Free Interactive Tool

Write Prompts That
Actually Work

Stop getting mediocre AI responses. Learn the CREATE framework and build powerful, structured prompts that deliver consistent, high-quality results.

Save TimeReduce back-and-forth editing cycles
Better ResultsStructured prompts outperform vague ones
RepeatableBuild a library of reusable prompts

The Difference a Good Prompt Makes

Weak Prompt

"Write a marketing plan for my business."

Generic, unfocused output that needs heavy editing
Engineered Prompt

"You are a senior digital marketing strategist with 10+ years of B2B SaaS experience. Create a 90-day content marketing plan for a fintech startup targeting mid-market CFOs. Include audience segmentation, content types, distribution channels, and a weekly publishing calendar. Use a professional tone. Before starting, ask up to 5 clarifying questions one at a time."

Targeted, structured output ready to use

The difference? Context, role, structure, and letting the AI ask questions. That's the CREATE framework.

The CREATE Framework

Six elements that transform basic prompts into powerful instructions. Click each to learn more.

C

Context

Provide relevant background information

Give the AI the background it needs to understand your situation. Strong context reduces guessing and improves accuracy.

  • Include the purpose, audience, goals, and constraints
  • Share only details that matter — avoid information overload
  • Think of it as the setup you'd give a human teammate
  • Good context anchors reasoning and reduces hallucinations
Example: "Our company is a B2B SaaS startup in the fintech space. We're launching a new expense management tool next quarter. Our target market is mid-size companies (100-500 employees) with distributed finance teams. Budget for the launch campaign is $50K."
R

Role

Define the AI's expertise and perspective

Assigning a role tells the AI what kind of expert to be. This shapes the depth, vocabulary, and perspective of the response.

  • Be specific: "seasoned digital marketing strategist with 10+ years in B2B SaaS" beats "marketing expert"
  • Try different roles for the same task to see how perspectives change
  • The role determines tone, jargon, and priorities
Example: "You are a senior content strategist with 12 years of experience in B2B technology marketing. You specialize in thought leadership content, demand generation, and marketing automation for SaaS companies."
E

Example

Show the format, tone, or structure you expect

Examples set expectations clearly. They show the AI what "good" looks like before it starts generating.

  • Include sample structure, formatting, or tone
  • Show what the output should look like
  • Even brief examples dramatically improve consistency
Example: "The output should follow this structure: an executive summary, three key sections with subheadings, a data table comparing options, and a clear recommendation with next steps. Use bullet points for key takeaways."
A

Ask

Let the AI ask clarifying questions first

This is a game-changer. By telling the AI to ask questions before starting, you get a much more tailored result.

  • Specify how many questions (e.g., "up to 5")
  • Ask for questions "one at a time" for a conversational flow
  • This mimics how a real expert would start a project
  • Reduces assumptions and fills in gaps you didn't think of
Example: "Before proceeding, ask up to 5 clarifying questions, one at a time — such as the target audience, key metrics, timeline, existing assets, and brand guidelines."
T

Task

Clearly describe the output you want

Be explicit about what you want the AI to produce. The clearer the task, the better the output.

  • Describe the deliverable: report, plan, email, code, etc.
  • Include length, format, and structure requirements
  • Mention specific sections or elements to include
  • State what success looks like
Example: "Create a 90-day content marketing plan that includes: audience personas, content pillar topics, a weekly publishing calendar, distribution strategy across LinkedIn and email, and KPIs for measuring success."
E

Enhance

Iterate and refine for better results

Prompt engineering is a loop, not a one-shot. The best prompts go through multiple rounds of testing and refinement.

  • Test your prompt in a fresh chat
  • Review: Does it match the goal? Is the tone right? What's missing?
  • Go back to your engineering chat to refine
  • Repeat until results are consistently what you want
  • Save the final version to your prompt library
Tip: Open a separate chat where the AI acts as your prompt engineer. Engineer the prompt there, then test it in a fresh chat.

