AI Agents

AGENTFORCE-SPEC · 70 questions

• After go-live, the Companion Care Group service agent answers "can I move my appointment to Thursday" with a friendly but generic explanation of the rescheduling policy and never actually changes the booking. The Appointment Management topic is being selected correctly. What is the most likely cause?
• A pet owner types "my dog is due for his rabies booster, can I get him in this week" into the Companion Care Group service agent. Several topics are configured. What does the reasoning engine evaluate first to decide how to respond?
• Two custom topics exist for the group's service agent: "Lab Result Follow-Up" and "Appointment Management." Owners asking "my dog's bloodwork came back, when should I bring him in" are routed to Appointment Management about half the time. What should the specialist change first?
• An Agentforce Specialist at a veterinary hospital group is documenting a custom topic for prescription refills so a colleague can maintain it. Which statement correctly separates the purpose of the topic's classification description from its scope?
• An Agentforce Specialist notices that one topic contains a single instruction of eleven sentences describing everything the agent should do for an insurance claim enquiry. Responses are inconsistent. What is the recommended remedy?
• A specialist writes this line into a topic: "Never provide a diagnosis, and never tell an owner whether a symptom is an emergency; hand the conversation to a licensed staff member instead." How is this line best characterised in Agentforce configuration terms?
• Which statement best describes what the reasoning engine does after a topic has been selected for a pet owner's request to check a claim's status?
• A pet owner asks the service agent "do you sell dog food subscriptions?" No configured topic covers retail products. What should a well-configured agent do?
• The Companion Care Group refill topic must check that the patient has been seen within the last six months, confirm the medication is not on the controlled-substance hold list, and only then create the refill request. Which action backing type best fits this requirement?
• A staff-facing agent should produce a short plain-language summary of a patient's last three visits for the front-desk team. Which action backing type is the natural fit?
• The group wants its staff-facing agent to answer "what is our post-operative discharge policy for orthopaedic patients" from approved documents rather than from the model's own knowledge. Which configuration concept addresses this requirement?
• A specialist must choose between a standard topic and a custom topic for handling routine appointment enquiries in a Service agent. What is the strongest reason to start from the standard topic?
• A topic for insurance claims has fourteen agent actions assigned, and the agent frequently invokes the wrong one. Apart from rewriting action descriptions, what structural change is most likely to improve action selection?
• The group wants the agent to hand a frustrated pet owner to a human representative when the owner explicitly asks for a person. What is the appropriate way to configure this?
• An agent action calls a Flow that requires an Appointment Id and a new date. The pet owner has said "move Bella's Thursday appointment to next Tuesday." What does the reasoning engine do about the required inputs?
• Which statement accurately describes standard agent actions in Agentforce?
• An Apex invocable action is being considered as the backing for an agent action. Which characterisation matches how the Agentforce Specialist exam expects you to treat it?
• A specialist adds an action to a topic but the agent never uses it, even when a test case clearly calls for it. The action works when tested on its own. What should be examined first?
• An agent action updates a prescription record. What is the recommended way to reduce the risk that the agent performs the update on the wrong prescription?
• Which scenario is the strongest case for backing an agent action with an external API or External Service rather than with a Flow?
• During a multi-turn conversation, an owner says "show me Bella's last visit," and two turns later asks "and her next appointment?" without repeating the patient's name. For the agent to resolve "her" correctly, what mechanism is holding the identified patient across turns?
• An agent action that lists available appointment slots keeps surfacing openings at hospitals across the whole region instead of the owner's home clinic, confusing pet owners who only want their local branch. Which configuration change most directly fixes this?
• The prescription-refill topic must never approve a refill if the patient has not been seen by a vet within the group's required window. A specialist stores that window as a reusable variable and references it in a filter on the refill action. Why is this preferable to writing the number into the topic's instructions as plain text?
• A specialist wants the appointment-booking topic to refuse any request to book a rabies-booster slot more than 30 days out, every time, with no exceptions. Rather than adding an instruction telling the model to check the date, what should the specialist configure?
• A new specialist assumes that when the service agent updates a Case on behalf of a pet owner, the update is recorded as having been made by that pet owner's Community/Experience user. Why is this assumption wrong?
• An agent action queries the Prescription object, but the Agent User's permission set was never granted field-level access to the "Controlled Substance" field. What is the most likely result when a pet owner asks about a refill that depends on that field?
• Refills for a controlled-substance medication must never be auto-approved by the agent, even if every other refill criterion is met. A specialist wants this to hold no matter how the instructions in the refill topic are later edited by a colleague. What is the most robust way to guarantee it?
