Putting It Together
Real examples across writing, analysis, research, and creative work. See the patterns in action.
You've learned the mental models. Now let's see them in practice. Each example shows the before and after—and why the "after" version gets better results.
The pattern you'll see repeated: Context + Task + Constraints + Quality bar
Example 1: Writing Task (Follow-up Email)
"Write a follow-up email"
Problem: No context, no goal, no tone guidance
"Write a follow-up email to a prospect I met at a conference yesterday.
Context: We talked about their team's struggle with manual data entry. They seemed interested in our automation tool but didn't commit.
Goal: Schedule a 20-minute demo call next week.
Tone: Professional but conversational. Reference our conversation.
Length: Under 150 words."
Why it works: Context (conference, their pain point), Task (schedule demo), Constraints (tone, length), Quality bar (reference conversation)
Example 2: Analysis Task (Sales Data)
"Analyze this data"
Problem: Which angle? For what purpose?
"Analyze this Q4 sales data. I'm presenting to the executive team tomorrow.
Specific questions:
- • Which product categories drove growth?
- • Did we hit our regional targets?
- • Any concerning trends in customer retention?
Format: 3 bullet points per question. Lead with the answer, then supporting data.
Focus on: Actionable insights, not just numbers. If something needs attention, flag it."
Why it works: Clear use case (exec presentation), Specific questions, Deliverable format, Permission to prioritize ("flag issues")
Example 3: Creative Task (Podcast Ideas)
"Give me podcast episode ideas"
Problem: For what show? What audience? What style?
"Generate 5 podcast episode ideas for a show about productivity for remote workers.
Show context: 30-minute episodes. Conversational, not preachy. Audience is knowledge workers tired of generic advice.
Recent topics we've covered: Deep work, async communication, avoiding burnout.
Format for each idea:
- • Episode title (punchy, max 6 words)
- • 2-sentence description
- • One counterintuitive angle or hook
Avoid: Anything we've done before. Generic 'tips and tricks' angles."
Why it works: Show context (style, audience), What to avoid (past topics), Deliverable format (title + description + hook), Quality bar (counterintuitive)
Example 4: Research Task (Decision Support)
"Tell me about solar panels"
Problem: Too broad. What's the decision?
"I'm deciding whether to install solar panels on my house in northern California.
My situation:
- • Average monthly electric bill: $180
- • Planning to stay in this house for 10+ years
- • Upfront budget: ~$20k
What I need:
- • Realistic payback timeline
- • Major risks or downsides I should know
- • Whether to buy or lease
Skip: History of solar technology, general environmental benefits (I already support the idea)."
Why it works: Specific situation (location, budget, timeline), Clear scope (decision factors), Explicit "skip this" (saves attention for what matters)
The Pattern
All four examples share the same structure. Once you see it, you can apply it to anything:
Context
What's the situation? Who's the audience? What constraints exist?
Task
What exactly do you need? Be specific about the deliverable.
Constraints
Format, length, tone, style. What are the boundaries?
Quality bar
What makes this "good"? What should it avoid? How will you know it's right?
Use this as a template. When you're not getting good results, check if you've covered all four pieces. One of them is probably missing.
Key Takeaway
Good prompts aren't magic. They're structured. They give the AI exactly what it needs to deliver what you want.
Context + Task + Constraints + Quality bar. Use this pattern, and you'll get consistent results across any domain.