Question 160
An organization delivering with a predictive approach discovers: A research pilot's AI prompts and data preparation choices are not reproducible by the receiving delivery team. Which observation provides the strongest evidence that the underlying problem was resolved?
Explanation
Correct Answer
“The receiving team reruns the test using documented inputs and obtains explainable comparable results.” is meaningful evidence for “A research pilot's AI prompts and data preparation choices are not reproducible by the receiving delivery team” because it demonstrates the practical outcome of “Capture datasets, assumptions, safeguards, and reproducible experiments in a controlled knowledge repository.” instead of merely recording that work took place.
Incorrect Answers
“Internal maintainers independently configure a test integration from the handed-over material.” could measure performance for “A vendor completes integration work but keeps interface configuration knowledge in personal notebooks.”, yet it is not direct evidence that “A research pilot's AI prompts and data preparation choices are not reproducible by the receiving delivery team” has been resolved through “Capture datasets, assumptions, safeguards, and reproducible experiments in a controlled knowledge repository.”.
“New team members can explain why key architecture tradeoffs were made without reconstructing chats.” could measure performance for “A distributed agile team stores architectural decisions only in transient chat threads.”, yet it is not direct evidence that “A research pilot's AI prompts and data preparation choices are not reproducible by the receiving delivery team” has been resolved through “Capture datasets, assumptions, safeguards, and reproducible experiments in a controlled knowledge repository.”.
“Support responders successfully complete representative recovery drills without project-team intervention.” could measure performance for “An operations handover is signed despite no evidence that support staff can recover the service after failure.”, yet it is not direct evidence that “A research pilot's AI prompts and data preparation choices are not reproducible by the receiving delivery team” has been resolved through “Capture datasets, assumptions, safeguards, and reproducible experiments in a controlled knowledge repository.”.