• A specialist is choosing between two ways to make sure the claims-status action only ever returns claims belonging to the account making the request: a filter on the action bound to the requesting account, or an instruction telling the model "only discuss claims for this owner." Which is the better choice, and why?
• A referral coordinator reports that the agent correctly has object and field permissions for Case, yet still cannot retrieve a specific Case during a conversation. The Case is owned by a queue the Agent User was never added to, and no sharing rule grants broader visibility. What explains the missing record?
• A specialist is reviewing two separate configurations: one masks an owner's home address before it reaches the underlying language model, and another determines whether the agent's action can query the Claim object at all. Which pairing correctly matches each configuration to its layer?
• A specialist notices the service agent can casually mention another owner's outstanding invoice balance when discussing an unrelated topic, because its Agent User was set up by cloning a System Administrator profile "to save time." What is the correct remediation?
• A staff-facing Employee agent is instructed to "always summarise the vet's internal notes when asked," but a custom field called Internal Vet Notes is intentionally excluded from the Agent User's field-level security. When a technician asks for those notes, what should happen, and why is that correct?
• The group wants an agent that pet owners use on the website and mobile app to check appointment times, ask about a refill, and follow up on a claim. Which agent type fits this requirement?
• A business-development representative is pursuing a discounted-services contract with a working-dog training partner and wants an agent to help draft outreach, track the opportunity, and summarise deal status. Which agent type matches this work?
• Veterinary technicians want an internal Slack-based assistant that answers "what is the post-operative discharge policy for orthopaedic patients" and similar staff-only policy questions, drawing on internal knowledge articles. Which agent type is the right fit, and why does the audience matter here?
• A veterinary group is defining who may modify the Agent User's permission set going forward. Why should this be restricted to a small, admin-level group rather than opened up to any specialist who wants to add access for convenience?
• A specialist argues that any agent type can be deployed to any channel, so channel alone should never be used to decide between Employee, Service, and Sales agent. What is the better basis for that decision?
• The group wants pet owners to reach the Service agent directly from the branded pet-owner portal they already log into to view records and pay invoices. Which channel connection fits this requirement?
• The group is building two separate scheduling assistants: one for front-desk staff to manage the day's appointment board internally, and one for pet owners to request their own appointments. Why should these be two different agents rather than one agent serving both audiences?
• A specialist configuring a brand-new Sales agent notices it arrives with different out-of-the-box standard topics and actions than a brand-new Service agent does, even before any custom configuration is added. What explains this?
• Veterinary technicians want to ask an internal assistant quick clinical-policy questions without leaving the messaging app they already use all day for team communication. Which channel and agent-type pairing fits?
• Pet owners frequently email the clinic's general inbox asking about invoice details or requesting a refill, and staff want the Service agent to read incoming messages, draft accurate replies grounded in the owner's records, and route anything sensitive to a human. Which channel configuration matches this requirement?
• Owners report that the agent gives noticeably different, sometimes conflicting, safety disclaimers depending on whether they message through email or the mobile app, even though it is meant to be the same agent. Investigation shows a colleague added separate guardrail instructions inside two different topic variants, one used only for email and one used only for the app. What is the correct fix?
• The Service agent is connected to both the pet-owner portal and email. A specialist wants to know whether the topics and actions available to the agent should be reconfigured separately for each channel. What is the correct understanding?
• The refill topic's "Get Upcoming Appointments" action sometimes returns appointments booked at a sister clinic across town instead of the clinic the owner actually called. Which change forces the action to always scope to the correct clinic, rather than trusting the reasoning engine to pick right?
• For controlled-substance refill requests, a specialist is deciding between two designs: (A) a topic instruction that says "never approve a refill for a controlled substance without staff review," or (B) a filter that excludes controlled-substance medications from the refill action's eligible record set unless a staff-override flag is set. Which design actually guarantees the restriction holds?
• A technician in Slack asks the Employee agent to pull up a colleague's HR-adjacent internal note, and the technician's own Slack identity happens to have broad channel visibility in Slack itself. Does the technician's Slack-level access expand what the agent's underlying action is allowed to retrieve from Salesforce?
• An owner tells the service agent "this is about Bella, my golden retriever" in the first message. Three turns later, without repeating Bella's name, the owner asks "were her lab results back yet?" What configuration lets the agent carry the identified patient forward reliably instead of re-inferring it from context each turn?
