SLF 1 - 4

Sunrays Limited (SL) is scaling its Agentforce deployment and needs to securely connect its AI agents to a growing number of external enterprise data systems and local developer environments. Instead of building custom integration logic and bespoke Application Programming Interfaces (APIs) for each new data source, the Agentforce Specialist recommends leveraging the Model Context Protocol (MCP). What is the primary purpose of using an open standard like MCP in this scenario?. To standardize the secure connection and delivery of context between the AI models and various local or remote data sources. To replace the need for Retrieval-Augmented Generation (RAG) by storing all external data natively within the large language model (LLM)'s weights. To allow the agent to autonomously negotiate task delegation with third-party supply chain agents.

Sunrays Limited (SL) wants its Agentforce Service Agent to retrieve up to date factual data from the internet. Which step must an Agentforce Specialist take to ensure the agent uses the web search capability properly?. Enable the Answer Questions with Knowledge standard action specifically on the General FAQ subagent (formerly known as a topic). Map the web search retriever to a custom search index in Data 360 before activating the agent. Remove the default General FAQ subagent (formerly known as a topic) from the agent and replace its function with a new General Web Search topic.

The compliance team at Sunrays Limited (SL) has determined that a specific customer service prompt must only process data through a secure, self-hosted Amazon Bedrock model and must be prevented from accessing the default OpenAI models. How should the Agentforce Specialist configure this specific model routing?. Assign a permission set to the dedicated Agent User that restricts Read access to the default Salesforce large language model (LLM) metadata records. Configure a Data 360 Private Connect endpoint that automatically reroutes all Agentforce requests away from the default large language model (LLM) Gateway. Register the model as a BYOLLM in Einstein Studio, then explicitly select this model when configuring the prompt template.

An Agentforce Specialist is testing an agent with Testing Center. The test results show that an agent correctly identifies the right subagent to handle an utterance, but it fails to select all necessary actions within that subagent. Which evaluation metric specifically identifies this failure?. Action Assertion. Response Evaluation. Subagent Assertion.

Sunrays Limited (SL)' agent must always look up the customer's account tier and open cases from Salesforce before deciding how to respond. Based on Agent Script flow of control, what is true about executing deterministic actions at the very start of a subagent (formerly known as a topic)?. Actions can only be guaranteed to run by placing them in the config block. Only before_reasoning can guarantee the large language model (LLM) is invoked before an action runs. The first instruction in reasoning.instructions always runs before the large language model (LLM) is invoked.

Sunrays Limited (SL) is configuring its Agentforce Testing Center to evaluate an agent handling customer complaints. The company wants to assess if the agent successfully demonstrates empathy and follows a proprietary "De-escalation Framework" before offering a resolution. Where should the Agentforce Specialist utilize a large language model (LLM)-as-judge to achieve an appropriate evaluation?. Adjust the fixed criteria of the standard Coherence quality evaluation to control the LLM-as-judge evaluation. Enable the default Instruction Adherence evaluation which natively uses LLM-as judge. Create a custom evaluation with a tailored prompt outlining the framework's criteria.

A company's support agent is inconsistently executing a mandatory fraud check action before processing refunds. In some conversations the fraud-check fires, but in others the agent skips directly to issuing the refund. Upon review, an Agentforce Specialist recommends placing the instruction guideline with 'run @actions.fraud check' followed by 'run @actions.process_refund'. What is the effect of this configuration change?. The fraud-check action will be more strongly suggested to the large language model (LLM), but it may still be skipped if the LLM determines it is not relevant to the conversation. The fraud-check action will execute after the large language model (LLM) check if it missed it in the process step execution, but the agent will lose its ability to use an empathetic tone. The fraud-check action will be forced to execute before the refund action in every conversation, because procedural instructions provide deterministic control over the execution order.

An Agentforce Specialist is evaluating an agent conversation. What is the reason why intents and session metrics within Observability are important to consider?. To evaluate the agent's overall performance. To evaluate why the agent was unable to answer a user's question. To evaluate the deflection rate of the agent.

Which business requirement best presents a good use case for using Prompt Builder?. Forecast future sales trends based on historical data. Send a quote proposal in reply to a request for a quote. Identify potential high-value leads based on lead score.

