AI Business Prompts

The Business Case for Structured AI Execution

Structured AI Prompts to Power Business Success

Two years ago, "AI business prompts" meant a Google Doc of clever ChatGPT tricks someone on LinkedIn swore would change your life. Maybe it was "47 prompts that will 10x your marketing." Maybe it was a bookmark folder no one opened twice.

That era is over.

According to PwC's 2025 Global AI Jobs Barometer — an analysis of nearly one billion job postings across six continents — workers with AI skills, including prompt engineering, now command a 56% wage premium over their peers, up from 25% the year prior. The World Economic Forum projects that 39% of workers' core skills will change by 2030, with AI and big data ranking as the number one skill priority for global employers.

This isn't a productivity hack. It's a labor market realignment.

And the skill at the center of it — the ability to communicate with AI systems in structured, repeatable, high-quality ways — is what the industry now calls prompt engineering.

The question for every business leader is no longer whether to adopt AI. It's whether your teams know how to use it with the precision, consistency, and governance that separate real performance gains from expensive improvisation.

What "AI Business Prompts" Actually Means

AI business prompts are structured instructions that turn a general-purpose AI model into a repeatable business process. A business-grade prompt carries context about the audience and objective, a specified output format, quality constraints, and governance metadata — which is what separates a deployment-ready deliverable from plausible-sounding text someone has to rewrite.

The distinction matters more than most people realize. Globally, "prompts for ChatGPT" alone generates over 70,000 searches per year, and the United States, India, and Germany lead the world in AI prompt search volume. The Asia Pacific AI prompt marketplace is growing at 23.9% CAGR, with India alone registering 24.7% CAGR through 2032. Text prompts dominate 54% of the entire global prompt marketplace.

This demand isn't uniform. What people search for differs dramatically by industry and region. But the underlying need is consistent: output that works on the first try.

The Hidden Problem: The Hidden Cost of Improvising AI

Here's what's actually happening inside most organizations right now:

Seventy-five percent of knowledge workers already use AI tools at work, according to both Microsoft and McKinsey's 2025 research. But only 39% have received any formal AI training from their employer. Only 25% of companies plan to offer it.

That gap produces a specific, measurable problem: most AI usage in business is improvised.

Every department has someone who's "pretty good with ChatGPT." Marketing has a folder of prompts that one person maintains. Sales reps copy-paste from blog posts. The finance team avoids AI entirely because the first output they got back sounded like an undergraduate wrote it. HR uses it quietly and tells no one.

The result isn't just inconsistent output. It's compounding waste:

Rework cycles multiply

When prompts lack context, structure, and formatting specifications, the output requires heavy editing. The promised efficiency gain evaporates in revision loops.

Knowledge stays siloed

When every individual builds their own prompts, there's no organizational learning. The same mistakes get made across departments. The same problems get solved from scratch, repeatedly.

Quality varies wildly

One team gets genuinely useful AI output. Another gets generic boilerplate. The difference isn't the AI model — it's the prompt. And no one is governing the prompts.

Compliance risk accumulates

In regulated industries, unstructured AI usage creates audit exposure. There's no version history, no approval trail, no way to prove what was generated, by whom, or why.

What Is Prompt Debt?

This is what the industry now calls Prompt Debt — the hidden operational liability that accumulates every time an organization uses AI without structure.

Like technical debt in software, prompt debt compounds silently until the cost of doing nothing exceeds the cost of fixing it.

Every prompt without version control, without audit trail, without format specification is an unrecorded liability on your organization's balance sheet.

And right now, most organizations are accumulating it faster than they realize.

The Productivity Evidence Is No Longer Anecdotal

The business case for AI proficiency — and specifically for structured prompt engineering — is now backed by research from institutions that don't traffic in hype.

Performance improvements are measurable and significant.

A landmark study from Harvard Business School, MIT, and Boston Consulting Group — published in Organization Science and known as the "Jagged Frontier" study — tested 758 BCG consultants with and without GPT-4 access. Consultants using AI completed 12.2% more tasks, 25.1% faster, with over 40% higher quality than control groups. Participants who received prompt engineering training performed even better — demonstrating that the skill of prompting, not just access to the model, drives the outcome.

Separately, Stanford and MIT researchers studying 5,000+ customer service agents found that AI assistance increased productivity by 15% on average, with a 34% improvement for novice workers. That's not a theoretical projection — it's a measured gain in issues resolved per hour.

The economic stakes are enormous.

McKinsey's Global Institute estimates that AI-driven productivity gains represent $4.4 trillion in annual economic value potential. PwC's data shows that industries most exposed to AI have seen productivity growth nearly quadruple — from 7% between 2018–2022 to 27% between 2018–2024 — alongside 3x higher revenue-per-employee growth compared to less AI-exposed sectors.

The workforce shift is already in progress.

The World Economic Forum's Future of Jobs Report 2025 projects 170 million new jobs created by 2030, with a net gain of 78 million after displacement. But the composition of those jobs will be different: 59% of the global workforce will need some form of reskilling, and the WEF report explicitly identifies advanced prompt-writing skills as a critical training priority.

Meanwhile, McKinsey reports that the number of workers in occupations requiring AI fluency has grown sevenfold in just two years — from approximately 1 million in 2023 to around 7 million in 2025. Gartner predicts 80% of the engineering workforce must upskill by 2027.

The demand is especially acute in India. Between March 2024 and March 2025, AI and data talent demand in India grew by an estimated 38% to 45% year-over-year. India's AI sector is projected to exceed $8 billion in value, and average prompt engineer salaries in India range from ₹8 lakh to ₹13 lakh per year. 93% of Indian enterprises signaled plans to boost AI budgets in 2025, with generative AI usage multiplying fivefold since 2023.

This is not about whether AI works. It's about whether your organization can execute with it consistently, at scale, with measurable outcomes.

The Difference Between Asking and Engineering

When most people search for "AI business prompts," they're looking for shortcuts — pre-written instructions they can paste into ChatGPT to get faster output. That's not wrong, exactly. But it's incomplete in a way that matters.

There's a material difference between:

Ad hoc prompts

One-off, unstructured requests that vary every time they're used. "Write me a marketing email." "Summarize this report." "Give me a sales script."

These produce unpredictable output because they contain no context about audience, format, tone, constraints, or objectives.

Structured prompts

Engineered instructions that include role definition, situational context, qualifying questions, output format specifications, quality constraints, and governance metadata.

These produce consistent, deployment-ready output because they treat prompting as a system, not a guess.

The Difference at a Glance

The distinction is the same one that separates a draft from a deliverable, a suggestion from a strategy, and an experiment from an operation.

Research supports this at a granular level. The prompt engineering market — valued at $505 million in 2025 — is projected to reach $6.7 billion by 2034, growing at a compound annual rate of 33%, according to Fortune Business Insights. Structured prompt processes have been shown to reduce AI errors by up to 76%. And 68% of firms now provide some form of prompt engineering training — an acknowledgment that this skill is no longer optional.

When we use the phrase "AI business prompts," we mean something specific: prompts engineered for business outcomes, organized by business function, governed for team-wide consistency, and designed to work across every major AI model.

That's not a template. It's infrastructure.

How Business Functions Deploy Structured AI Prompts

The value of structured prompts isn't abstract. It shows up in specific workflows, across specific departments, producing specific results.

What follows isn't a list of things to try. It's a map of how organizations apply AI business prompts to real operational challenges.

Click any category below to open its full prompt library — ready-to-use prompts with usage instructions. Copy, replace the bracketed fields, and paste into ChatGPT, Claude, Gemini, or Perplexity.

You are a B2B SaaS copywriter. Write a cold outreach email to [JOB TITLE] at [COMPANY TYPE] about [PRODUCT].

Context to weave in:
- Likely pain point: [PAIN]
- Our differentiator: [DIFFERENTIATOR]
- Proof point: [CUSTOMER RESULT OR STAT]

Rules:
- Subject line under 50 characters, no clickbait, no emojis
- Body under 120 words
- Open with a specific observation about their business
- One CTA only: 15-minute call
- No "I hope this finds you well"
- No adjectives like "revolutionary" or "cutting-edge"

Output: Subject line, blank line, body, then a one-line P.S. with the proof point.

Act as a content strategist. Build a content brief for an article titled "[WORKING TITLE]".

Target audience: [AUDIENCE — role, seniority, industry]
Primary keyword: [KEYWORD]
Search intent: [informational / commercial / transactional]
Business goal: [leads / brand / SEO ranking]

Produce:
1. A one-sentence angle that differentiates this article
2. H2/H3 outline (5-8 sections) with a one-line purpose per section
3. Five questions a skeptical reader would ask, with answers
4. Internal links to suggest (topics, not URLs)
5. A 100-word meta description
6. Two CTA options

No fluff. No "in today's fast-paced world."

Create a 30-day social media calendar for [BRAND] on [PLATFORM: LinkedIn / Instagram / X].

