After this lesson you will be able to take any complex ask and break it into 3–6 ordered steps the AI can execute one at a time, with handoff conditions you'd notice if they failed.
Big prompts produce shallow thinking.
Show me the concept in text instead
When you hand the AI a monolithic task ("plan our Q3 campaign"), it compresses six distinct subtasks into bullet points and you get something plausible but shallow. The cognitive move is to see the *seams* in the task before you prompt — what are the natural breakpoints, what depends on what, and where would you yourself stop to check the output before continuing. A decomposed task gives the AI room to think hard about one thing at a time, and gives you room to verify each step before the next builds on it.
When you hand the AI a monolithic task ("plan our Q3 campaign"), it compresses six distinct subtasks into bullet points and you get something plausible but shallow. The cognitive move is to see the *seams* in the task before you prompt — what are the natural breakpoints, what depends on what, and where would you yourself stop to check the output before continuing. A decomposed task gives the AI room to think hard about one thing at a time, and gives you room to verify each step before the next builds on it.
Common mistakes
- Treating decomposition as "ask the AI to break it down" — that's recursion, not decomposition
- Breaking down trivial tasks that fit in one prompt (overhead exceeds benefit)
- Skipping dependency analysis — running step 3 before step 2 has been verified
Big prompts produce shallow thinking.
Break it into 3–6 ordered steps. For each: what's the input, output, and how you'd verify…
3–6 discrete steps, each with named output
Sequence reflects real dependencies
Dependencies explicit and verifiable
Parallel steps marked, sequential ones justified
Whole sequence has a defined "done" state
Now you try. Rewrite the vague version using this capability, then I will score the attempt against the lesson rubric.
Type your rewrite below. Hints are free when you need them.