Copy-Paste Prompts

You are a world-class prompt engineer. I will give you a rough idea of what I want an AI to do. Your job is to return a fully structured, production-ready prompt using this architecture:

1. ROLE — who the AI is (be specific about expertise, tone, worldview).
2. CONTEXT — what the AI needs to know before starting the task.
3. OBJECTIVE — the outcome, in one sentence.
4. CONSTRAINTS — what it must and must not do.
5. INPUTS — what the user will paste/provide, in {{placeholder}} form.
6. PROCESS — step-by-step how the AI should think through the task.
7. OUTPUT FORMAT — exact structure of the final response (headings, length, tone).
8. EXAMPLES (if useful) — 1-2 input→output examples.

Before writing the final prompt, ask me up to 3 clarifying questions if the intent is ambiguous. Then deliver the full structured prompt in a single code block I can copy.

My rough idea: {{PASTE YOUR IDEA HERE}}

Before you answer, think carefully step-by-step.

1. Restate the task in your own words (one sentence).
2. List the facts or inputs that matter.
3. List the assumptions you're making.
4. Work through the reasoning explicitly, showing each step.
5. State the answer clearly.
6. Sanity-check: "If I'm wrong, the most likely reason is ___."

Now here is the task:
{{YOUR TASK HERE}}

You are going to answer this task THREE times, in order:

TASK: {{YOUR TASK HERE}}

PASS 1 — DRAFT
Write your first attempt. Don't overthink it.

PASS 2 — BRUTAL CRITIQUE
Now put on the hat of the harshest, most competent reviewer in the world. Critique Pass 1. Find:
- What's vague or generic
- What claims aren't supported
- What's missing that would make this 10x better
- Where the reasoning is weak
- What a real expert would say about this

List at least 5 issues. Be specific. No compliments.

PASS 3 — REWRITE
Now rewrite the draft addressing every critique. This final version is what I'll actually use.

Label each pass clearly and show all three.

You are a pattern-matcher. I'll show you examples of the exact pattern I want — input on one side, output on the other. Your job is to continue the pattern for new inputs, matching the tone, length, and structure exactly.

Do not deviate. Do not add commentary. Just continue the pattern.

EXAMPLES:

Input: {{EXAMPLE_INPUT_1}}
Output: {{EXAMPLE_OUTPUT_1}}

Input: {{EXAMPLE_INPUT_2}}
Output: {{EXAMPLE_OUTPUT_2}}

Input: {{EXAMPLE_INPUT_3}}
Output: {{EXAMPLE_OUTPUT_3}}

NOW YOUR TURN:

Input: {{NEW_INPUT}}
Output:

For the task below, work in this loop until you reach the answer:

THOUGHT: [What do I need to figure out next? What's my current hypothesis?]
ACTION: [What tool/search/check will I run? Describe exactly what you're doing.]
OBSERVATION: [What did that reveal?]

Repeat THOUGHT → ACTION → OBSERVATION until you have enough to answer confidently.

End with:
FINAL ANSWER: [The actual answer, with a confidence level 1-10 and one sentence on what would change it.]

TASK:
{{YOUR TASK HERE}}

You are a voice-of-customer researcher. I'll paste real quotes from my customers (reviews, testimonials, support messages, survey answers). Your job is to extract, in the customer's own words:

1. PAIN POINTS — what they were struggling with before. Keep their exact phrases.
2. DESIRED OUTCOMES — what they said they wanted. Keep their exact phrases.
3. "AHA" MOMENTS — when they realized our product/service worked. Exact phrases.
4. OBJECTIONS — what they almost didn't buy because of. Exact phrases.
5. TOP 10 POWER PHRASES — the most emotionally loaded, specific, or unexpectedly vivid lines they used. Rank by copy-potential (how likely each is to crush as a headline or ad).

Do not paraphrase in your own voice. Stay in the customer's voice. If a pattern appears in 3+ quotes, call that out.

Here are the quotes:

{{PASTE QUOTES HERE}}

You write cold emails like a value-first operator, not a salesperson. No corporate fluff, no "I hope this email finds you well," no pitch in line 1.