Brand voice: [3-4 adjectives, e.g. "direct, technical, warm, no hype"]
Audience: [WHO]
Business goal: [awareness / leads / community]
Content pillars: [PILLAR 1], [PILLAR 2], [PILLAR 3], [PILLAR 4]

For each day, provide:
- Post format (text / carousel / reel / poll / link)
- Hook (first line, under 12 words)
- Core idea in one sentence
- CTA (if any)

Balance: 60% value, 30% engagement, 10% promotion. No more than 2 posts per week that promote directly.

You are an SEO content strategist. I want to rank for "[KEYWORD]".

Top-ranking pages currently cover: [PASTE 3-5 BULLET SNIPPETS FROM SERP OR SAY "unknown"]
My angle: [YOUR UNIQUE POV OR DATA]

Produce:
1. Search intent in one sentence
2. What every competitor misses
3. A differentiated H1 (under 65 characters)
4. Full H2/H3 outline that covers intent fully but adds 2 sections competitors don't have
5. Five semantic keywords to weave in naturally
6. Featured snippet opportunity — write a 45-word answer block

Avoid keyword stuffing. Write for a smart reader, not a crawler.

Write paid ad copy for [PRODUCT] running on [GOOGLE SEARCH / META / LINKEDIN].

Product: [ONE-LINE DESCRIPTION]
Primary audience: [WHO]
Core benefit: [BENEFIT]
Pain point: [PAIN]
Offer: [TRIAL / DEMO / DISCOUNT]
Landing page promise: [WHAT THEY GET]

Deliver:
- 5 Google headlines (max 30 characters each)
- 5 Google descriptions (max 90 characters each)
- 3 Meta primary texts (max 125 characters)
- 3 Meta headlines (max 40 characters)

Rules: no emojis in Google. Numbers beat adjectives. One idea per line. No "Learn more" as primary CTA unless nothing else fits.

Analyze the following content samples from [BRAND] and produce a brand voice guide.

Samples:
[PASTE 3-5 PARAGRAPHS OF REAL BRAND CONTENT]

Produce:
1. Voice in 3 adjectives, each with a one-line definition
2. What we sound like (2 examples)
3. What we don't sound like (2 counter-examples)
4. Sentence length preference and structure patterns
5. Words and phrases we use
6. Words and phrases we avoid
7. How the voice shifts by channel (blog vs. email vs. social)

Base everything on the samples. Do not invent tone traits not present in the content.

Write landing page copy for [PRODUCT/SERVICE].

Offer: [WHAT THEY GET]
Audience: [WHO + WHAT THEY CURRENTLY DO INSTEAD]
Objections: [TOP 3 OBJECTIONS]
Proof: [STATS, LOGOS, TESTIMONIALS]
Primary CTA: [ACTION]

Produce copy for each section:
1. Hero — headline (under 12 words) + subhead (under 25 words) + CTA
2. Problem section — 3 pain points in the customer's own language
3. Solution section — how it works in 3 steps, each one line
4. Proof section — 3 proof formats (stat, quote, logo wall)
5. Objection section — reframe each of the 3 objections
6. Final CTA — headline + urgency line (no fake scarcity)

Every line must earn its place. No filler.

Build a 5-email nurture sequence for leads who downloaded [LEAD MAGNET].

Audience: [WHO]
Goal: [BOOK DEMO / TRIAL SIGNUP / PURCHASE]
Sales cycle: [DAYS / WEEKS]
Common objections: [OBJECTIONS]

For each email provide:
- Send day (Day 1, Day 3, Day 7, Day 12, Day 20)
- Subject line (under 45 chars)
- Preview text (under 90 chars)
- Body (100-150 words)
- One CTA
- Purpose in the sequence (educate / build trust / handle objection / create urgency / ask)

Sequence arc: Email 1 delivers the promise. Emails 2-3 build authority. Email 4 handles the main objection. Email 5 asks directly. No "just following up."

Act as a positioning strategist. Analyze how [YOUR BRAND] is positioned against [COMPETITOR 1], [COMPETITOR 2], and [COMPETITOR 3].

For each competitor, describe:
- Their stated positioning (from their homepage headline)
- Who they appear to target
- Their apparent strength
- Their apparent weakness
- The gap they leave in the market

Then produce:
1. A positioning map with 2 axes you choose and defend
2. Where [YOUR BRAND] can own space nobody else occupies
3. Three positioning statements to test
4. Two claims we should stop making because competitors also make them

Be specific. Vague positioning analysis is useless.

Build a 6-week product launch plan for [PRODUCT].

Launch date: [DATE]
Primary audience: [WHO]
Secondary audience: [WHO]
Key differentiator: [DIFFERENTIATOR]
Assets available: [BLOG / VIDEO / DEMO / CASE STUDY / ETC]
Budget: [LOW / MEDIUM / HIGH]
Team size: [N]

Produce a week-by-week plan covering:
- Pre-launch (weeks -6 to -1): awareness and list building
- Launch week: coordinated push across owned + paid + earned
- Post-launch (weeks +1 to +2): momentum and follow-up

For each week list: channel, asset, owner, KPI. Then list the 3 things most likely to go wrong and how to prevent them.

Write a cold call script for [PRODUCT] targeting [ICP].

Context:
- Who I'm calling: [TITLE, COMPANY TYPE]
- Why now: [TRIGGER EVENT OR TIMING]
- What I know about them: [RESEARCH POINT]
- My goal for the call: [MEETING / QUALIFY / INTRO]

Produce:
1. Opening (under 15 seconds) — no "how are you today"
2. Permission ask
3. Value hypothesis (1 sentence, tied to their world, not ours)
4. Qualifying question that reveals pain
5. Three objection branches with responses:
   - "Not interested"
   - "Send me an email"
   - "We already use [COMPETITOR]"
6. Close for the next step
7. Voicemail script (under 20 seconds)

Tone: peer-to-peer. Not scripted-salesy.

Generate a discovery call question set for selling [PRODUCT] to [BUYER ROLE].

Deal context:
- Deal size: [RANGE]
- Typical buying committee: [ROLES]
- Common trigger: [TRIGGER]

Produce 15 questions organized by:
1. Current state (what they do today)
2. Pain and cost of inaction (quantified where possible)
3. Desired future state
4. Decision process and stakeholders
5. Timeline and urgency
6. Budget and approval

For each question, add a follow-up probe starting with "Tell me more about..." or "What happens if...".

Avoid: yes/no questions. Avoid: leading questions that telegraph the answer.

Create an objection handling playbook for [PRODUCT] sold to [BUYER ROLE].

For each objection below, provide:
- What the buyer is actually saying (surface)
- What they might actually mean (subtext)
- A response framework (acknowledge → reframe → evidence → question)
- One proof point to use
- What NOT to say

Objections:
1. "It's too expensive"
2. "We're already using [COMPETITOR]"
3. "We don't have time to implement"
4. "I need to talk to my team"
5. "We tried something like this before and it didn't work"
6. "Send me some information"

Keep responses under 60 words each. No scripts — frameworks.

Draft a sales proposal for [PROSPECT].

Inputs:
- Prospect company: [NAME, SIZE, INDUSTRY]
- Their stated problem: [PROBLEM IN THEIR WORDS]
- Their desired outcome: [OUTCOME]
- Our solution: [WHAT WE'LL DELIVER]
- Pricing: [RANGE OR STRUCTURE]
- Implementation timeline: [WEEKS]
- Decision criteria: [WHAT THEY SAID MATTERS]

Produce a proposal with these sections:
1. Executive summary (under 150 words, in their language)
2. Understanding of their situation
3. Proposed approach (phases, deliverables)
4. Timeline
5. Investment (options if applicable)
6. Why us (proof, not claims)
7. Next steps

Rules: no "we are a leading provider." No filler. Every section answers "so what?" for the buyer.

Write a 4-email follow-up sequence after a product demo with [PROSPECT].

From the demo, I know:
- Their top priority: [PRIORITY]
- Their main concern: [CONCERN]
- Their timeline: [TIMELINE]
- Stakeholders involved: [WHO]
- What they said they liked: [LIKED]

Emails:
1. Day 0 — recap + next steps (send within 4 hours)
2. Day 2 — address their concern with a resource or proof
3. Day 5 — share a relevant case study or metric
4. Day 10 — direct ask for decision or next meeting

Rules:
- Every email subject under 45 characters
- Every email under 100 words
- No "just checking in" or "circling back"
- Each email must add new value, not repeat

Build a detailed buyer persona for [PRODUCT] targeting [PERSONA NAME].

Inputs:
- Job title: [TITLE]
- Company size: [SIZE]
- Industry: [INDUSTRY]
- Reports to: [WHO]
- Measured on: [METRICS]

Produce:
1. A day-in-the-life narrative (150 words)
2. Their top 3 business pressures this quarter
3. What they're evaluated on and how it maps to our product
4. What they're afraid of (career risk, team pushback, change fatigue)
5. Where they get information (sources, communities, analysts)
6. The language they actually use (5 phrases)
7. How they build internal business cases
8. What disqualifies a vendor instantly

Be specific. No generic "wants to save time and money."

