol.

Coral Cloud Resorts needs to ensure its booking agent executes actions in a specific sequence: First retrieve available sessions, then verify customer eligibility, and finally create the booking. The current implementation allows the large language model (LLM) to execute these actions in any order, causing booking failures. Which approach should an Agentforce Specialist implement?. Create custom variables that store completion status for each step, then implement conditional filters on subsequent actions requiring previous variables to be populated, ensuring deterministic execution order. Configure topic, classification description, and action instructions with priority levels and sequence indicators to guide the reasoning engine in selecting the correct action order automatically. Write comprehensive topic instructions detailing the exact sequence of actions using numbered steps and explicit ordering requirements for the reasoning engine to follow during booking workflows.

An administrator at Universal Containers has successfully deployed a new agent from a sandbox to production using a change set. The agent uses a prompt template that invokes a Salesforce flow to perform a complex calculation. In production, when users interact with the agent, it fails with an error message every time the flow is supposed to run. The flow was included in the change set and is present in production. What is the most likely cause of this issue?. The flow was not manually activated in the production org after the deployment. The user in production does not have permission to run the flow. The change set did not include the dependent Apex classes for the flow.

Coral Cloud Resorts wants to cover a broad range of user phrasing when testing its FAQ agent. Which Testing Center feature meets that need?. AI-generated synthetic test utterances based on natural language variations. Uploading only a small set of manually written prompts. Relying on live customer logs to capture phrasing diversity after deployment.

Coral Cloud Resorts is uploading thousands of new HTML knowledge articles files for a resort launch. To ensure Agentforce retrieves accurate responses quickly, which chunking strategy should be used when creating a new index?. Semantic-based passage extraction. Conversation-based chunking. Section-aware chunking.

Coral Cloud Resorts (CCR) sees the agent forgot the dietary/activity preferences gathered earlier. They need those preferences to persist throughout the session. What should CCR implement?. Configure custom variables to capture/store customer preferences from action outputs. Rely on natural conversation memory and instruct the agent to look back. Create a context variable to capture/store customer preferences as action outputs.

When creating a custom retriever in Einstein Studio, which step is considered essential?. Select the search index, specify the associated data model object (DMO) and data space, and optionally define filters to narrow search results. Define the output configuration by specifying the maximum number of results to return, and map the output fields that will ground the prompt. Configure the search index, choose vector or hybrid search, choose the fields for filtering, the data space and model, then define the ranking method.