I spent the last week testing every “AI hack” on the internet. Honestly, most of them are a waste of time. But then I found this specific framework, and it changed everything for me. I used to get generic, boring answers from ChatGPT. Now? It feels like I have a team of experts working for me.
If you’ve ever felt frustrated because AI just “doesn’t get it,” this guide is for you. We’re going to break down a simple AI prompt formula for beginners called RICECO. It stands for Role, Instruction, Context, Examples, Constraints, and Output.
Trust me, once you start using this, you’ll never go back to one-sentence prompts again.
Table of Contents
The Secret Shortcut: The ICC Method

If you are in a rush, you don’t always need all six steps. I found that using just the “Big Three” gets you 80% of the way there. I call this the ICC method: Instruction, Context, and Constraints.
Step 1: Give the AI a “Role” (The R in RICECO)
Assigning a role is like flipping a switch. It tells the AI which “brain” to use. I tested this by asking for sleep advice. When I told the AI it was a “board-certified doctor,” the answer was medical and professional. When I said it was a “sleep-deprived parent,” it became relatable and encouraging.
- How to do it: Start your prompt with “You are a [Job Title]…”
- Why it works: It changes the tone and depth of the answer instantly.
Step 2: Write Specific Instructions (The I in RICECO)
This is the core task. The biggest mistake I see? Being too vague. If you ask for something “engaging,” the AI guesses what that means. Usually, it guesses wrong.
- Don’t say: “Write a cool YouTube script.”
- Do say: “Write a 60-second script using a curiosity gap hook.”
Note: I made the mistake of being vague for months. Specificity is your best friend. Tell the AI exactly what job it needs to do.
Step 3: Provide the Context (The first C in RICECO)
Context is the background story. It’s the difference between a generic blog post and one that actually helps your readers.
- Who is it for? (e.g., small business owners).
- What is the goal? (e.g., explain the value of a tool).
- What is the platform? (e.g., LinkedIn or a personal blog).
Step 4: Show, Don’t Just Tell (The E in RICECO)
This is a game-changer. If you want the AI to write like you, give it a sample of your writing. I call this the “cheat sheet” method.
- Action: Paste a previous newsletter or article and say, “Use this as a reference for tone.”
- Result: The AI imitates your style instead of sounding like a robot.
Step 5: Set Your Boundaries (The second C in RICECO)
Constraints are the rules. They stop the AI from falling into bad habits, like using too many emojis or being too wordy.
- Avoid buzzwords: Tell it to avoid words like “innovative” or “cutting-edge.”
- Set limits: “Keep it under 100 words.”
- Style rules: “Don’t use hashtags” or “Use a warm tone.”
Warning: I found that if you don’t set constraints, the AI tends to get very “corporate” and boring. Always tell it what not to do!
Step 6: Choose Your Output Format (The O in RICECO)
Finally, tell the AI how to present the information. You don’t have to settle for a wall of text.
- Try a Table: “Present this as a three-column table.”
- Try Markdown: “Use H2 headers and bullet points.”
- Try Data: “Format this as a JSON code snippet.”

What Usually Goes Wrong
Even with a RICECO prompt framework guide, things can get messy. Here is what I’ve noticed:
- The Context Dump: If you give too much background, the main instruction might get lost. Keep the core task clear.
- Trusting the First Draft: I never take the first answer. I always ask the AI, “What did you miss?” or “Critique your own work.”
- Being Too Polite: You don’t need to say “please.” Just be direct and clear.
Summary
Using the RICECO formula isn’t about being a “tech person.” It’s just about being a better communicator.
- Role: Who is the AI?
- Instruction: What is the task?
- Context: What is the background?
- Examples: Can you show it a sample?
- Constraints: What are the rules?
- Output: How should it look?
Have you tried using a framework like this before? It totally changed my workflow, and I’d love to hear if it works for you too. Let me know in the comments!
FAQs
Do I really need to use all six parts of RICECO every time?
No. I use the full formula for big, important tasks like writing a long blog post or a complex email. For quick, simple questions, just stick to the ICC shortcut (Instruction, Context, Constraints). It saves time while still giving you a much better answer than a basic prompt.
How do I provide “Examples” without making the prompt too long?
I found a great trick for this! Instead of pasting a 2,000-word article, just paste 2 or 3 paragraphs that perfectly capture your “voice.” You can also tell the AI, “Analyze the tone of the following text, then use that tone for the rest of this conversation.” This keeps your prompt clean and focused.
What should I do if the AI ignores my “Constraints”?
This happens sometimes, and it’s annoying! If the AI ignores a rule (like “Don’t use emojis”), I usually just reply with: “Note: You included emojis. Please rewrite the last response without them.” Usually, the AI “wakes up” and follows the rule on the second try.
Can I use RICECO with free AI tools like the basic ChatGPT or Claude?
Yes! I’ve tested this across almost every major tool. Because RICECO is based on how Large Language Models (LLMs) think, it works on both free and paid versions. However, the paid versions (like ChatGPT Plus) are generally better at following strict Constraints.
What is the most important part of the RICECO formula?
If I had to pick just one, it’s Context. Without context, the AI is just guessing. When you tell the AI who the audience is and why you are writing, the quality of the output jumps instantly.