Act as a sales manager reviewing a deal. Analyze the following and give me a candid assessment.

Deal context:
- Company: [NAME, SIZE]
- Deal value: [VALUE]
- Stage: [STAGE]
- Champion: [WHO, HOW STRONG]
- Economic buyer: [WHO, LEVEL OF ACCESS]
- Competition: [WHO ELSE]
- Timeline: [WHEN THEY SAY THEY'LL DECIDE]
- What we know: [FACTS]
- What we don't know: [GAPS]

Produce:
1. Honest probability of close (0-100%) with reasoning
2. The single biggest risk
3. The single biggest opportunity
4. Three questions I should ask in the next call
5. Whether I should invest more time or walk

Don't be polite. Be useful.

Build a battle card for selling against [COMPETITOR].

Our product: [OURS]
Their product: [THEIRS]

Produce:
1. How they position themselves (their homepage headline verbatim)
2. Their 3 strongest claims
3. Where each claim is vulnerable
4. Our 3 strongest differentiators (only if real)
5. Where they win legitimately — be honest
6. Three landmine questions to ask in a deal where they're present
7. Two customer stories where we replaced them
8. What NOT to say about them

One page. No marketing language. Assume the reader is a rep who needs this in 60 seconds.

Write a renewal + upsell pitch for existing customer [NAME].

Account context:
- Current plan: [PLAN]
- Usage: [USAGE METRICS]
- Time as customer: [MONTHS/YEARS]
- Champion: [WHO]
- What they said last review: [FEEDBACK]
- Where they've outgrown current plan: [SPECIFIC]
- Upsell target: [NEXT TIER / ADD-ON]

Produce:
1. A one-paragraph value recap (their results, not our features)
2. The natural upgrade narrative tied to their growth
3. Three ROI points specific to the upgrade
4. Two objection responses (price and change management)
5. A proposed renewal call agenda

Do not start with "We're excited to..." Start with their outcome.

Summarize this week's pipeline for my Monday review.

Deals:
[PASTE LIST: Deal name, stage, value, next step, last activity, risk]

Produce:
1. Total pipeline by stage
2. Deals that moved forward this week (with reason)
3. Deals that stalled (with reason and next action)
4. Top 3 deals to focus on this week and why
5. Deals at risk — recommended action for each
6. Forecast: commit / best case / pipeline

Keep it to one page. Use a table. No narrative fluff.

Produce a monthly variance analysis report for [COMPANY / DEPARTMENT].

Inputs:
- Actuals: [PASTE LINE ITEMS WITH ACTUAL NUMBERS]
- Budget: [PASTE BUDGET NUMBERS]
- Prior month: [PRIOR MONTH NUMBERS]
- Materiality threshold: [e.g. 5% or $50,000]

For each line with a variance above threshold:
1. State the variance in dollars and percentage
2. Identify the top 2-3 operational drivers (not accounting artifacts)
3. Classify as timing, structural, or one-time
4. State whether it recurs next month
5. Recommend action if structural

Format: Table with columns Line, Actual, Budget, Variance $, Variance %, Drivers, Classification, Action.

Do not explain variances as "due to timing" without specifying what timing.

Write the financial narrative section for a board presentation.

Inputs:
- Period: [Q / MONTH]
- Revenue: [ACTUAL VS PLAN VS PRIOR]
- Gross margin: [%]
- EBITDA: [ACTUAL VS PLAN]
- Cash: [BALANCE + RUNWAY]
- Key wins this period: [WINS]
- Key misses: [MISSES]
- Forward outlook: [GUIDANCE]

Produce:
1. Headline: one sentence a board member will remember
2. Performance vs plan — what happened, why, in plain English
3. Two things going well, with numbers
4. Two things not going well, with honest explanation and remediation
5. Forward view — 3 numbers that matter next quarter
6. Risk table: 3 risks, likelihood, impact, mitigation

Rules: no hedging language. No "slightly below." Use exact figures. Board members read fast.

Build a budgeting framework for [COMPANY] for [FISCAL YEAR].

Context:
- Revenue last year: [AMOUNT]
- Growth target: [%]
- Headcount now: [N]
- Planned hires: [N + WHEN]
- Major initiatives: [LIST]
- Cost structure: [FIXED VS VARIABLE SPLIT]

Produce:
1. Revenue build-up by month (assumptions stated)
2. Headcount plan by quarter with fully-loaded cost
3. Department opex budget by category
4. Cash flow projection with minimum cash floor
5. Three sensitivity scenarios: base, downside, upside
6. Top 5 assumptions that matter most

State every assumption explicitly. If you don't know, list as a driver to be confirmed.

Write a monthly investor update email for [COMPANY].

Inputs:
- Period: [MONTH]
- Highlights: [3 WINS WITH NUMBERS]
- Lowlights: [2 HONEST MISSES]
- Key metrics: [MRR, BURN, RUNWAY, GROWTH]
- Asks: [INTROS / HIRES / ADVICE]

Structure:
1. Subject line: "[Company] — [Month] Update"
2. One-paragraph summary (numbers in first two sentences)
3. Highlights (bulleted, quantified)
4. Lowlights (honest, with what we're doing)
5. Metrics table
6. Asks (specific, easy to action)

Tone: direct and confident, not promotional. Investors read this for signal, not spin.

Create an audit preparation checklist for [COMPANY] ahead of [AUDIT TYPE: external / internal / regulatory].

Inputs:
- Industry: [INDUSTRY]
- Auditor: [FIRM OR "internal team"]
- Period under audit: [DATES]
- Prior audit findings: [ISSUES]

Produce:
1. Document request list, grouped by statement line
2. Accounts with highest inherent risk in our industry
3. Five areas most likely to get flagged this year
4. Owner assignments (by role, not name)
5. Timeline: what's due 6 weeks before, 3 weeks before, 1 week before
6. Common last-minute requests to prepare for

Include a risk-rated priority column (High/Medium/Low) for each item.

Analyze cash flow for [COMPANY] and forecast the next 13 weeks.

Inputs:
- Current cash: [AMOUNT]
- AR aging: [BUCKETS WITH AMOUNTS]
- AP due next 13 weeks: [AMOUNTS BY WEEK]
- Payroll: [WEEKLY/MONTHLY AMOUNT]
- Committed outflows: [LIST]
- Expected inflows (weighted): [LIST]

Produce:
1. 13-week cash flow table
2. Minimum cash position and when
3. Sensitivity: what if AR is 15% slower
4. Top 3 actions to protect cash
5. Trigger points for line of credit

Flag any week where cash drops below [MINIMUM FLOOR].

Build a cost reduction proposal for [COMPANY].

Context:
- Target reduction: [AMOUNT OR %]
- Current spend categories: [LIST WITH AMOUNTS]
- Constraints: [NO LAYOFFS / PRESERVE R&D / ETC]
- Timeline: [WHEN SAVINGS NEED TO HIT]

Produce:
1. Opportunity map — every category with realistic savings range
2. Quick wins (0-60 days, low risk)
3. Structural changes (3-12 months, higher impact)
4. Tradeoffs for each major cut — what capability is lost
5. Implementation sequence
6. Risks and how to mitigate

For each item: savings estimate, effort, time-to-impact, risk level.

Do not propose cuts that only look good on paper. Show the second-order effect.

Analyze pricing strategy for [PRODUCT / SERVICE].

Inputs:
- Current pricing: [TIERS AND PRICES]
- Cost per unit / delivery: [COST]
- Customer segments: [SEGMENTS]
- Competitor pricing: [COMPETITOR PRICES]
- Willingness-to-pay signal: [ANY RESEARCH]
- Churn at current price: [%]
- Conversion at current price: [%]

Produce:
1. Margin analysis by segment
2. Where we're underpriced (evidence required)
3. Where we're overpriced (evidence required)
4. Three pricing model alternatives (usage, tiered, flat)
5. Recommended change with expected impact
6. Risks of the change and mitigation

Be candid. "Raise prices" is easy to say and hard to do — show the path.

Design a capital allocation framework for [COMPANY].

Inputs:
- Available capital: [AMOUNT]
- Competing uses: [LIST WITH ASK AMOUNTS]
- Strategic priorities: [TOP 3]
- Risk tolerance: [LOW / MEDIUM / HIGH]
- Time horizon: [SHORT / MEDIUM / LONG]

Produce:
1. Evaluation criteria with weights
2. Score each use of capital against criteria
3. Recommended allocation with rationale
4. What we're explicitly NOT funding and why
5. Decision review cadence

Format as a scoring table plus one page of rationale.

Create a financial due diligence checklist for evaluating [TARGET COMPANY].

Inputs:
- Target industry: [INDUSTRY]
- Deal size: [RANGE]
- Strategic rationale: [WHY WE'RE INTERESTED]
- Deal structure: [ASSET / STOCK / MERGER]

Produce:
1. Financial diligence request list (grouped by statement line)
2. Quality of earnings focus areas
3. Working capital normalization items
4. Hidden liabilities to look for
5. Customer concentration analysis
6. Integration cost estimate framework
7. Red flags to escalate immediately

Include a table with columns: Item, Why it matters, Owner, Status.

