Structured Prompt Builder
Start with the fundamentals — persona, task, and context.
Build
Examples
Guide
Help
Define expertise level, industry, and perspective.
Use verbs: Write, Analyze, Summarize, Compare, Generate, Evaluate.
Who's involved? What's the situation? What constraints exist?
Example Input
Example Output
1
2
1
2
A
B
Level 1: Structure the Prompt
Structured Prompt: Change Management Message
❌ WEAK PROMPT
"Help me write something about change management."
✓ STRONG PROMPT
"You are a senior change strategist. Write a 250-word message to frontline managers explaining an ERP migration. Tone: direct and reassuring."
COMPARE THE RESULTS:
"Change management is important for organizations. When implementing new systems, communication helps. Managers should talk to their teams about changes. This helps people understand what's happening..."
Generic, vague, no clear audience or purpose.
"Team, I want to be direct: we're migrating to a new ERP system in Q2. Your teams will ask questions—here's what matters. This isn't about efficiency theater. Our current system can't scale with regional growth..."
Specific, actionable, appropriate tone and length.
Few-Shot Prompt: Product Descriptions
❌ WEAK PROMPT
"Write product descriptions for me."
✓ STRONG PROMPT
"Input: Wireless mouse → Output: Ergonomic wireless mouse with 6 programmable buttons. 18-month battery life. Now write for: Standing desk"
COMPARE THE RESULTS:
"This is a standing desk that allows you to work while standing. It can be adjusted to different heights and is made of quality materials. Great for offices and home use."
Doesn't match the example pattern; verbose and generic.
"Height-adjustable standing desk with electric motor. 48-72" range. Built-in cable management and memory presets."
Matches the concise, feature-focused pattern perfectly.
Chain-of-Thought: Project Risk Analysis
❌ WEAK PROMPT
"Analyze this project's risks for me."
✓ STRONG PROMPT
"Step 1: Identify root causes. Step 2: Map stakeholder impacts. Step 3: Evaluate solution options. Step 4: Recommend mitigation approach."
COMPARE THE RESULTS:
"Some potential risks include timeline delays, budget overruns, and team capacity issues. Consider these carefully."
Surface-level list without structured analysis.
"Step 1 - Root Causes: Key dependency on vendor API (not yet GA). Step 2 - Stakeholder Impact: Affects product launch timeline... Step 3 - Options: Build abstraction layer vs. wait for GA..."
Systematic, follows reasoning steps, actionable insights.
Level 2: Structure the Thinking
Task Decomposition: Strategy Document
❌ WEAK PROMPT
"Create a comprehensive AI strategy document."
✓ STRONG PROMPT
"Stage 1: Analyze current capabilities. Stage 2: Define success metrics. Stage 3: Identify use cases. Stage 4: Draft implementation roadmap."
COMPARE THE RESULTS:
"An AI strategy should include vision, goals, implementation plan, and success metrics. Consider stakeholders, technology, and organizational readiness..."
Jumbled structure, attempts everything at once.
"Stage 1 - Current Capabilities: [detailed analysis building foundation] Stage 2 - Success Metrics: [builds on Stage 1] Stage 3 - Use Cases: [informed by metrics]..."
Each stage builds on previous; coherent narrative flow.
Decision Framework: Vendor Selection
❌ WEAK PROMPT
"Help me decide between these two vendors."
✓ STRONG PROMPT
"Option A: Build in-house. Option B: Buy vendor solution. Criteria: Cost, time-to-market, control, scalability. Method: Weighted scoring (cost 40%, time 30%, control 20%, scalability 10%)"
COMPARE THE RESULTS:
"Both vendors have pros and cons. Vendor A is cheaper but Vendor B has better features. Consider your budget and needs."
No structured evaluation; opinion without framework.
"Weighted Analysis: Option A - Cost: 4/5 (40%=1.6), Time: 2/5 (30%=0.6)... Total: 2.9/5. Option B - Cost: 2/5 (0.8), Time: 5/5 (1.5)... Total: 3.7/5. Recommendation: Option B scores 28% higher..."
Quantified, transparent methodology, clear rationale.
Framework Guide
Level 1 — Structure the Prompt
Structured Prompt
The foundation. Always include: Persona (who), Task (what), Context (why/for whom). Add Constraints and Format for control.
Use for: Any task. This is your starting point.
