Question
How do I practice delegation decision framework?
Quick Answer
List every recurring task you performed this week. For each task, score three dimensions on a 1-5 scale: (1) Irreversibility — how costly is it to fix a poor outcome? (2) Identity-centrality — does this task define who you are or develop a skill only you should develop? (3) Cognitive uniqueness — does this task require judgment, context, or creativity that only you possess?
Multiply the three scores to get a Delegation Resistance Score (DRS) ranging from 1 to 125. Tasks scoring below 30 are strong delegation candidates. Tasks scoring 30-60 are candidates for partial delegation or automation. Tasks scoring above 60 should stay with you. Identify your three lowest-scoring tasks and draft a delegation plan for each: who or what system receives it, what specification they need, and what verification you will use.
Common pitfall: Treating delegation decisions as binary — either you do everything or you hand off everything. The framework collapses when people skip the scoring and rely on gut feel, which is biased toward keeping tasks that feel comfortable and delegating tasks that feel unfamiliar, regardless of strategic value. Another common failure: scoring tasks once and never re-scoring. Your comparative advantage shifts as you grow, as your team develops capability, and as tools improve. A task that scored 90 on identity-centrality six months ago may score 40 today because you have already mastered it and your growth edge has moved elsewhere.
This practice connects to Delegation Patterns — building it as a repeatable habit compounds over time.
Learn more in these lessons
- The delegation decision framework
Use clear criteria to decide what to delegate, what to automate, and what to keep.
Go beyond the answer
This topic is part of Delegate to AI Without Losing Judgment — a guided sequence of 16 lessons.
After completing this path you will run a working delegation boundary: a criteria framework for what goes to AI versus what stays with your judgment, specifications clear enough that AI output comes back usable, a verification practice that catches errors without redoing the work, and the warning signs of over-delegation before your judgment atrophies.
Start the path →