5 Tips To Create The PERFECT AI Prompt (10x Your Results!)

If you’ve spent five frustrating minutes trying to get an AI to write a simple email, only to receive five pages of philosophical musings, you’re not alone.

The world’s best AI tools—including our own SalesApe AI Agents—are powerful, but they operate on a fundamental principle: Garbage In, Garbage Out. If your prompt is vague, your output will be vague, wasteful, or just plain useless.

To get 10x the results from any Large Language Model (LLM) like ChatGPT, Gemini or Perplexity, you need to think less like a user and more like a director.

Here is our 5-Step Formula for crafting the perfect prompt that turns vague ideas into powerful, actionable output.

The Formula: R-G-C-F-C

To build the perfect prompt, you must clearly define five components for the AI. Think of it as Role, Goal, Context, Format, and Constraint.

Step 1: The Role (R) – Who Are You?

The most common mistake people make is treating the AI like a search engine. You must treat it like an actor, assigning it a specific, powerful identity. This immediately elevates the quality of the response by dictating style, tone, and knowledge base.

Why this works: Defining the role forces the AI to activate its training on that specific persona, using industry jargon, appropriate tone, and relevant examples.

Step 2: The Goal (G) – What Do You Need?

Be ruthless about clarity. If you ask for a summary, you'll get a summary. If you ask for a summary that compels an immediate action, you get a result that delivers business value.

The Goal must be a single, measurable outcome.

• Vague Goal: "Write some ideas for a blog post."
• Clear Goal: "Generate five click-bait titles and three one-paragraph outlines for a blog post targeting CTOs that addresses AI adoption anxiety."

The more constraints you put on the goal, the better the final output will be.

Step 3: The Context (C) – Why and How?

The AI needs the necessary background to make its output relevant to your specific situation. Don't assume the AI knows your business or audience (even if you’ve already shared this info in a previous conversation).

You must provide the "Who, What, and Why" that dictates the content.

• Include: Target audience demographics, your company's unique value proposition (UVP), competitive information, or the specific problem you are trying to solve.
• Example Context: "The email must focus on how our product reduces churn by 15%. Our main competitor is Company X, so do not mention cost savings, but focus purely on efficiency gains."

This step is essential for making generic AI output sound like it came from your specialized team.

Step 4: The Format (F) – Structure is King

AI models love structure. If you leave the format open-ended, you risk getting a massive wall of text. By defining the format, you make the output instantly usable.

Always specify the structure you want:

• "Respond in a Markdown table with four columns."
• "The response must be a three-paragraph email (no more, no less)."
• "Output must be a Python script with inline comments for each function."
• "Use five bullet points starting with action verbs."

Bonus Power Move: You can even use this to establish tone, e.g., "Write the entire response in the style of a 1940s detective."

Step 5: The Constraint (C) – The Secret Sauce

The Constraint is what forces the AI to work harder and more creatively, often eliminating obvious or bland responses. This step is about adding a final layer of difficulty.

Use Constraints to:

• Avoid Clichés: "Do not use the words 'synergy,' 'paradigm,' or 'innovative.'"
• Limit Length: "The word count must be between 100 and 120 words."
• Enforce Style: "Use a casual, conversational tone, and end with a strong, open-ended question."

By imposing this final constraint, you ensure the AI focuses on quality over quantity and forces it to think outside the box it might have been trained in.

Put it All Together

Here is a full, powerful prompt using all five steps:

"[R] Act as a Senior Product Manager at a B2B SaaS company that sells workflow automation software. [G] Write a compelling, 300-word product description for a landing page that focuses on driving sign-ups for a free trial. [C] Our target audience is mid-level HR managers who are burnt out on administrative tasks. The main pain point is onboarding new staff. [F] Use three clear headlines followed by short paragraph descriptions. [C] Do not use any technical jargon, and ensure the tone is empathetic but action-oriented."

Follow this R-G-C-F-C structure, and you'll go from wasting time to 10x-ing your AI output in every single conversation. It's the same rigorous process we use to train our SalesApe Agents to perfection.

Think Robot, Not Human

Whilst running through your R-G-C-F-C formula, remember you’re talking to a machine. You’re not going to hurt its feelings, it’s not going to report you to HR if you ask it to redraft a piece of work, it’s not going to think you’re rude because you didn’t ask how its kid’s little league team is doing. 

Your human copy writer might not appreciate being micromanaged but your LLM likes as much detail as possible. 

❓ Frequently Asked Questions (FAQs)

1. Does this 5-step formula work for all AI models (text, image, and code)?

The R-G-C-F-C (Role, Goal, Context, Format, Constraint) framework is universal for all generative AI.

• For text-based LLMs, it is exactly as described.
• For image generators (like Midjourney or DALL-E), the "Role" might become the style (e.g., "Act as a 19th-century watercolorist") and the "Format" becomes the ratio/style (e.g., "Output as a 16:9 cinematic shot").
• For code generation, the "Role" is the language (e.g., "You are a Python expert") and the "Goal" is the function/script you need.

The key is that you are always providing specific constraints to drive higher quality.

2. How long should a good prompt be? Is shorter or longer better?

Length doesn't matter; density does. A short prompt with high information density (e.g., "Act as a CTO and write a one-sentence, punchy UVP for our product") is better than a long, rambling one. A longer prompt is necessary when you need to provide extensive Context (C) and specific Constraints (C). Always aim for the minimum number of words required to clearly and precisely convey the R-G-C-F-C information.

3. What is "Zero-Shot" vs. "Few-Shot" prompting, and how does it relate to the formula?

This is where prompt engineering gets technical!

• Zero-Shot Prompting is the most common—you give the AI the prompt (using the R-G-C-F-C formula) and expect an immediate, complete answer.
• Few-Shot Prompting relates to adding Context (C). It involves giving the AI a few examples of the desired input/output before asking it to perform your task. For instance, you provide three examples of a perfectly written sales email, and then ask the AI to write a fourth one. This drastically improves quality, as the AI has a much clearer, immediate pattern to follow.

‍