Most people start using AI like ChatGPT or Claude by treating it like a magical mind-reader. They type: "Write me an email to a client," and then get annoyed when the result feels robotic, generic, and full of fake corporate fluff.
Here is the truth: An AI is not a human brain. It is the world's most powerful autocomplete machine.
If you give it vague instructions, it gives you a vague answer. But when you give it clear boundaries and specific details, it produces work that looks like it took an expert hours to craft.
The 4-Step Prompt Formula (RTCF)
Whenever you write a prompt for work, school, or business, make sure it has these 4 simple ingredients:
- 1. Role (Who is talking?): Give the AI a perspective. For example: "You are a friendly customer service lead" or "You are a clear technical editor." (Avoid empty buzzwords like "world's greatest genius"—be practical).
- 2. Task (What must it do?): State exactly one clear action using verbs like: "Summarize," "Rewrite," "Calculate," or "Extract."
- 3. Context (What information is it using?): Provide the source text, email, or background data wrapped in quotation marks or tags so the AI knows what to read.
- 4. Format (How should the answer look?): Specify the shape of the output: "3 bullet points under 20 words each," "A 2-column table," or "A clean email with a subject line."
The Law of Clear Execution
When you provide specific constraints, the AI doesn't need to guess your intent. You get high-signal, production-grade output on the very first try.