一个提示词公式（角色＋任务＋背景＋格式＋约束）＋ 5 套即拿即用模板 ＋ 3 张空白卡——把一句模糊的话补完整，再通过几轮对话拿到一个能检查的小成果。
家家鸽 Jagger
万象 AI 联合创始人｜碧丽斯传媒创始人｜AI 技能咨询师
Forget magic spells. Learn one formula that makes your request complete, then let AI work the task out step by step. This lesson takes one real task and walks you through saying what you want clearly, then reaching a checkable small result over a few rounds of dialogue.
| 1 formula | 5 templates | 3 blank cards | 1 small result |
|---|---|---|---|
| Role + task + context + format + constraints | From defining the result to checking and correcting — copy and use | Fill in before, during, and before delivery | The one you ran yourself, and can check |
Three-part delivery promise: hands-on practice + prompt template pack + homework.
You write a long paragraph of instructions, and AI still comes back long, hollow, and missing the part you actually needed.
The problem usually is not that you wrote too little — it is that the request was incomplete: what role AI should take, what exactly to do, what it needs to know, what the output should look like, and what it must not do. Today, no pile of fancy spells. First learn one formula that makes the request complete, then reach a checkable small result through a few rounds of dialogue.
Four key concepts in plain language (result, prompt, formula, multi-turn dialogue); telling apart five easily-confused things (goal, function, tool, step, acceptance); an overview of 8 common work scenarios; the soul of the lesson — the formula role + task + context + format + constraints, broken down part by part with examples.
The five-step order for multi-turn dialogue and the six questions that make one task clear; a main case — turning a long course transcript into reviewable study material — with prompts A (state the result), B (define this round), and C (take the result, check it, decide the next round), all in copy-ready full text.
5 ready-to-copy templates (task brief / this round / meeting minutes / long material / task breakdown); 3 blank cards you can print and fill in; 3 general red lines + 5 lesson-specific ones; appendix on choosing tools.
3 lessons + 1 assignment, covering the full path from concepts to three rounds of hands-on work to reusable templates.
Four key concepts in plain language; telling apart five easily-confused things (goal, function, tool, step, acceptance); 8 common work scenarios; the soul of the lesson — role + task + context + format + constraints, broken down part by part with examples; and an advanced trick: ask AI to ask you three questions before starting.
The easiest five-step order for multi-turn dialogue; the six questions that make one task clear; five things that stay human decisions even as AI gets stronger; a main case — turning a long course transcript into reviewable study material — with prompts A / B / C in full text, plus the checkpoints to watch during the demo.
5 templates (one-page task brief / defining this round / meeting minutes / long material / task breakdown); 3 blank cards (before starting, in progress, before delivery); 3 general red lines + 5 lesson-specific ones; a closing line and an appendix on choosing tools.
In one sentence: AI drafts and fills the gaps; you judge and decide.
Treat AI as a very smart junior colleague — you are the lead, and your job is to brief the work well and make the call.
Remember: the formula is a skeleton that helps you think clearly, not a spell that produces good results the moment you recite it. What really decides quality is whether you have thought through what must be delivered, who will use it, and what counts as usable.
