1. Why can an AI feature not be specified the way conventional software can?
2. What three things replace the specification in an AI product?
3. In an AI product, what is the status of behaviour at the edges — when the system is unsure or the input is out of scope?
4. Which of these must come from a system of record rather than a generative model? Select all that apply.
5. An assistant states a balance and a suggestion in one sentence, in the same voice. What is the problem?
6. What is the detectability test for an AI product surface?
7. Which responsibilities are new to the AI product manager relative to conventional product management? Select all that apply.
8. Two models score within noise of each other. On what grounds should the choice be made?
9. A medical triage classifier would be worthless or harmful below a quality bar. What kind of product is this?
10. What does a manual prototype establish that building the system does not?