Problem decomposition, root-cause analysis, assumptions, alternatives, risk analysis, scenario thinking and no-action comparison.
The Problem-Solving Prompt System
Twenty reasoning patterns for turning vague questions into structured, verifiable decisions — rebuilt for modern AI instead of “magic prompt” folklore.
Useful foundation. Outdated packaging.
The original infographic contains many sound analytical moves, but it treats them as twenty separate “must-have prompts.” A better 2026 approach is modular: specify the objective, context, constraints, verification standard and output — then invoke only the reasoning patterns the problem actually needs.
“Act as an expert,” “creative solutions,” and “fastest path” need criteria, evidence discipline, trade-offs and uncertainty handling.
The original largely omits fact / inference separation, source checking, confidence calibration and explicit triggers for changing the recommendation.
20 reasoning patterns
Search, filter, expand and copy only what you need. The point is not to use all twenty every time.
Build one serious prompt.
Use a single structured brief instead of chaining twenty isolated instructions.
Complete the fields above, then tap “Generate prompt.”
From prompt collection to decision system
The verification discipline
For consequential work, instruct the model to distinguish supplied facts, established facts, inference, assumptions and speculation. Ask what evidence would materially change the answer. This is more valuable than adding theatrical phrases such as “act as the world’s best expert.”
Recommendation → strongest reasons → trade-offs → confidence → unresolved uncertainties → immediate next action → conditions that would change the recommendation.
Better prompts are not longer.
They are better specified.
Use the patterns as modular reasoning tools, not as ritual. The strongest prompt is the shortest one that captures the objective, evidence, constraints, decision criteria and required output.
Visit ImageFirm.ai