Sunrays Limited (SL) has successfully chunked and vectorized its unstructured meeting notes into a Data 360 search index. An Agentforce Specialist needs to connect this data to a prompt template to dynamically refine search criteria and retrieve the most relevant information. How should the specialist achieve this?. Create a Flex template referencing a data lake object (DLO). Create a Flex template referencing a retriever in the prompt template. Create a Flex template referencing the data model object (DMO).

Sunrays Limited (SL) is implementing an Agentforce Service Agent to assist customers. The agent must be able to retrieve information from proprietary policy documents stored as PDFs and ensure its responses are grounded exclusively in this approved company data, rather than generic large language model (LLM) knowledge. The project team requires the fastest and least complex setup possible, minimizing manual configuration of backend data components. Which approach should the Agentforce Specialist take to satisfy these requirements?. Upload the PDFs to Salesforce Files and configure the agent's Topic Instructions to dynamically read the files at run time. Manually uploaded the PDFs into Data 360 as Unstructured Data Model Object (UDMO) type and select create a new Agentforce Data Library mapping option. Create an Agentforce Data Library (ADL) and upload the PDF policy documents directly to the library.

Sunrays Limited (SL) is building an Agentforce Service Agent to handle order cancellations. The Agentforce Specialist needs to ensure that a critical "Check Cancellation Eligibility" action is executed deterministically on every relevant turn, without relying on the reasoning engine's discretion to choose the tool. During a code review, a junior developer asks why the Agentforce Specialist called the action using the run @actions.name command instead of simply listing the action under the reasoning.actions: block. What must the Agentforce Specialist explain regarding the difference between these two invocation methods?. Both patterns execute the action on every turn automatically; the difference is purely syntactic, as the run command is simply the newer notation for Agent Script. The run command is only valid when nested inside reasoning.actions: blocks to pass parameters; the Agentforce Specialist must list it under reasoning.actions: for the large language model (LLM) to access it. Listing an action under reasoning.actions: makes it a subjective tool that the large language model (LLM) decides whether to call, calling it with the run command forces guaranteed execution every time.

Sunrays Limited (SL) uses an agent to handle customer service inquiries. SL recently partnered with an external logistics provider that operates its own distinct, autonomous AI agent. When a customer requests a complex international shipping reroute, the SL agent needs to securely communicate, negotiate routing terms, and delegate the execution of the reroute directly to the logistics provider's AI agent. Which open standard multi-agent protocol is specifically designed to facilitate this autonomous task delegation and negotiation between independent AI agents?. Agent to Agent (A2A) Protocol. Model Context Protocol (MCP). OpenAPI Specification (OAS).

Global Finance Corp (GFC) is expanding its Agentforce rollout from a basic customer service agent to a suite of specialized agents handling Fraud Detection, Loan Origination, and Billing. GFC operates entirely within a single, global Salesforce instance. The CIO wants to ensure that as the number of specialized agents scales, the company maintains strict, centralized control over security guardrails and user context, ensuring customers do not have to repeat themselves when their request spans multiple departments. What is a reasonable architectural approach to achieve this level of scalability and control?. Implement a Multi Org, Multi Agent (MOMA) architecture connected via the Agent to Agent (A2A) protocol to safely isolate each department's agent. Deploy a Single Org, Multi Agent (SOMA) architecture using a primary Orchestrator agent to manage shared context natively and route sub-tasks to the specialized agents. Use the Model Context Protocol (MCP) to federate multiple external, third-party agents directly into the existing Agentforce Service console.

An administrator at Sunrays Limited (SL) has successfully deployed a new Service Agent from a sandbox to production using a change set. The Service Agent uses a prompt template that invokes a Salesforce flow to perform a complex calculation. In production, when users interact with the Service 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 user in production does not have permission to run the flow. The change set did not include the dependent Apex classes for the flow. The flow was not manually activated in the production org after the deployment.

What is the correct process to leverage Prompt Builder in Salesforce Org?. Select the appropriate prompt template type to use, develop the prompt within the prompt workspace, select resources to dynamically insert CRM-derived grounding data, pick the model to use, and test and validate the generated responses. Enable the target object for generative prompting, develop the prompt within the prompt workspace, select records to fine tune and ground the response, enable the Trust Layer, and associate the prompt to an action. Select the appropriate prompt template type to use, select a Salesforce standard prompt, determine the object to associate the prompt, select a record to validate against, and associate the prompt to an action.