Review questions

Review questions

Test your understanding with these questions. Marking and saved results are available with a free account.

1. Question 1A team reports 92% overall accuracy for an application triage aid but does not separate errors on urgent applications. What is the strongest step before allowing it to delay reviews?

   Approve it because overall accuracy exceeds 90%. Measure urgent misses against urgent applications and assess delay harm. Only confirm that the same version processed every application. Rename the output advice rather than decision without changing workflow.
2. Question 2The vendor holds urgent-case measurements, while the product owner authorizes use of the triage aid inside the organization. Which allocation is best for release?

   The vendor alone is accountable for use impact because it holds test data. The end user owns release because they see the output. The vendor supplies test evidence; the product owner owns acceptance after relevant review. Everyone at the meeting is jointly accountable with no named decision owner.
3. Question 3Which metric card best tests control of missed urgent applications?

   Urgent misses with numerator, denominator, period, owner and breach action. Only the monthly number of tool users. A fairness commitment in the policy. Aggregate accuracy without error distribution or response.
4. Question 4A complaint summary becomes automatic closure without a model change. Which gate is required?

   No gate because weights are unchanged. Review the new purpose and decision impact, appeal and oversight. Only rename the feature on the product page. Only measure old summarization speed.
5. Question 5A team is buying a tool to rank urgent complaints; the supplier offers only overall accuracy. What evidence is most relevant first?

   The vendor's worldwide customer count. The vendor's responsible-AI slogan. Urgent misses on suitable cases and evidence of review capability. Summary generation speed without testing ranking.
6. Question 6An appeal dashboard counts only requests received. A reviewer finds the appeal button missing in the Arabic interface. What is the first assurance judgment?

   The channel works because every recorded request closed. Praise closure speed without testing access. Suspend the effectiveness claim and test channel access and the eligible-decision denominator. Delete Arabic requests from the report.

Create a free account to mark your answers and save your progress.