Deloitte
Q1. Deloitte Australia
What was the main problem with Deloitte’s AI-assisted report?
A. It took too long to complete
B. It contained incorrect references and a fabricated quotation
C. It was too expensive to print
B. It contained incorrect references and a fabricated quotation
Reveal: Deloitte's case illustrates why completing work does not guarantee that the result is reliable.
Q3. Apple Intelligence
What went wrong with Apple’s AI-generated news summaries?
A. Some summaries contained false information
B. The summaries were too long
C. News notifications stopped appearing
A. Some summaries contained false information
Reveal: Apple's case illustrates "Ignoring the Unhappy Outcome" because the AI workflow did not adequately handle inaccurate or misleading summaries.
Q5. Wayfair
Why did Wayfair use AI to improve product tags?
A. To replace all warehouse employees
B. To help customers find suitable products more easily
C. To generate more advertising content
B. To help customers find suitable products more easily
Reveal: The purpose was linked to customer product discovery rather than simply increasing AI activity.
Q2. Deloitte Australia
What outcome should the client have valued most?
A. Producing reports as quickly as possible
B. Using AI in more consulting tasks
C. Receiving an accurate and reliable report
C. Receiving an accurate and reliable report
Reveal: The client needed information it could trust, not simply a faster report.
Q4. Apple Intelligence
What should Apple prioritize when measuring success?
A. The number of summaries generated
B. The speed of generating summaries only
C. Accurate information and user trust
C. Accurate information and user trust
Reveal: Quick reading is useful, but inaccurate summaries can undermine the outcome users actually value.
Q7. Overall understanding : Activity vs. Outcome
Which of the following best measures the success of an AI project?
A. Number of tasks automated
B. Amount of time saved
C. Improvement in customer outcomes
C. Improvement in customer outcomes
Reveal: Automation and time savings are useful operational measures, but they do not establish customer value by themselves.
Q6. Wayfair
Which metric would best show whether customers benefited?
A. Number of product tags processed
B. Number of AI tools installed
C. Percentage of customers who successfully find relevant products
C. Percentage of customers who successfully find relevant products
Reveal: Processing tags measures activity; successful product discovery is closer to the customer outcome. It is a proposed metric, not a reported Wayfair result.
Q8. Guardrail metrics
If an AI project's primary KPI is the speed of generating news summaries, which guardrail metric should the team track?
A. Number of summaries generated
B. Percentage of summaries containing factual errors
C. Average summary length
B. Percentage of summaries containing factual errors
Reveal: Factual error rate is a proposed guardrail metric that checks whether faster summaries are compromising accuracy. This is not a reported Apple metric.