Each task is scored on two independent axes. The combined scores place it into one of four quadrants, each suggesting a different AI deployment posture.
X axis
Problem structure
How well-defined are the inputs, methods, constraints, and evaluation criteria?
Y axis
Outcome convergence
Would independent engineers produce the same result from the same inputs?
Scoring dimensions
Problem structure (X)
1.
Input completeness — Are all required inputs available?
2.
Problem clarity — Can the task proceed without interpretation?
3.
Procedural availability — Does a clear procedure exist?
4.
Evaluation criteria — Can correctness be determined?
5.
Tacit knowledge — Can it be done without undocumented expertise?
Outcome convergence (Y)
1.
Solution uniqueness — One correct answer or many?
2.
Independent reproducibility — Same result from different engineers?
3.
Assumption sensitivity — Stable outcome under varied assumptions?
4.
Validation autonomy — Verifiable without expert judgment?
5.
Reasoning transparency — Can the logic be made explicit?