• The group wants an internal tool, available only to authenticated staff inside Salesforce, that lets a receptionist ask "which exam rooms are free at the Riverside clinic in the next hour" and get an answer drawn from internal scheduling data. Which agent type fits this requirement?
• A Service agent tells an owner "I could not find any lab results for Bella," even though the front-desk staff can see three released results for that patient in the same org. Testing confirms the action and topic are configured correctly. What is the most likely root cause?
• To avoid repeated troubleshooting tickets, a specialist grants the Agent User a permission set with "View All" on every clinical object across all six regional hospitals. What is the main risk this introduces?
• A new specialist asks what actually determines which Salesforce records an Agentforce agent can see and act on at run time, separate from whatever the pet owner or staff member chatting with it can normally access. What is the correct answer?
• The group's growth team wants an agent that engages website visitors who fill out a "join our wellness plan" form, asks qualifying questions, and creates or updates a Lead record for a human account manager to follow up on. Which agent type is designed for this job?
• Pet owners reach the group through the website, the mobile app, and a Slack channel staff monitor, asking things like "is my dog's incision supposed to look like this" or "can you check my claim status." These interactions are logged and tracked as Cases. Which agent type is the natural fit for this front-facing work?
• The group is expanding into a new region and wants an agent that proactively engages people who requested information about a pet-insurance partnership, tracked as Leads, to determine fit before a rep calls. A Service agent is already deployed for existing-client case handling. Should the group extend the existing Service agent to cover this, or build separately?
• Vets at the group want to ask "what was Bella's last recorded weight" from Slack during rounds, without opening Salesforce. A specialist connects the internal agent to a Slack channel used only by clinical staff. What is an important consideration specific to deploying an agent into Slack?
• The group's shared refill-request inbox is connected to a topic that drafts a reply confirming eligible refills and flags ineligible ones for a pharmacist. Compared to the live web-chat channel, what most distinguishes how this agent behaves over the email channel?
• A specialist copies a topic built for the group's Service agent into a new Employee agent, expecting it to work identically. Some of the standard actions referenced by the topic are missing from the new agent. What best explains this?
• Pet owners log into the group's self-service portal, built on Experience Cloud, to chat with the agent about their own pets only. What governs whether an owner can only see their own pet's records through that portal-embedded agent, rather than any pet in the org?
• A guardrail correctly triggers escalation when an owner describes a possible emergency, and on the web widget this hands the conversation to a live human agent. The same guardrail fires on the email channel, where there is no live agent to transfer to. What should the specialist configure for this case?
• An owner's Contact record has a "Preferred Contact Language" field set to Spanish. The group wants every agent response for that owner rendered in Spanish without asking the model to re-check the field on every single turn. What is the deterministic way to configure this?
• Two topics in the same agent, one for routine appointment booking and one for insurance-claim status, both consistently respond in the same warm, plain-language tone associated with the group's brand, even though a different specialist wrote each topic. What is the most likely reason the tone is consistent across both?
• A specialist is defining a new custom topic and must decide what belongs in the topic's scope statement versus the agent's company description. Which pairing correctly separates the two?
• A pathology system posts bloodwork results into Salesforce with a status of "Pending Vet Review" before a vet signs off, then "Released" afterward. The group never wants the agent to disclose a pending result to an owner, even if the model reasons the owner is asking urgently. What is the most reliable configuration to enforce this?
• The refill topic must let a staff-flagged "chronic condition" pet skip the six-month check-in requirement, while every other pet is still blocked from refilling without one. The specialist wants this enforced without adding another instruction sentence for the model to interpret. What combination achieves this?
• The group sets Patient object-wide default sharing to Private so each clinic sees only its own patients. An Employee agent used by an operations manager who oversees all six clinics needs to query patients across every location. Granting the Agent User object-level "Read" permission alone is not enough. What else is required?
• Owners asking "can you move my Thursday visit" are sometimes routed to the "Lab Result Follow-Up" topic instead of "Appointment Management." Both topics' classification descriptions mention "visit" and "next steps." What is the best fix?
• A topic's instructions currently say only "answer the owner's question about lab results." Owners have occasionally been given results after only loosely confirming which pet they mean. What should the specialist add to the instructions to fix the sequencing, without relying on a guardrail or a filter?
• The group wants clinic managers to ask, from a private Slack channel restricted to managers, "how many open exam rooms does Riverside have this afternoon," pulling from an internal scheduling object. Which combination of agent type and channel is the correct fit?
• At a high level, once a Service agent has been built and tested with its topics and actions, what is the correct general process for making it available to owners through a new channel such as the group's mobile app, without pinning the answer to a specific menu path or button label?