
Four questions reveal any task’s automation potential:
- Can we define “good” objectively?
- Can we create test cases with known correct answers?
- Is the task self-contained or context-dependent?
- How long is the feedback loop?
High scores = automation candidate. Low scores = remains human.