Write a job description for [ROLE] at [COMPANY].

Inputs:
- Level: [JUNIOR / MID / SENIOR / LEAD]
- Location: [REMOTE / HYBRID / ON-SITE + CITY]
- Comp range: [RANGE]
- Reports to: [TITLE]
- Team size: [N]
- Must-have skills: [LIST]
- Nice-to-have: [LIST]

Produce:
1. Role summary (3 sentences, no buzzwords)
2. What you'll actually do (5 bullets starting with a verb)
3. What success looks like at 30 / 60 / 90 days
4. Required qualifications (only what's truly required)
5. Preferred qualifications
6. About the team and company

Rules:
- No "fast-paced environment" or "wear many hats"
- No "rockstar" or "ninja"
- Use "you" not "the candidate"
- List required vs preferred honestly — long required lists shrink the applicant pool for no reason

Draft a performance review for [EMPLOYEE NAME / ROLE].

Inputs:
- Review period: [DATES]
- Strengths observed: [SPECIFIC EXAMPLES]
- Areas for growth: [SPECIFIC EXAMPLES]
- Goals from last review: [GOALS + OUTCOMES]
- Peer feedback themes: [THEMES]
- Next-period goals: [GOALS]

Produce:
1. Overall summary (3 sentences, honest)
2. What's going well, with 2 specific examples
3. Where to focus next, with 2 specific examples
4. Rating justification if a rating is used
5. 3 SMART goals for the next period
6. Development suggestions

Rules:
- Every claim backed by an observable example
- No vague praise or vague criticism
- No personality judgments — behaviors and outcomes only
- Language that would survive being read by the employee and their lawyer

Draft an HR policy on [TOPIC: remote work / PTO / expenses / etc.] for [COMPANY].

Inputs:
- Jurisdiction: [COUNTRY / STATE]
- Company size: [N]
- Company culture: [3 ADJECTIVES]
- Legal constraints to respect: [ANY]
- Current practice: [WHAT WE DO TODAY]

Produce:
1. Purpose
2. Scope (who this applies to)
3. Policy statement
4. Specific rules and procedures
5. Exceptions process
6. Roles and responsibilities
7. Effective date and review cycle

Rules:
- Plain language, 8th-grade reading level
- Active voice
- No legalese unless quoting statute
- Flag any clause that should be reviewed by employment counsel

Build a 30-60-90 day onboarding plan for a new [ROLE] at [COMPANY].

Inputs:
- Team: [TEAM]
- Reports to: [TITLE]
- Key deliverables in first 6 months: [LIST]
- Key stakeholders: [LIST]
- Tools they'll use: [LIST]

Produce:
Day 1-30 (Learn):
- People to meet, questions to ask, materials to read
- Success criteria for Day 30

Day 31-60 (Contribute):
- First deliverables
- Stakeholder relationships to deepen
- Success criteria for Day 60

Day 61-90 (Own):
- Full ownership areas
- First measurable outcomes
- Success criteria for Day 90

Include a weekly check-in template and a list of common early mistakes to avoid in this role.

Help me prepare for a difficult feedback conversation with [TEAM MEMBER ROLE].

Situation:
- What happened: [FACTS]
- Impact: [BUSINESS OR TEAM IMPACT]
- Pattern: [ONE-OFF OR REPEATED]
- Prior conversations: [WHAT WAS SAID]
- Desired outcome: [WHAT SHOULD CHANGE]

Produce:
1. Opening line (direct, not softened)
2. The observation (facts only, no interpretation)
3. The impact (specific, business-relevant)
4. The ask (specific behavior change)
5. Two responses to likely pushback
6. How to close with commitment on next steps

Tone: direct and respectful, not aggressive. Do not use sandwich feedback. Do not use "I feel" statements for behavior correction.

Generate exit interview questions for a departing [ROLE] at [COMPANY].

Context:
- Tenure: [DURATION]
- Reason stated: [REASON]
- Team: [TEAM]
- Voluntary or involuntary: [TYPE]

Produce questions grouped by:
1. Role and work (what they did, what they'd change)
2. Manager and team (without leading)
3. Company and culture
4. Compensation and growth
5. What would have made them stay
6. Advice for the next person in their seat

For each question: what we're trying to learn, and what signals to listen for.

Avoid yes/no questions. Avoid questions where the answer is already known.

Build a compensation benchmarking brief for [ROLE] at [COMPANY].

Inputs:
- Role: [TITLE]
- Level: [LEVEL]
- Location: [CITY / REGION]
- Company stage: [SEED / SERIES / PUBLIC]
- Industry: [INDUSTRY]
- Current comp: [BASE + EQUITY + BONUS]

Produce:
1. Market range for base (25th, 50th, 75th percentile)
2. Typical equity structure at this level
3. Bonus structure common in this market
4. Total comp positioning
5. Where we're competitive, where we're not
6. Recommendation with reasoning

Flag: sources and assumptions. This is directional, not a substitute for a paid survey.

Design a training program on [TOPIC] for [AUDIENCE].

Inputs:
- Audience: [ROLE + SENIORITY]
- Business goal: [OUTCOME]
- Current skill level: [BASELINE]
- Format: [SELF-PACED / COHORT / LIVE / BLENDED]
- Duration available: [HOURS / WEEKS]
- Budget: [LOW / MEDIUM / HIGH]

Produce:
1. Learning objectives (measurable)
2. Module structure with timing
3. Delivery methods per module
4. Assessment approach
5. Reinforcement plan (post-training)
6. Success metrics

Rules: outcomes-first. No content for content's sake. Every module maps to a learning objective.

Draft a Performance Improvement Plan for [ROLE].

Inputs:
- Employee role: [ROLE]
- Performance gaps: [SPECIFIC, OBSERVABLE]
- Prior feedback: [WHAT WAS SAID, WHEN]
- Expected standards: [MEASURABLE STANDARDS]
- PIP duration: [30 / 60 / 90 days]
- Support offered: [TRAINING, MENTORING, ETC]

Produce:
1. Purpose of the PIP
2. Specific performance gaps (facts, not opinions)
3. Expected performance standards (measurable)
4. Support and resources provided
5. Check-in schedule
6. Consequences of not meeting standards
7. Signature block for employee and manager

Language must be factual, documented, and legally defensible. No speculation. No personality commentary.

Design an employee engagement survey for [COMPANY].

Inputs:
- Company size: [N]
- Industry: [INDUSTRY]
- Known pain points: [ISSUES]
- Prior survey results: [THEMES]
- Goal: [DIAGNOSE / TRACK / MEASURE INTERVENTION]

Produce:
1. 25 questions grouped by theme (manager, growth, comp, workload, purpose, belonging)
2. Likert scale format recommendations
3. 3 open-text questions
4. Anonymity and privacy protections
5. Distribution plan
6. Post-survey action communication template

Rules: no double-barreled questions. No leading questions. Keep to 15 minutes max.

Write an SOP for [PROCESS NAME].

Inputs:
- Process owner: [ROLE]
- Frequency: [DAILY / WEEKLY / AD HOC]
- Trigger: [WHAT STARTS THE PROCESS]
- Systems used: [TOOLS]
- Compliance requirements: [IF ANY]
- Known failure modes: [ISSUES]

Structure:
1. Purpose
2. Scope
3. Roles and responsibilities (RACI)
4. Step-by-step procedure (numbered, imperative voice)
5. Decision points with criteria
6. Exceptions and escalation
7. Records and retention
8. Revision history

Each step must be actionable by someone new to the role. Include what "done" looks like at each step.

Analyze this process and propose optimizations.

Current process:
[DESCRIBE STEPS OR PASTE PROCESS]

Inputs:
- Volume: [TRANSACTIONS PER PERIOD]
- Cycle time: [CURRENT]
- Error rate: [%]
- Team size: [N]
- Cost per transaction: [AMOUNT]

Produce:
1. Value stream map — step, time, owner, value-add (yes/no)
2. Bottlenecks and wait states
3. Failure modes and their cost
4. Automation candidates, ranked by ROI
5. Recommended future-state flow
6. Expected impact: time, cost, quality
7. Implementation sequence

Quantify everything. "Improve efficiency" is not an outcome — "reduce cycle time from 4 days to 1" is.

Build a project plan for [PROJECT].

Inputs:
- Objective: [WHAT SUCCESS LOOKS LIKE]
- Deadline: [DATE]
- Budget: [AMOUNT]
- Team: [ROLES + AVAILABILITY]
- Dependencies: [EXTERNAL FACTORS]
- Known risks: [RISKS]
- Stakeholders: [WHO + INTEREST]

Produce:
1. Work breakdown structure (WBS)
2. Task list with owner, duration, dependencies
3. Critical path
4. Milestones with dates
5. Risk register (probability, impact, mitigation)
6. Communication plan
7. RACI matrix
8. Definition of done

Format the WBS as a table. Sequence phases: initiation, planning, execution, monitoring, closure.