Few-Shot Prompt
Teach by example. Show 1-3 input/output pairs to establish the pattern you want. Model matches your examples.
Use for: Style matching, format replication, pattern recognition.
Chain-of-Thought
Force step-by-step reasoning. Define 3-5 explicit thinking steps. Prevents shortcuts on complex problems.
Use for: Analysis, planning, multi-step problems, decision support.
Level 2 — Structure the Thinking
Task Decomposition
Break big tasks into sequential stages. Each stage produces input for the next. Reduces cognitive load.
Use for: Long reports, complex strategies, multi-phase projects.
Decision Framework
Structured comparison of 2+ options. Define options, criteria, and evaluation method. Forces systematic thinking.
Use for: Go/no-go decisions, vendor selection, strategic choices.
Common Mistakes
• Vague personas: "Expert" → "Senior behavioral economist who explains concepts to executives"
• Weak verbs: "Help me with..." → "Write / Analyze / Compare / Generate"
• Missing constraints: Always specify length, tone, and audience
• No examples: For style-sensitive work, show 1-2 examples
• Skipping context: Model can't read your mind — state what matters
Help & FAQ
Getting Started
How do I create my first prompt?
1. Select "Structured Prompt" from the sidebar
2. Fill in Persona, Task, and Context
3. Watch the Live Preview update
4. Check Quality Signals for feedback
5. Click Copy when ready
2. Fill in Persona, Task, and Context
3. Watch the Live Preview update
4. Check Quality Signals for feedback
5. Click Copy when ready
What do the colors mean in Quality Signals?
● Green = Good to go!
● Yellow = Could be more specific
● Red = Missing critical information
● Yellow = Could be more specific
● Red = Missing critical information
Which framework should I use?
Start with Structured Prompt for any task. Use Few-Shot when showing examples is easier than explaining. Use Chain-of-Thought for analysis or multi-step problems. Level 2 frameworks are for advanced use cases.
Library & Backup
How do I save my prompts?
1. Fill in your prompt fields
2. Scroll to "Save to Library" at bottom
3. Enter a name and tag
4. Click Save
Your prompts are automatically stored in your browser.
2. Scroll to "Save to Library" at bottom
3. Enter a name and tag
4. Click Save
Your prompts are automatically stored in your browser.
Why should I export my library?
Browser data can be lost if you clear cookies, use a different browser, or experience technical issues. Click 📚 Library → ⬇ Export Library to download a backup file. Store it safely and import when needed.
My library disappeared - what happened?
Likely causes: Incognito/Private browsing, cleared browser data, or different browser/device. If you exported your library, click 📚 Library → ⬆ Import Library to restore. Otherwise, prompts are lost (this is why we recommend regular exports).
Best Practices
What makes a good Persona?
Be specific: "Senior change strategist with 10 years in Fortune 500 transformation" beats "Expert". Include expertise level, perspective, and relevant background.
How specific should my Task be?
Very specific! "Write a 250-word email announcing the ERP migration to frontline managers" beats "help me write something about change management". Use action verbs and define the deliverable.
What should I put in Context?
Answer: Who's the audience? What's happening? What constraints matter? Use the • Bullets toggle for clarity. Example: "• Audience: Senior executives • Situation: Q4 planning • Constraint: 2-week timeline"
Features & Tips
What's the bullet toggle in Context?
Click • Bullets button to automatically format your context with bullet points. Click again (shows ✓) to remove bullets. Makes context easier to read and scan.
How do I use the Examples tab?
The Examples tab shows weak vs. strong prompts with actual AI outputs. See concrete before/after comparisons to understand what makes prompts effective. Great for learning!
Does this work on mobile?
Yes! The app is fully responsive. On mobile, sections stack vertically. Desktop is preferred for intensive work, but mobile works great for quick edits or reviewing saved prompts.
Troubleshooting
Quality Signals aren't updating
Make sure you're filling fields for the active framework (selected in sidebar). Framework-specific fields only appear when that framework is selected.
Fields aren't showing up
Each framework shows different fields. For example, "Few-Shot" shows Example Input/Output boxes. "Chain-of-Thought" shows Reasoning Steps. Select the framework you want from the sidebar first.
Browser compatibility issues?
ETC works on Chrome 90+, Firefox 88+, Safari 14+. Update your browser if experiencing issues. Incognito/Private mode won't save library data.
Still need help?