Question 90
At Yardley Homes, a property warranty contact center is improving warranty service before expanding the internal rollout (SF-A24-01). The administrator observes that an agent would close a high-severity case directly without the required human approval. Any fix must ensure that service changes follow approved warranty conditions. Which observable result would best demonstrate that the correction actually worked?
Explanation
Correct Answer
This is the direct acceptance signal for the reported failure: In a controlled Case acceptance test, a high-severity case remains open until required approval is recorded; the record is retained under Yardley Homes's rollout (SF-A24-01). The chosen correction was to require an approved authorization and confirmation step before the privileged action; the measurement therefore tests the relevant source, routing or authorization outcome rather than unrelated activity.
Incorrect Answers
The proposed observation “A cross-account record access test fails under the agent's runtime security context; confirm this using a representative sandbox fixture” would check a different original case: A service agent action runs with a privileged user context that grants excessive record visibility. The correction must ensure that plant access boundaries are not weakened by chatbot replies. Its mechanism is: The user identity executing an action determines which permissions can apply. In the present case, an agent would close a high-severity case directly without the required human approval. The correction must ensure that service changes follow approved warranty conditions; the more direct acceptance evidence is “A high-severity case remains open until required approval is recorded; confirm this and compare with approved baseline behavior” because sensitive agent operations need deterministic approval gates beyond conversational requests.
The proposed observation “The agent declines or safely escalates an unauthorized object request; confirm this and confirm the resulting trace” would check a different original case: An action attempts to read a protected custom object based only on a user's natural-language request. The correction must ensure that member records cannot be exposed to other members. Its mechanism is: An agent conversation is not an authorization bypass for CRM data. In the present case, an agent would close a high-severity case directly without the required human approval. The correction must ensure that service changes follow approved warranty conditions; the more direct acceptance evidence is “A high-severity case remains open until required approval is recorded; confirm this and compare with approved baseline behavior” because sensitive agent operations need deterministic approval gates beyond conversational requests.
The proposed observation “Two consecutive credit requests produce at most one approved credit action; confirm this and compare with approved baseline behavior” would check a different original case: An agent issues a customer credit twice during one support session after a repeated request. The correction must ensure that safety-critical changes need an attributable human approval. Its mechanism is: A mutable state flag can enforce a rule independent of LLM wording. In the present case, an agent would close a high-severity case directly without the required human approval. The correction must ensure that service changes follow approved warranty conditions; the more direct acceptance evidence is “A high-severity case remains open until required approval is recorded; confirm this and compare with approved baseline behavior” because sensitive agent operations need deterministic approval gates beyond conversational requests.