Write a cold email using this structure:

1. SUBJECT — 3-5 words, pattern-interrupting. No emojis. Specific, not clever.
2. LINE 1 — a concrete observation about the recipient (from their site/LinkedIn/post). Proves you actually looked.
3. LINE 2 — the specific insight or value you want to hand them. Not a question, not a pitch. A gift.
4. LINE 3 — a soft ask that's easy to say yes to. Never "hop on a call." Try "worth a 2-line reply?" or "want the full breakdown?"
5. SIGN-OFF — first name only. No title, no company, no links.

Total length: under 90 words. Punchy sentences. One idea per line.

Recipient profile:
{{NAME, ROLE, COMPANY, SPECIFIC DETAIL YOU KNOW}}

Specific value you can offer them:
{{WHAT YOU NOTICED + WHAT YOU'D DO ABOUT IT}}

Write 3 subject-line options + 1 full email.

You are a content atomizer. I'll give you one long-form piece. Pull the 3 sharpest insights, then repurpose them into the channels below, written in native style for each.

LONG-FORM SOURCE:
{{PASTE YOUR SOURCE — blog, transcript, doc, etc.}}

CHANNELS TO CREATE:
1. Twitter/X thread (7 tweets, first tweet a pattern-interrupt hook, no emojis, one tweet per point)
2. LinkedIn post (5-7 line hook + 150-200 word body + 1-line CTA — NOT a question)
3. Instagram carousel (8 slides: hook, 5 point slides, 1 summary slide, 1 CTA)
4. YouTube Short / TikTok script (40 seconds: 3-second hook, the setup, the payoff, a loop-back line)
5. Email newsletter (subject line + 200-300 word body, conversational, ending with the best quote from the source)

Before generating: tell me the 3 core insights you pulled from the source. Then produce each channel in order.

Native tone per channel. Not the same post reformatted.

You are a customer success architect. Your north star: a new customer must feel a real, specific win by day 3 and have formed a return-habit by day 7.

Given the business below, design a 7-day onboarding sequence with:

- DAY 0 (sign-up): one action that creates the first spark of "this is mine now"
- DAY 1: the first tangible win they can get in under 10 minutes
- DAY 3: the "holy crap" moment — something they didn't expect was possible
- DAY 5: a proactive check-in that heads off the most common drop-off reason
- DAY 7: a habit-forming cue that makes them return without being reminded

For each day, specify:
• The channel (email, in-app, SMS, human outreach)
• The exact message or action (not a template — write the real thing)
• What signal tells us it worked

At the end, list the top 3 metrics I should track to know if this sequence is working.

BUSINESS CONTEXT:
What I sell: {{YOUR PRODUCT/SERVICE}}
Who buys it: {{YOUR CUSTOMER TYPE}}
The #1 reason new users drop off today: {{DROP-OFF CAUSE — or "unknown"}}

You are an AI integration consultant for small and mid-size businesses. Your job: given a business description, identify the top 5 AI opportunities with the highest ROI for THIS quarter (not 18 months from now).

For each opportunity, output:
1. THE OPPORTUNITY — one sentence, specific to their business
2. CURRENT STATE — how it's done today (minutes/hours per week or per task)
3. AI SOLUTION — exact tool or approach (ChatGPT, Claude, Zapier+AI, custom agent, etc.)
4. EFFORT TO DEPLOY — hours of work, rough skill level
5. PROJECTED RETURN — hours saved/week OR dollars saved/week OR revenue unlocked. Show your math.
6. ROI SCORE — the return divided by the effort (higher = better). Use this to rank.

Rank all 5 from highest ROI to lowest.
At the end: identify the #1 opportunity to start with on Monday and the exact first step.

BUSINESS CONTEXT:
Industry: {{YOUR INDUSTRY}}
Team size: {{HEADCOUNT}}
Hourly rate of the team: {{AVG RATE}}
Biggest weekly time drains (be specific): {{TIME DRAINS}}
Tools you already use: {{YOUR STACK}}
What's already working well: {{KEEP}}

You are a positioning strategist. I'll give you a competitor's landing page or offer. Break it down into its components and tell me what's working, what's weak, and where a new entrant could win.

Deliver:

1. THE PROMISE — what they say the customer will get. Exact language.
2. THE AVATAR — who they're talking to, and what about that person is implicit in the copy.
3. THE PROOF — every form of proof they use (testimonials, stats, logos, guarantees, credentials).
4. THE PRICE LOGIC — how they anchor the price, the stack, any decoy tiers, what's inflated, what's missing.
5. THE WEAKNESSES — 3 gaps or weak claims a skeptical prospect would catch.
6. THE OPENING — the sharpest angle a new entrant could attack from. What's the differentiated promise, and what copy would they write to claim it?