Build a supplier evaluation scorecard for [CATEGORY].

Inputs:
- Category: [WHAT WE'RE BUYING]
- Annual spend: [AMOUNT]
- Criticality: [LOW / MEDIUM / HIGH]
- Current suppliers: [N]
- Business requirements: [QUALITY, LEAD TIME, ETC]

Produce:
1. Evaluation criteria with weights (quality, cost, delivery, service, risk, ESG)
2. Scoring rubric (1-5) with definitions per level
3. Required documentation per criterion
4. Site visit checklist
5. Reference check questions
6. Red flags to disqualify immediately
7. Recommendation format for leadership

Format as a scoring sheet leaders can review in 10 minutes.

Create a quality management checklist for [PROCESS / PRODUCT].

Inputs:
- Industry standard: [ISO / FDA / INTERNAL]
- Product / service: [WHAT]
- Customer requirements: [SPECS]
- Regulatory requirements: [RULES]
- Current quality metrics: [NUMBERS]

Produce:
1. Quality objectives (measurable)
2. Critical-to-quality (CTQ) characteristics
3. Inspection points and methods
4. Acceptance criteria per inspection
5. Non-conformance handling procedure
6. Corrective and preventive action (CAPA) process
7. Documentation requirements
8. Review cadence

Include a table: Inspection point, Method, Criteria, Owner, Frequency.

Document the workflow for [PROCESS].

Inputs:
- Trigger: [WHAT STARTS IT]
- End state: [WHAT DONE LOOKS LIKE]
- Systems: [TOOLS]
- Roles involved: [ROLES]
- Volume: [PER PERIOD]

Produce:
1. Swimlane description by role
2. Step-by-step flow with decision points
3. Inputs and outputs per step
4. Handoff points (where work transfers between roles)
5. Time per step (if known)
6. Automated vs manual steps
7. Pain points and improvement opportunities

Format as a numbered flow. Include "if X then Y" branches explicitly.

Build a capacity plan for [TEAM / FUNCTION].

Inputs:
- Current headcount: [N] with roles
- Current volume: [UNITS PER PERIOD]
- Growth forecast: [% OR UNITS]
- Productivity per person: [RATE]
- Absenteeism: [%]
- Ramp time for new hires: [WEEKS]

Produce:
1. Current capacity vs demand
2. Gap or surplus by month for next 12 months
3. Hiring or reduction plan
4. Ramp schedule for new hires
5. Overtime assumptions
6. Cost impact
7. Trigger points — when to hire, when to pause

Format as a monthly table: Demand, Capacity, Gap, Action.

Write an incident response plan for [INCIDENT TYPE].

Inputs:
- Risk scenario: [WHAT COULD HAPPEN]
- Business impact: [WHAT'S AFFECTED]
- Regulatory notification: [YES / NO + TIMELINE]
- Team available: [ROLES]
- Systems affected: [LIST]

Produce:
1. Severity classification (Sev 1-4)
2. Detection methods
3. First-response checklist (first 30 minutes)
4. Escalation matrix (who to call, when)
5. Internal communication templates
6. External communication templates
7. Recovery procedures
8. Post-incident review process

Include a "do not" list — actions that make things worse.

Write a continuous improvement proposal for [PROCESS].

Inputs:
- Current state: [HOW IT WORKS TODAY]
- Pain points: [WHAT'S BROKEN]
- Proposed change: [WHAT YOU WANT TO DO]
- Expected benefit: [TIME / COST / QUALITY]
- Investment required: [COST + EFFORT]
- Risks: [WHAT COULD GO WRONG]

Produce:
1. Problem statement (quantified)
2. Root cause analysis (5 whys or fishbone)
3. Proposed solution
4. Alternatives considered and rejected
5. Cost-benefit analysis
6. Implementation plan
7. Success metrics
8. Risks and mitigation

Format: one page, decision-ready. No filler.

Design a KPI dashboard for [FUNCTION / TEAM].

Inputs:
- Audience: [EXECUTIVE / MANAGER / TEAM]
- Business objective: [WHAT WE'RE OPTIMIZING]
- Current metrics tracked: [LIST]
- Data sources: [SYSTEMS]
- Review cadence: [DAILY / WEEKLY / MONTHLY]

Produce:
1. Top 5-7 KPIs (not 20)
2. Leading vs lagging split
3. Definition of each KPI (formula, source, owner)
4. Target and threshold for each
5. Visualization recommendation per KPI
6. Alert thresholds
7. What to do when each KPI goes off-track

Rules: if a KPI can't drive a decision, drop it. Every metric must have an owner and a response playbook.

Write technical documentation for [SYSTEM / FEATURE / API].

Inputs:
- Audience: [DEVELOPERS / OPS / END USERS]
- Purpose: [SETUP / REFERENCE / TROUBLESHOOTING]
- Prerequisites: [LIST]
- Key concepts: [LIST]

Produce:
1. Overview (what this is, when to use it)
2. Prerequisites and setup
3. Quickstart (working example in under 5 minutes)
4. Core concepts
5. Common tasks (with code or commands)
6. Troubleshooting (top 5 issues)
7. Reference links

Rules:
- Every code sample must be complete and runnable
- No "simply" or "just" — those words hide complexity
- Assume the reader is competent but unfamiliar

Write an Architecture Decision Record for [DECISION].

Inputs:
- Decision: [WHAT WE DECIDED]
- Context: [WHY THIS CAME UP]
- Options considered: [LIST]
- Chosen option: [WHICH ONE]
- Tradeoffs: [WHAT WE'RE GIVING UP]
- Stakeholders: [WHO]

Format:
1. Title and status (proposed / accepted / superseded)
2. Context — the forces at play
3. Decision — what we chose
4. Consequences — positive, negative, neutral
5. Alternatives considered and why rejected
6. References

Rules: concise. One page. Written for a future engineer who wasn't in the room.

Write a blameless post-mortem for [INCIDENT].

Inputs:
- What happened: [TIMELINE OF EVENTS]
- Impact: [USERS AFFECTED, DOWNTIME, DATA LOSS]
- Root cause: [TECHNICAL CAUSE]
- Detection: [HOW WE FOUND OUT]
- Response: [WHAT WE DID]
- Resolution: [HOW IT WAS FIXED]

Produce:
1. Summary (3 sentences)
2. Impact (quantified)
3. Timeline (UTC timestamps)
4. Root cause (5 whys)
5. Contributing factors
6. Detection analysis — why didn't we catch this sooner
7. Response analysis — what went well, what didn't
8. Action items with owners and due dates
9. Lessons learned

Blameless: focus on systems, not people. "Human error" is not a root cause.

Build a migration plan for moving from [OLD SYSTEM] to [NEW SYSTEM].

Inputs:
- Systems: [FROM → TO]
- Data volume: [SIZE]
- Downtime tolerance: [ZERO / HOURS / DAYS]
- Users affected: [N]
- Dependencies: [UPSTREAM / DOWNSTREAM SYSTEMS]
- Compliance constraints: [IF ANY]

Produce:
1. Migration phases (assessment, planning, pilot, cutover, validation)
2. Data mapping requirements
3. Rollback plan
4. Rollback triggers — when to abort
5. Communication plan
6. Testing strategy
7. Cutover runbook
8. Success criteria
9. Post-migration monitoring

Include a "what could go wrong" list with mitigations.

Write API documentation for [ENDPOINT / SERVICE].

Inputs:
- Purpose: [WHAT IT DOES]
- Authentication: [METHOD]
- Rate limits: [LIMITS]
- Endpoints: [LIST]
- Request format: [SCHEMA]
- Response format: [SCHEMA]

For each endpoint produce:
1. Method and path
2. Purpose (one line)
3. Authentication required (yes/no)
4. Request parameters (name, type, required, description)
5. Request example (curl + JSON)
6. Response example (success)
7. Error responses (codes + meanings)
8. Rate limit notes

Rules: every parameter documented. Every error code explained. Runnable examples.

Write a runbook for [OPERATIONAL TASK / ALERT].

Inputs:
- Trigger: [ALERT / SCHEDULE / REQUEST]
- Systems involved: [LIST]
- Prerequisites: [ACCESS, TOOLS]
- Expected outcome: [WHAT DONE LOOKS LIKE]
- Rollback: [HOW TO UNDO]

Produce:
1. Purpose
2. When this runbook applies
3. Prerequisites
4. Step-by-step procedure (numbered, with exact commands)
5. Verification — how to confirm success
6. Troubleshooting (common failures)
7. Rollback procedure
8. Escalation (who to contact, when)

Every command should be copy-pasteable. No "log into the server" — specify which server, which user.

Structure a security audit report for [SYSTEM / APPLICATION].

