Meridian's ops team tells the PM they want to "use AI for maintenance." What should the PM do first, during Business Understanding, to turn this into an actionable AI-specific business question?

Formulating AI-specific business questions is the first enabler of the CPMAI Business Understanding phase, and it means converting a vague aspiration into a specific, measurable, answerable question that data and a model could actually address. "Can we predict which component is likely to fail within the next 30 flight hours using sensor data?" is concrete enough to guide dataset identification, define success metrics, and eventually support a Go/No-Go decision. Jumping straight to selecting a specific machine-learning algorithm for prototyping is a technical implementation choice made before the problem itself is defined, which risks building the wrong thing efficiently. Purchasing a maintenance-management software license is a procurement action that doesn't address the underlying scoping problem at all. Hiring a data scientist before any goal has been defined repeats a common project-management error: staffing a specialized role before there is a defined problem for them to solve wastes their time and often leads to scope drift. The discipline of translating a vague stakeholder request into a specific, testable business question is what distinguishes a well-run AI Business Understanding phase from one that jumps straight to technology.