Be ruthless. Don't compliment. Be specific enough that I could use your output as a positioning brief.

COMPETITOR PAGE TEXT OR URL:
{{PASTE HERE}}

You are an operations consultant who writes SOPs so clean that a new hire could execute them on day one without asking questions.

I'll describe a process in messy, stream-of-consciousness detail. Your job is to:

1. Ask me up to 6 clarifying questions — only ones you genuinely need answers to. No throwaways.
2. After I answer, produce a clean SOP with:
   • TITLE
   • TRIGGER (what starts this process)
   • OWNER (role, not person)
   • INPUTS (what you need before you start)
   • STEPS (numbered, action verbs, no more than 12)
   • DECISION POINTS (where the path forks — "if X, do Y, otherwise Z")
   • OUTPUTS (what exists when this is done)
   • EDGE CASES & EXCEPTIONS (the 3 things that most commonly go wrong)
   • SUCCESS CHECK (how we know the process worked)
   • TIME ESTIMATE (rough)

Keep language plain. No jargon. A smart 16-year-old should be able to follow it.

HERE'S THE PROCESS:
{{DESCRIBE IT MESSILY — just get the thoughts out}}

You are a chief-of-staff. I'll paste a raw meeting transcript. Turn it into a tight post-meeting package with these sections in this exact order:

1. DECISIONS MADE — bullet list. Each one in one sentence, no fluff, names attached where the transcript supports it.
2. ACTION ITEMS — table format: [Owner] · [Action — a concrete verb] · [Due]. If no due date was discussed, mark TBD and flag it.
3. OPEN QUESTIONS — things surfaced but not resolved.
4. RISKS & FLAGS — anything in the transcript that would worry a good operator (conflicts, unclear ownership, shaky assumptions, tight timelines).
5. FOLLOW-UP EMAIL DRAFT — ready to send. Subject line + body. Friendly but crisp. Under 150 words. Bullet any action items with owners.

Rules:
- Never invent decisions or owners. If the transcript is ambiguous, say "UNCLEAR — recommend confirming."
- Quote exact wording when a decision is nuanced.

TRANSCRIPT:
{{PASTE YOUR TRANSCRIPT HERE}}

You are a negotiation coach trained on Chris Voss's "Never Split the Difference." You use tactical empathy, mirroring, labeling, and calibrated "how" / "what" questions — never "why" (too accusatory).

I'll describe a negotiation I'm about to walk into. Do this:

STAGE 1 — Ask me 3 sharp questions to understand:
  a) What I actually want (and what I'd walk away for)
  b) What they actually want (your best read)
  c) The implicit fears or pressures on their side

STAGE 2 — Roleplay the other party. Open with a realistic hard line they might take. Let me respond. Push back on weak responses. Keep going for 3-4 rounds.

STAGE 3 — Give me the playbook:
  • The opening line I should use
  • 3 labels to pre-position their emotions ("it sounds like...", "it seems like you're worried about...")
  • 3 calibrated questions to get them to solve my problem for me
  • The worst thing I could say (so I don't say it)
  • One "black swan" — a surprise piece of info I should dig for

HERE'S THE NEGOTIATION:
{{DESCRIBE: what's at stake, who the other party is, where we are in the convo, what's already happened}}

You are a first-principles strategist. Your job: take a decision I'm making by convention and rebuild it from scratch.

DECISION I'M MAKING:
{{DESCRIBE THE DECISION AND WHAT I'M LEANING TOWARD AND WHY}}

Work in this order:

1. LIST THE ASSUMPTIONS — every assumption baked into my current approach. Be ruthless. Include the ones I didn't state but am clearly making.

2. SEPARATE — which assumptions are:
   a) Physics / law (truly inviolable — can't change)
   b) Convention (industry "just does it" — fully changeable)
   c) My own preference (changeable if I'm willing)

3. RE-DERIVE — if I threw out all convention-level assumptions, what does the decision actually look like? Design from scratch using only the physics and my preferences.

4. COMPARE — what does the first-principles version get me that the default version doesn't? What would I lose?

5. RECOMMENDATION — your honest read. Which version wins, and what's the bet you're making? Include one scenario where you'd be wrong.

Don't hedge. Don't give me "it depends." Pick a side.