Inputs:
- Scope: [WHAT'S AUDITED]
- Standards: [OWASP / SOC2 / ISO 27001]
- Findings: [VULNERABILITIES FOUND]
- Severity ratings: [CRITICAL / HIGH / MEDIUM / LOW]

Produce:
1. Executive summary (for non-technical readers)
2. Scope and methodology
3. Findings table: ID, Description, Severity, Evidence, Remediation, Owner
4. Risk rating rationale
5. Remediation roadmap (30 / 60 / 90 days)
6. Positive controls already in place
7. Next audit cadence

Rules: severity must be defensible. "Critical" is not a matter of opinion — justify each rating.

Write user stories for [FEATURE].

Inputs:
- Feature: [WHAT WE'RE BUILDING]
- User type: [WHO]
- Business outcome: [WHY]
- Constraints: [TECHNICAL / REGULATORY]
- Known edge cases: [LIST]

For each story produce:
1. Story format: "As a [user], I want [action], so that [outcome]"
2. Acceptance criteria (Given / When / Then)
3. Edge cases to consider
4. Out of scope
5. Dependencies
6. Definition of done

Rules:
- One story per user value, not per technical task
- Acceptance criteria must be testable
- No "user-friendly" or "fast" without a threshold

Assess technical debt in [SYSTEM / CODEBASE].

Inputs:
- System age: [YEARS]
- Team size: [N]
- Recent incidents: [LIST]
- Development velocity trend: [FASTER / SLOWER / FLAT]
- Known problem areas: [MODULES / SERVICES]

Produce:
1. Debt categories (code, architecture, test, documentation, infrastructure)
2. For each: examples, business impact, remediation effort
3. Prioritization matrix (impact vs effort)
4. Recommended sequence
5. Estimated cost of inaction
6. Quick wins (under 1 week)
7. Long-term projects

Quantify where possible. "Poor code quality" is not an assessment — "the payments module takes 3x longer to modify than peer modules" is.

Create a deployment checklist for releasing [VERSION / FEATURE].

Inputs:
- Release type: [MAJOR / MINOR / PATCH]
- Systems affected: [LIST]
- Downtime expected: [YES / NO + DURATION]
- Rollback plan: [HOW]
- Stakeholders to notify: [WHO]

Produce:
Pre-deployment:
- Code freeze confirmation
- Test coverage thresholds met
- Security scan passed
- Database migrations tested
- Rollback tested
- Stakeholders notified

Deployment:
- Step-by-step with verification between steps
- Exact commands
- Success signals
- Abort signals

Post-deployment:
- Smoke tests
- Monitoring dashboards to check
- Metrics thresholds
- Rollback criteria
- Communication

Include a "who to call if X fails" section.

Build a contract review checklist for [CONTRACT TYPE: NDA / MSA / SaaS / etc.].

Inputs:
- Our position: [CUSTOMER / VENDOR / PARTNER]
- Jurisdiction: [COUNTRY / STATE]
- Contract value: [AMOUNT]
- Risk tolerance: [LOW / MEDIUM / HIGH]
- Non-negotiables: [CLAUSES]

Produce a clause-by-clause review guide covering:
1. Parties and definitions
2. Scope of services / goods
3. Payment terms and invoicing
4. Term and termination
5. Liability and indemnification
6. IP ownership
7. Confidentiality
8. Data protection / privacy
9. Warranties and disclaimers
10. Force majeure
11. Governing law and disputes
12. Assignment and change of control

For each: what to look for, common red flags, our preferred position, fallback positions.

Flag any clause that requires escalation to outside counsel.

Draft an internal policy on [TOPIC].

Inputs:
- Jurisdiction: [COUNTRY / STATE]
- Company size: [N]
- Industry: [INDUSTRY]
- Regulatory requirements: [RULES]
- Current practice: [WHAT WE DO NOW]

Structure:
1. Purpose and scope
2. Definitions
3. Policy statement
4. Roles and responsibilities
5. Procedures
6. Exceptions and approvals
7. Non-compliance consequences
8. Related policies
9. Review cycle
10. Effective date and version

Language: plain, readable, active voice. Flag any clause that should be reviewed by counsel for jurisdictional fit.

Summarize regulatory changes affecting [INDUSTRY] in [JURISDICTION] for [PERIOD].

Inputs:
- Regulatory bodies: [LIST]
- Regulations of interest: [LIST]
- Business impact areas: [OPERATIONS AFFECTED]

For each change:
1. Regulation name and reference
2. Effective date
3. What changed (plain English)
4. Who it affects in our business
5. Compliance action required
6. Deadline
7. Risk if we miss it
8. Owner (by role)

Format as a table. Prioritize by urgency and impact.

Write a legal risk assessment memo on [ISSUE].

Inputs:
- Situation: [FACTS]
- Jurisdiction: [WHERE]
- Stakeholders: [WHO]
- Potential outcomes: [WHAT COULD HAPPEN]
- Timeline: [WHEN IT MATTERS]

Structure:
1. Executive summary (3 sentences for a non-lawyer)
2. Facts
3. Legal framework (relevant statutes / case law themes)
4. Analysis — apply law to facts
5. Risk assessment: likelihood x impact
6. Mitigation options
7. Recommendation
8. Open questions for counsel

Rules: distinguish legal analysis from business judgment. Flag where we need outside counsel.

Draft a privacy policy outline for [PRODUCT / SERVICE].

Inputs:
- Jurisdictions: [GDPR / CCPA / DPDP / OTHER]
- Data collected: [CATEGORIES]
- Purpose of processing: [WHY]
- Third parties: [WHO RECEIVES DATA]
- Retention: [HOW LONG]
- User rights honored: [LIST]

Produce sections:
1. Introduction and scope
2. Data we collect
3. How we use data
4. Legal basis (GDPR)
5. Sharing and disclosure
6. International transfers
7. Retention
8. Security
9. Your rights
10. Cookies and tracking
11. Children's privacy
12. Changes to this policy
13. Contact

Flag sections requiring jurisdiction-specific language (DPDP India, CCPA California, GDPR EU).

Draft a mutual NDA between [PARTY A] and [PARTY B].

Inputs:
- Purpose: [WHY THEY'RE TALKING]
- Jurisdiction: [GOVERNING LAW]
- Duration: [YEARS]
- Definition of confidential info: [BROAD / NARROW]
- Permitted disclosures: [CARVE-OUTS]
- Remedies: [INJUNCTION / DAMAGES]

Sections:
1. Parties
2. Purpose
3. Definition of Confidential Information
4. Obligations of receiving party
5. Exclusions (public domain, prior knowledge, etc.)
6. Term
7. Return / destruction
8. No license
9. No warranty
10. Remedies
11. Governing law
12. Entire agreement

Standard mutual NDA. Flag any clause likely to be negotiated.

Build a compliance audit preparation plan for [STANDARD / REGULATION].

Inputs:
- Framework: [ISO / SOC2 / HIPAA / GDPR / etc.]
- Audit type: [INTERNAL / EXTERNAL / CERTIFICATION]
- Timeline: [WEEKS / MONTHS]
- Current state: [MATURE / DEVELOPING / EARLY]
- Known gaps: [ISSUES]

Produce:
1. Controls matrix: Control ID, Requirement, Current state, Evidence needed, Owner, Status
2. Gap analysis
3. Remediation plan with timeline
4. Evidence collection checklist
5. Auditor Q&A preparation
6. Walkthrough scripts
7. Pre-audit self-test

Format as a working spreadsheet structure. Assign real owners.

Summarize the following Terms of Service in plain English.

[PASTE TERMS OR KEY SECTIONS]

Produce:
1. What you're agreeing to (3 sentences)
2. What the company can do with your data
3. What you can and cannot do
4. What happens if you breach
5. How disputes are resolved
6. How the terms can change
7. Anything unusual or noteworthy

Rules: no legal jargon. Write for a smart non-lawyer. Flag sections that would concern a typical user.

Structure a legal research memo on [QUESTION].

Inputs:
- Jurisdiction: [COUNTRY / STATE]
- Legal question: [PRECISE QUESTION]
- Facts: [RELEVANT FACTS]
- Business context: [WHY THIS MATTERS]

Produce:
1. Question presented
2. Brief answer
3. Facts
4. Discussion — statutes, regulations, case law themes
5. Analysis applied to our facts
6. Counterarguments
7. Conclusion
8. Open questions for counsel

Rules: cite general legal principles, not specific cases (verify everything with counsel). Flag uncertainty honestly.

Develop an IP protection strategy for [COMPANY / PRODUCT].

Inputs:
- What we've created: [IP ASSETS — code, brand, content, inventions]
- Key markets: [JURISDICTIONS]
- Budget: [LOW / MEDIUM / HIGH]
- Business model: [HOW WE MONETIZE]
- Competitors: [WHO]

Produce:
1. IP inventory (what we have, what's protected, what's not)
2. Protection options per asset (patent, trademark, copyright, trade secret)
3. Cost and timeline per option
4. Priority sequence
5. What to protect vs. what to publish
6. Employee and contractor IP assignment requirements
7. Monitoring and enforcement plan

Format as a portfolio table. Recommend what to file this year vs. defer.