Prompt Builder

Answer a few questions and we'll assemble a powerful prompt using the CREATE framework.

Goal Voice C R E A T E
Getting Started

What do you need AI help with?

Describe your goal in a sentence or two. Don't worry about being perfect — we'll refine it.

Quick start:
Brand Voice

What voice should the output use?

A strong brand voice ensures consistency across all AI-generated content. Define your own or skip for a neutral tone.

Custom Voice

Define your brand voice, communication style, or personality for the AI to adopt in the generated prompt.

No Voice

Skip voice styling. The AI will use its default, neutral tone.

C — Context

What's the background?

Help the AI understand the situation. Think: who, what, why, and any constraints.

R — Role

Who should the AI be?

Pick or customize the expert persona. This shapes the depth and perspective of the response.

E — Example

What should the output look like?

Describe the format, structure, or tone you want. This sets clear expectations.

A — Ask

Should the AI ask questions first?

Letting the AI ask clarifying questions before starting produces much more tailored results.

T — Task

What exactly should the AI produce?

Be specific about the deliverable. The clearer you are, the better the output.

E — Enhance

Your Engineered Prompt

Here's your assembled prompt. Copy it, test it, and come back to refine.

Your AI-Engineered Prompt
AI Powered

How to Enhance

  1. Test it — Paste this prompt into ChatGPT, Claude, or Gemini in a fresh chat
  2. Review the output — Does it match your goal? What's missing?
  3. Refine here — Go back to any step and adjust, then re-generate
  4. Save it — Once it works consistently, save it to your prompt library

Example Prompts

See the CREATE framework in action. Click any example to load it into the builder.

Business Proposal

A compelling business proposal with executive summary, scope, pricing, timeline, and expected outcomes.

Try This →

Marketing Plan

A 90-day content marketing strategy with audience personas, content pillars, distribution channels, and KPIs.

Try This →

Board Presentation

A structured presentation outline for a board meeting with data points, key metrics, and strategic recommendations.

Try This →

Sales Outreach Email

A professional cold outreach email that opens a conversation without being pushy or generic.

Try This →

Job Description

A compelling job posting with responsibilities, qualifications, culture section, and competitive benefits.

Try This →

Project Scope

A project scope document with objectives, deliverables, timeline, dependencies, and success criteria.

Try This →

Pro Tips

Advanced techniques to level up your prompt engineering.

Chain of Thought

Add "Walk through your reasoning step by step before giving the final answer" to improve accuracy on complex tasks like planning, analysis, or math.

Reduce Hallucinations

Tell the AI: "If you're not confident, say so instead of guessing" and "Cite reputable sources with publication dates." Run a research step first for factual work.

Talk, Don't Type

Use voice input — it's 3-4x faster than typing. AI models love context, and speaking naturally gives them more to work with.

Engineer First, Use Second

Open a separate chat where the AI acts as your prompt engineer. Build and refine the prompt there, then test it in a fresh chat.

Choose the Right Model

Use powerful models (GPT-4, Claude) for complex reasoning. Use lighter models for rewrites and brainstorming. Use research modes for factual work.

Human in the Loop

For high-stakes decisions (legal, financial, medical), always have an expert review. Treat AI as a fast researcher, not the final decision-maker.

The Prompt Engineering Workflow

Prompt engineering is a loop, not a one-shot. Follow this process for the best results.

1

Draft

Write a simple version of your prompt

2

Engineer

Open a separate chat — have AI use CREATE to improve it

3

Test

Run the engineered prompt in a fresh chat

4

Review

Does it match the goal? What's missing or off?

5

Enhance

Return to the engineering chat, refine, retest

6

Save

Lock in the final prompt to your library for reuse

Ready to transform how you use AI?

The CREATE framework is just one part of how MyZone AI helps businesses work smarter. Explore our workshops, coaching, and AI implementation services.