The Prompt Cheat Sheet

Must Include · Nice to Include · Expert-Level

Print this. Tape it near your keyboard. In a month you won’t need it.

MUST INCLUDE — every prompt, every time

1. The task. Say exactly what you want produced. “Write a two-paragraph email” beats “help me with an email.”

2. The context. Who it’s for, what it’s about, what’s already happened. The AI knows nothing about your situation until you tell it. The single biggest beginner mistake is under-explaining.

3. The format. What should the output look like? An email, a list, a table, a schedule, 200 words, one page. If you don’t say, you get the AI’s guess.

Weak: “Help me with a fundraiser email.” Strong: “Write a short, warm email to parents of my son’s Cub Scout pack announcing our popcorn fundraiser. Kickoff is Sept 12, orders due Oct 3. Include a friendly ask for two parent volunteers. Keep it under 150 words.”

NICE TO INCLUDE — when the result matters

Tone. Friendly, formal, brief, encouraging. One word changes everything.

Audience details. “These are parents who are already stretched thin” produces a different email than no audience at all.

An example. Paste something you liked (“match the tone of this”) and the AI will mirror it.

The raw material. Don’t describe the email thread — paste it. Don’t summarize your notes — paste them. The AI works far better with the real thing.

Constraints. What to avoid, length limits, reading level, “don’t use exclamation points,” “no emojis.”

EXPERT-LEVEL — the habits that separate confident users

Ask-me-first. End your prompt with: “Before you answer, ask me any questions you need to make this as good as possible.” The AI will surface the details you forgot you knew. This one habit improves results more than any other trick on this page.

Show-your-work. Add: “After your answer, briefly explain your approach and any assumptions you made.” You’ll see where it guessed, catch wrong assumptions early, and steadily build a feel for how these tools think. (Honest note: this is the AI describing its approach, not a literal window into its circuitry — but it’s exactly the view you need to steer it.)

Meta-prompting. Paste your prompt and say: “Improve this prompt first, show me the improved version, then answer it.” You’re using the AI to get better at using the AI.

Iterate like an editor. The first output is a draft. Give specific feedback: “warmer,” “cut the second paragraph,” “make the ask clearer.” Two rounds of feedback beat one perfect prompt.

Ask for options. “Give me two versions — one brief and businesslike, one warm and personal.” Choosing between options is easier than describing what you want from scratch.

Self-check. For anything factual or high-stakes: “Before finalizing, review your answer for errors or anything I should verify.” Then verify the things it flags — and the things it doesn’t.

Save what works. When a prompt nails it, save it. A collection of proven prompts — and eventually reusable “skills” your AI applies automatically — is how you stop starting from scratch. (Chapter 10 shows you how.)

The one rule that overrides everything: you are the editor. The AI drafts; you judge, verify, and decide. That’s what being AI Native means.

Part of “Practical Understanding and Use of AI/LLMs” — developed with AI assistance and reviewed by a human editor. May contain minor errors.