Introduction
Creating content with AI is no longer a futuristic fantasy—it’s today’s superpower. But here’s the kicker: your results are only as good as the prompt you give. Garbage in, garbage out.
You can have the most powerful AI engine in the world, but if your prompt is vague, flat, or unstructured, you’ll get mediocre results. On the flip side, a well-crafted prompt can give you outputs that are clear, compelling, and conversion-ready.
So how do you write better prompts? The secret lies in using structured models, iterative prompting, and a few mental reframes that professional creators are now adopting. Let’s dive into a practical, step-by-step guide you can follow.
Why Prompt Quality Matters
A prompt is more than a question. It’s a blueprint. A good one tells AI what role to assume, what tone to use, what structure to follow, and what outcome to produce.
“Think of a prompt like giving directions to a talented intern. The more specific and clear your instruction, the better the outcome.”
Here’s what separates good prompts from bad ones:
-
Bad prompt: “Write a sales page for my course.”
-
Better prompt: “Act as a veteran copywriter. Write a sales page for a productivity course aimed at 45+ solopreneurs. Use AIDA structure. Keep the tone inspiring but practical.”
Which one do you think will get better results?
Let’s look at models and methods to help you write prompts like a pro.
1. Use Role Framing: “Act As…” or “Assume the Role Of…”
This is one of the most powerful ways to sharpen a prompt.
Why it works: It gives the AI a perspective, tone, and skill set to emulate.
Examples:
-
“Act as a world-class email copywriter and write a welcome sequence for my lead magnet.”
-
“Assume the role of a productivity coach. Create a 5-day plan to help mid-life solopreneurs focus and launch a product.”
- Assume the role of an expert prompt engineer and create a prompt that will generate [TOPIC] content.
The prompt should be optimised for clarity and effectiveness. (- this prompt creates a prompt which should be tweaked to output the desired content.)
Pro Tip: Stack roles if needed: “Act as a direct response copywriter, experienced solopreneur, and spiritual mindset coach…”
2. Use Prompt Engineering Models (Just like AIDA)
Sales copywriters use frameworks like AIDA, PAS, and FAB. You can use similar prompt-writing frameworks to guide the AI.
The PRISM Model (for Better Prompts)
Purpose – What do you want the output to do?
Role – Who should the AI act as?
Input – What context or background does it need?
Structure – What format do you want? (e.g. bullet points, steps, script)
Mode – What tone, style, or voice should it use?
Example using PRISM:
“Write a blog post (Purpose) as a marketing consultant (Role) based on the notes below (Input). Use a listicle format (Structure) and keep it witty and concise (Mode).”
3. Use the “Prompt the Prompt” Method
Instead of asking for content, ask ChatGPT to give you a prompt first. You can then refine it and rerun it for better quality.
Example:
“Act as an expert prompt engineer. Create a prompt that would generate a compelling lead magnet idea for solopreneurs over 45.”
Then, edit that prompt before using it. This gives you two passes: one for thinking, one for creating.
4. Give Examples in Your Prompt
AI LOVES examples. It learns from them in real time.
Bad: “Write an Instagram caption about consistency.”
Better: “Write an Instagram caption about consistency. Use this example as inspiration: ‘No trainer. No hype. Just consistency — for 20 months.’ Make yours short, punchy, and in that style.”
5. Tell It What You DON’T Want
Sometimes AI needs boundaries to avoid vague fluff.
Add “Do Not” Constraints:
-
Don’t use bullet points.
-
Avoid corporate language.
-
Don’t exceed 300 words.
This helps especially with voice and structure.
6. Use Iterative Prompting: “V1, V2, V3…”
Don’t expect to hit gold with one prompt. Ask for 3 versions and pick the best, then refine it.
Example Prompt:
“Write three variations of a tweet promoting my eBook about launching your first digital product. Make each one different in style—1. straight value, 2. story-driven, 3. curiosity hook.”
This gives you options, not a dead end.
7. Use the WIIFM Lens (“What’s in it for me?”)
Before you run a prompt, ask: “Is this customer-focused or me-focused?”
Better prompts focus on the end result for the reader or viewer, not just on what you want to say.
Bad: “Write a blog post about my coaching program.”
Better: “Write a blog post for 45+ solopreneurs struggling with consistency. Introduce a coaching program that helps them build daily habits that lead to income. Make the reader feel seen.”
8. Leverage “Stacked Prompts” or “Prompt Chains”
These are sequences where one prompt leads into another, each refining the output.
Example Chain for a Lead Magnet:
-
“Give me 5 lead magnet ideas for mid-life solopreneurs who want to use AI.”
-
“Take the best one and write a detailed outline.”
-
“Now write a landing page using the outline above.”
Think of it as a workflow of prompts instead of a single one-shot request.
9. Use Templates: Pre-Built Prompt Blueprints
You can create your own prompt templates over time.
Here’s one for blog posts:
Save these templates in Notion or a Prompt Library.
