Prompting is briefing, not guessing

Back to: Prompt Like a Pro

Replace one-shot prompt tricks with a conversation that has a clear outcome, evidence, and review step.

The idea

A professional brief reduces avoidable ambiguity. It tells a collaborator what result matters, what material is authoritative, what boundaries apply, and how the work will be judged. Prompting works the same way. There is no universal phrase that guarantees quality across tasks, models, and situations.

Think in three phases: brief, draft, and review. The brief defines the job. The model produces a draft or set of options. Your review compares that output with explicit criteria. A follow-up then targets the largest gap instead of asking vaguely for something “better.”

This approach keeps you in control. The assistant can contribute speed and variation, while you provide the purpose, source material, priorities, and final judgment.

Try it now

1. Choose a familiar task and write the first short prompt that comes to mind.
2. Before running it, write three criteria that would make the answer genuinely useful.
3. Run the prompt. Score the answer against those criteria from 1 to 5.
4. Write one follow-up that names the largest gap, includes an example, and preserves what already works.

BRIEF: Produce [outcome] for [audience/use].\nSOURCE OF TRUTH: Use only [material].\nSUCCESS CRITERIA:\n1. [criterion]\n2. [criterion]\n3. [criterion]\n\nREVIEW: Score the draft against each criterion. Identify the largest gap before revising.

Quality check

• The outcome is observable, not merely a topic.
• My criteria describe quality before I see the draft.
• The follow-up targets one gap instead of restarting blindly.