How to know if your AI use case is actually worth building.

The truth is, most AI projects are already headed for failure before anyone has even started.

Because no one stopped to ask: should we even be doing this?

Before I scope any AI project, I run four questions past the team:

  1. Is there a real business problem here?
    Not "AI could probably help with this." A specific problem that costs you time or money, and keeps coming back.

  2. Do you have the data to support it?
    AI needs clean, accessible, sufficient input. Most teams haven't checked this before committing to a project.