10. Use Meta-Prompts for Evaluation
Once AI gives you content, you can ask it to critique itself.
Examples:
-
“How could this email be made more persuasive?”
-
“Give this blog post a punchier intro and tighter conclusion.”
-
“What’s missing from this listicle?”
This is a powerful way to collaborate with the AI, not just consume its output.
Bonus: A Few Prompting Power Tips
-
Be conversational. Write prompts like you’d speak to a team member.
-
Avoid generic words like “great,” “interesting,” or “content.” Use specifics.
-
Timebox requests. e.g. “Make this a 5-minute read.”
-
Use analogies to shape tone: “Write this like Hemingway meets Seth Godin.”
Final Word: Prompting Is a Skill, Not a Hack
Writing great prompts isn’t about tricking the machine—it’s about clarifying your own thinking. AI reflects your intent back to you. The clearer the signal, the sharper the result.
If you’re a creator, solopreneur, or coach, this is a skill worth practicing every day. Build a “prompt muscle,” and you’ll create faster, better, and with more impact.
“The prompt is the product. When you get that right, everything downstream gets easier.”
Call to Action
I have created a GPT called Prompt Builder Pro that uses 8 Prompt Models, including PRISM. It chooses which model to use after you answer a few questions, before creating the most optimised prompt for your circumstances. Look out for it in my store.
And there’s more…
Here’s a curated list of prompt-structuring models similar to PRISM, including a few inspired by proven copywriting and creative thinking frameworks:
🔶 1. RAISE Model – For Problem-Solving or Strategy Prompts
Use when: You want AI to help you solve problems or create a system/strategy.
-
Role – Who should the AI be?
-
Audience – Who are you trying to help or speak to?
-
Issue – What is the core problem or opportunity?
-
Solution – What type of solution are you asking for? (framework, list, story)
-
Expression – In what format and tone?
Example Prompt:
“Act as a business strategist (R) helping solopreneurs over 50 (A) who struggle with inconsistent income (I). Suggest a step-by-step framework (S) in a clear, motivating tone (E).”
🔷 2. CREATE Model – For Generating Creative Assets (blogs, videos, hooks)
-
Channel – What platform is it for? (Blog, Twitter, Email)
-
Role – What persona should the AI adopt?
-
Emotion – What feeling should the output evoke?
-
Audience – Who is this for?
-
Theme – What’s the main topic or idea?
-
End Goal – What do you want the reader to do/feel next?
Example Prompt:
“Act as a viral Twitter ghostwriter (R). Create a thread (C) for 45+ creators (A) about building habits (T) that evokes hope and empowerment (E). End with a CTA to join a free email list (E).”
🔶 3. IDEAL Model – For Educational or Explainer Prompts
-
Intent – What’s the goal? (explain, teach, persuade)
-
Depth – Beginner, intermediate, expert?
-
Examples – Should it include examples or analogies?
-
Angle – What unique spin or perspective?
-
Layout – What structure? (list, story, tutorial)
Example Prompt:
“Explain blockchain (I) to beginners (D) using simple analogies (E), focusing on how it decentralizes power (A), as a 5-step tutorial (L).”
🔷 4. SOFT Model – For Tone, Style, and Format Customization
-
Style – Conversational, professional, playful, etc.
-
Outcome – What’s the desired reader action or feeling?
-
Format – Email, bullet list, story, Q&A, etc.
-
Theme – Central topic or message
Example Prompt:
“Write a casual-style (S) email (F) on the theme of procrastination (T), designed to make readers feel understood and take action (O).”
🔶 5. CLEAR Model – For Clarity-First Prompts
-
Context – What’s the backstory or setup?
-
Limitations – Any word count, tone, exclusions?
-
End goal – What should the output achieve?
-
Audience – Who’s it for?
-
Role – Who should AI act as?
This model is great for filtering vague thoughts into usable instructions.
🔷 6. FAB Prompting (Adapted from Features-Advantages-Benefits for Copywriting)
Use this to create persuasive or customer-focused prompts.
-
Feature – What does the product/idea/tool do?
-
Advantage – What makes it unique or better?
-
Benefit – How does it help the end user emotionally or practically?
Prompting Example:
“Write a short sales page for a Notion planner (F) that works offline and syncs with calendars (A), and helps overwhelmed solopreneurs finally get control of their day (B).”
🔶 7. SCOPE Model – For Comprehensive Idea Generation
-
Subject – What’s the general topic or field?
-
Context – What’s the use case or need?
-
Objective – What result do you want?
-
Perspective – Any unique angle or worldview?
-
Expression – Desired format (headline, list, tweet, etc.)
Bonus Tip: Combine These Models Like LEGO Blocks 🧱
You can stack these models depending on the complexity of the prompt. For example:
-
Use PRISM to set structure
-
Add RAISE to identify the pain point
-
Then overlay SOFT to fine-tune the tone
Leave A Comment