Draft a response to this customer message.

Customer message:
[PASTE MESSAGE]

Context:
- Customer: [NAME, TIER, TENURE]
- Issue category: [BILLING / BUG / HOW-TO / COMPLAINT]
- Resolution available: [YES — WHAT / NO — WHY]
- Brand voice: [3 ADJECTIVES]
- SLA: [TIMEFRAME]

Produce:
1. Subject line (if email)
2. Greeting
3. Acknowledgment of their specific issue (repeat their words)
4. Answer or resolution
5. What happens next and by when
6. Offer of further help
7. Sign-off

Rules:
- No "I understand your frustration" clichés
- Match their formality level
- Take ownership — no "the system did this"
- Never promise what we can't deliver

Write an escalation protocol for [SUPPORT TEAM].

Inputs:
- Support tiers: [TIER 1 / 2 / 3]
- Escalation triggers: [WHAT CAUSES ESCALATION]
- Stakeholders: [ROLES]
- SLA per tier: [TIMEFRAMES]
- Communication channels: [EMAIL / SLACK / PHONE]

Produce:
1. Escalation criteria by severity
2. Who escalates to whom and when
3. Information required at escalation
4. Communication templates (internal)
5. Customer communication per stage
6. Escalation closure criteria
7. Post-escalation review

Include a matrix: Severity, Response time, Owner, Escalation path, Communication cadence.

Write a knowledge base article on [TOPIC].

Inputs:
- Topic: [WHAT IT ANSWERS]
- Audience: [END USER / ADMIN / DEVELOPER]
- Trigger: [WHY USERS ASK THIS]
- Steps required: [PROCESS]
- Common mistakes: [WHAT GOES WRONG]

Structure:
1. Title (search-friendly, describes the outcome)
2. Summary (one sentence)
3. Applies to (versions, plans, roles)
4. Prerequisites
5. Step-by-step instructions with screenshots placeholders
6. Verification (how to know it worked)
7. Troubleshooting
8. Related articles
9. Feedback prompt

Rules:
- Written for search, not for narrative
- One task per article
- Use "you" not "the user"
- Number the steps

Analyze this batch of customer feedback.

Feedback:
[PASTE FEEDBACK — SURVEYS, TICKETS, REVIEWS]

Context:
- Product: [WHAT]
- Time period: [DATES]
- Volume: [N RESPONSES]

Produce:
1. Top themes (with frequency counts)
2. Sentiment by theme
3. The 3 most urgent issues (severity × frequency)
4. The 3 most repeated requests
5. What we're doing well (per customers, not per us)
6. Direct quotes worth sharing with the team
7. Recommended actions with owners
8. What we should NOT act on (noise, not signal)

Rules: quantify. Separate loud from important. Don't average away the outliers.

Draft a response to this customer complaint.

Complaint:
[PASTE COMPLAINT]

Context:
- What went wrong: [ROOT CAUSE]
- What we're doing to fix it: [REMEDIATION]
- Compensation offered: [IF ANY]
- Customer history: [VALUE, TENURE]
- Severity: [LOW / MEDIUM / HIGH]

Produce:
1. Direct acknowledgment — no deflection
2. Apology that is specific, not generic
3. Explanation — what happened, briefly
4. Remediation — what we're doing
5. What we're doing to prevent recurrence
6. Compensation (if applicable)
7. Follow-up commitment with a name and date

Rules: no "we're sorry for any inconvenience." Be specific about what we're sorry for. Never blame the customer.

Design a chatbot conversation flow for [USE CASE].

Inputs:
- Purpose: [WHAT THE BOT HANDLES]
- User intent: [WHAT USERS WANT]
- Fallback: [WHEN TO HAND OFF TO HUMAN]
- Brand voice: [3 ADJECTIVES]
- Common questions: [TOP 5]

Produce:
1. Welcome message
2. Main menu options
3. Branch flows per intent
4. Qualifying questions
5. Resolution messages
6. Handoff trigger and message
7. Error states and recovery
8. Closing messages

Rules:
- No more than 3 options per menu
- Every path must have an exit to human support
- No dead ends
- Test every path mentally

Draft a refund and return policy for [PRODUCT / SERVICE].

Inputs:
- Jurisdiction: [COUNTRY / STATE]
- Product type: [PHYSICAL / DIGITAL / SUBSCRIPTION]
- Return window: [DAYS]
- Condition requirements: [CONDITIONS]
- Exceptions: [WHAT'S NOT RETURNABLE]
- Process: [HOW CUSTOMERS RETURN]

Structure:
1. Overview
2. Eligibility window
3. Condition requirements
4. What cannot be returned
5. How to initiate
6. Refund method and timing
7. Shipping (who pays)
8. Exchanges
9. Damaged or defective items
10. Contact

Rules: plain language. No hidden conditions. Flag any consumer-protection law requirements for our jurisdiction.

Draft proactive outreach to customers affected by [ISSUE / CHANGE].

Inputs:
- What happened: [ISSUE]
- Who is affected: [SEGMENT]
- What we're doing: [REMEDIATION]
- What customer needs to do: [ACTION OR NOTHING]
- Timeline: [WHEN]
- Compensation: [IF ANY]

Produce:
1. Subject line (honest, not alarming)
2. Greeting
3. What happened (factual, brief)
4. Who is affected and how
5. What we're doing
6. What they need to do
7. When it will be resolved
8. How to reach us
9. Apology that lands

Rules: reach out before they discover the issue. No marketing language in an incident message.

Create a training scenario for new support agents.

Inputs:
- Skill being trained: [SKILL]
- Scenario type: [TICKET / CALL / CHAT]
- Difficulty: [BEGINNER / INTERMEDIATE / ADVANCED]
- Product: [WHAT]

Produce:
1. Scenario setup (customer situation)
2. Customer persona and tone
3. Opening message
4. Hidden information the agent must uncover
5. Common wrong responses
6. Best-practice response path
7. Debrief questions
8. Evaluation rubric

Rules: realistic, not cartoonish. Include ambiguity like real customers. Test judgment, not just knowledge.

Build a CSAT improvement plan for [SUPPORT TEAM].

Inputs:
- Current CSAT: [SCORE]
- Target: [SCORE]
- Volume: [TICKETS PER PERIOD]
- Top complaints: [THEMES]
- Team size: [N]
- Timeframe: [WEEKS / MONTHS]

Produce:
1. Root cause analysis of current CSAT drivers
2. Quick wins (under 30 days)
3. Structural improvements (30-90 days)
4. Metrics to track weekly
5. Training and coaching plan
6. Process changes required
7. Expected impact per action
8. Owner per action

Rules: address the top 2 root causes first. Don't spread effort thin across everything.

Write the narrative for a board meeting on [DATE].

Inputs:
- Period covered: [Q / MONTH]
- Key wins: [LIST WITH NUMBERS]
- Key misses: [LIST WITH HONEST EXPLANATION]
- Strategic topics to discuss: [LIST]
- Decisions needed from board: [LIST]
- Forward guidance: [NUMBERS OR RANGE]

Produce:
1. Opening framing — one paragraph a board member will remember
2. Performance summary (facts, no spin)
3. Wins and what they signal
4. Misses and what we're doing
5. Strategic topics with background and options
6. Board asks — decisions we need
7. Closing summary

Rules: no hedging. Board members want clarity, not comfort. Lead with the hardest news.

Build a strategic plan for [COMPANY] for [PERIOD].

Inputs:
- Current state: [REVENUE, TEAM, PRODUCTS, MARKETS]
- Vision: [WHERE WE WANT TO BE]
- Time horizon: [12 / 24 / 36 months]
- Constraints: [CAPITAL, TALENT, MARKET]
- Competitive landscape: [WHO + HOW]

Produce:
1. Situation assessment (internal + external)
2. Strategic priorities (3-5, ranked)
3. Where we will play / where we will NOT play
4. Key bets with reasoning
5. Resource requirements
6. Milestones and metrics
7. Leading indicators to track
8. Risks and mitigation

Format as a one-page strategy on a page + supporting detail. Force choices — no "and" strategies.

Draft investor communication for [EVENT: earnings / raise / milestone].

Inputs:
- Event: [WHAT]
- Audience: [EXISTING / PROSPECTIVE / BOTH]
- Key messages: [3 POINTS]
- Numbers: [REVENUE, BURN, RUNWAY, GROWTH]
- Asks: [INTROS / CAPITAL / ADVICE]
- Tone: [CONFIDENT / CAUTIOUS / CELEBRATORY]

Produce:
1. Subject line
2. Opening (numbers first)
3. What happened and why it matters
4. What's next
5. Specific asks
6. Closing

Rules: no spin. Investors can smell it. Lead with what matters most even if it's not the best news. Numbers in the first two sentences.

Build a decision matrix for [DECISION].

Inputs:
- Decision: [WHAT WE'RE DECIDING]
- Options: [OPTION A, B, C, D]
- Criteria: [WHAT MATTERS — list with weights]
- Constraints: [TIME, BUDGET, PEOPLE]
- Stakeholders: [WHO DECIDES, WHO INFLUENCES]

Produce:
1. Criteria with weights (must total 100)
2. Scoring rubric per criterion
3. Matrix: options × criteria
4. Weighted score per option
5. Sensitivity analysis — what if weights shift
6. Recommendation with rationale
7. What would change the recommendation
8. Reversibility — can we undo this

Format as a table. Be explicit about assumptions.

Analyze [TARGET COMPANY] as a potential acquisition.

Inputs:
- Target: [NAME, INDUSTRY, SIZE]
- Strategic rationale: [WHY WE'RE INTERESTED]
- Public info: [WHAT WE KNOW]
- Deal size range: [AMOUNT]
- Integration model: [ABSORB / STANDALONE / HYBRID]

Produce:
1. Strategic fit assessment
2. Financial profile (based on available info)
3. Synergy hypotheses with estimates
4. Integration risks
5. Cultural fit considerations
6. Alternative structures (asset, stock, merger, minority stake)
7. Go / no-go recommendation with reasoning
8. Diligence priorities if we proceed

Flag all assumptions. This is analysis, not advice.

Write an executive briefing on [TOPIC].

Inputs:
- Audience: [CEO / BOARD / ELT]
- Decision needed: [YES / NO / WHAT]
- Time available: [2 / 5 / 15 minutes]
- Background: [CONTEXT]
- Key facts: [DATA POINTS]
- Options: [LIST]
- Recommendation: [WHAT YOU THINK]

Structure:
1. Bottom line up front (3 sentences)
2. Why now
3. Key facts (bulleted, quantified)
4. Options with tradeoffs
5. Recommendation with rationale
6. Decision needed by [DATE]
7. Next steps if approved

Rules: one page max. BLUF format. No setup — lead with the conclusion.

Assess the market opportunity for [PRODUCT / MARKET].

Inputs:
- Market: [WHAT]
- Geography: [WHERE]
- Buyer: [WHO]
- Current alternatives: [WHAT THEY USE TODAY]
- Our angle: [DIFFERENTIATOR]

Produce:
1. Market size (TAM, SAM, SOM) with reasoning
2. Growth rate and drivers
3. Competitive landscape (direct, indirect, substitutes)
4. Buyer behavior and buying criteria
5. Barriers to entry
6. Our right to win (or not)
7. Go-to-market implications
8. Go / no-go recommendation

State every assumption. Flag where estimates are soft.

Build an enterprise risk register for [COMPANY].

Inputs:
- Industry: [INDUSTRY]
- Company size: [N]
- Key dependencies: [CUSTOMERS / SUPPLIERS / SYSTEMS]
- Recent incidents: [LIST]
- Regulatory environment: [RULES]

Produce a table with:
- Risk ID
- Category (strategic, operational, financial, compliance, cyber, reputational)
- Description
- Likelihood (1-5)
- Impact (1-5)
- Risk score
- Current controls
- Residual risk
- Owner
- Mitigation action
- Review date

Top 10 risks only. Rank by residual risk score. Include at least one risk most companies in our position underestimate.

Prepare crisis communication for [CRISIS SCENARIO].

Inputs:
- Scenario: [WHAT HAPPENED / COULD HAPPEN]
- Stakeholders affected: [CUSTOMERS / EMPLOYEES / INVESTORS / REGULATORS / PUBLIC]
- Severity: [LOW / MEDIUM / HIGH / SEVERE]
- Facts confirmed: [WHAT WE KNOW]
- Facts unknown: [WHAT WE DON'T KNOW]
- Legal constraints: [IF ANY]

Produce:
1. First statement (within 2 hours — acknowledge, don't explain)
2. Holding statement for ongoing situation
3. Full statement when facts are clear
4. Internal communication (employees first)
5. Customer communication
6. Investor / regulator communication
7. Media Q&A with prepared responses
8. What NOT to say
9. Update cadence

Rules: no speculation. No blaming. No "thoughts and prayers." Lead with actions, not feelings.

Structure an annual strategy review for [COMPANY].

Inputs:
- Year covered: [YEAR]
- Goals set vs. achieved: [LIST]
- Key learnings: [WHAT WE LEARNED]
- Market changes: [WHAT SHIFTED]
- Team changes: [HIRES / DEPARTURES]
- Financials: [REVENUE, MARGIN, CASH]

Produce:
1. Year in review — what happened, honestly
2. What worked and why
3. What didn't work and why
4. Market changes we missed
5. Strategic implications
6. Priorities for next year (3-5)
7. What we will stop doing
8. Resource reallocation
9. 90-day action plan

Rules: candor over comfort. The "what we will stop" section is the hardest and most valuable.

The Real Competitive Advantage Isn't Access to AI. It's How You Execute With It.

This is the section most "AI business prompts" pages skip entirely. Because the real insight — the one that separates organizations experimenting with AI from organizations executing with it — isn't about finding better prompts.

It's about building systems.

Consider the data:

• 88% of organizations now use AI in at least one business function (McKinsey, 2025).
• 92% plan to increase AI investment over the next three years.
• But only 1% of leaders describe their organizations as "mature" in AI deployment.

That 91-point gap between investment and maturity is a systems problem. It's not a model problem, a budget problem, or a talent problem. It's the absence of infrastructure that turns individual AI usage into organizational capability.

That infrastructure includes:

Standardization

When prompts are engineered once, governed centrally, and deployed consistently, output quality becomes predictable across the organization — not dependent on which individual happens to be good at ChatGPT.

Governance

Version control, approval workflows, audit trails, and usage tracking transform AI from an unmanaged experiment into a governed operation. This is especially critical in regulated industries, but valuable everywhere.

Workflow integration

Prompts that exist in isolation — in bookmark folders, Google Docs, or someone's head — don't compound. Prompts that are embedded in team workflows, connected to business functions, and measured for effectiveness become organizational knowledge.

Execution intelligence

Understanding what prompts are used, by whom, how often, and with what outcomes turns AI adoption from a faith-based initiative into a data-informed strategy.

This is the shift that matters. Not from "no AI" to "using AI" — most organizations have already crossed that line. But from improvised AI to governed AI. From scattered prompts to prompt systems. From individual experiments to execution infrastructure.

What Global Communities Are Actually Asking For

The most reliable signal of what business users need from AI prompts doesn't come from vendor marketing. It comes from the questions people ask when no one is selling them anything.

Reddit: Frameworks, not tricks

The highest-engagement prompts on business subreddits share one trait: they force structured decision-making. A prompt that asks the AI to "challenge my assumptions and suggest alternatives" consistently outperforms prompts that ask for ideas. A prompt that assigns the AI a ruthless mentor persona — "be brutally honest, no motivational fluff" — generates more actionable output than one that asks for a "helpful analysis".

Quora: Context is non-negotiable

Quora's business prompt discussions reveal a consistent frustration: generic prompts produce generic results. The solution users converge on is the same framework: define who you are, who your audience is, and what you're trying to accomplish. A Quora answer about prompt structure puts it bluntly: "When most people type a prompt like 'marketing plan for a bakery,' it defaults to the most generic, average response possible. But if you give it the Context-Task-Constraint framework, it will deliver exactly what you need".

Pinterest: Visual + strategic

Pinterest's business prompt demand centers on market opportunity scanning and content strategy. The most-shared prompt frameworks focus on identifying profitable micro-niches with high search demand but low creator saturation, generating board ideas that attract target audiences, and creating captions that stop the scroll without sacrificing brand voice. This is prompt engineering applied to visual discovery platforms — a growing category as businesses realize that Pinterest is a search engine, not just a mood board.

India: Speed + localization

India's prompt demand has a distinct character: urgency and scale. With 93% of enterprises increasing AI budgets and 40% of enterprise apps expected to feature task-specific AI agents by 2026, the need isn't for sophisticated prompt architecture theory — it's for prompts that work immediately, in Indian business contexts, across English and regional language workflows. The Asia Pacific prompt marketplace's defining features — localization, multilingual adaptation, and workflow-ready assets — are most pronounced in India, where businesses are skipping the "experimentation" phase and moving directly to deployment.

Start Free. Scale When You're Ready.

You can begin immediately with the prompt libraries above — organized by business function. No credit card. No demo request. Copy any prompt, replace the bracketed fields, and paste into your AI tool. Compare the output to what you've been getting with ad hoc prompts.

Most teams follow a natural path:

1. Browse the prompt library for your specific function.
2. Test a few structured prompts with your existing AI tools.
3. Compare the output to what you've been getting with ad hoc prompts.
4. Adopt team-wide with governance features when you're ready to scale.

Frequently Asked Questions

How to Know If Your Organization Is Ready to Scale AI

Don't summarize. Give a judgment standard.

One-sentence positioning summary: Competitors sell prompts. The real solution is ending Prompt Debt. Own the concept of "debt" and you own the category.