75+ ChatGPT Prompts for Small Business Owners (2026 Edition)
The most practical ChatGPT prompt library for small business owners in 2026. 75+ tested prompts across marketing, sales, support, hiring, finance, and operations — with a framework for writing your own.
Most "ChatGPT prompts for business" lists are useless. They give you 30 variations of "write a marketing plan for my business" and call it a prompt library. The prompts in this guide are different — every one has been tested in actual small business workflows, each comes with a note on why it works, and they're organized so you can grab the right prompt for the right job without scrolling for 10 minutes.
We'll cover 75+ prompts across nine categories: marketing, sales, customer service, social media, HR, finance, operations, strategy, and SEO. Before the prompts, a short section on the formula that makes any prompt better.
If you want to go beyond prompts and have AI actually do the work for you, see our AI Agents for Small Business buyer's guide.
The prompt formula that makes everything else easier
Before you copy a single prompt below, internalize this. Every good business prompt has four parts:
1. Role — Who should the AI act as? ("Act as a senior copywriter specializing in DTC e-commerce.")
2. Context — What does it need to know about your business? (Industry, audience, current situation, brand voice.)
3. Task with constraints — What exactly do you want, with limits? (Word count, format, tone, what to include or avoid.)
4. Output format — How should it deliver? (Table, bullet list, paragraph, JSON, markdown.)
This is sometimes called the R-C-T-O framework. Once you've used it a few times, you'll find yourself naturally writing better prompts without thinking about the formula.
The single biggest upgrade you can make to any prompt: Add the phrase "Before you start, ask me any clarifying questions you need so you can give me a better response." This turns ChatGPT from a one-shot answer machine into a conversation partner that actually gathers what it needs.
Marketing & Content Creation (12 prompts)
1. Customer avatar deep-dive
Act as a senior marketing strategist. I run [BUSINESS TYPE] serving [GENERAL AUDIENCE].
Build a detailed customer avatar including:
- Demographics (age, income, location, role)
- Top 3 pain points related to my offer
- Top 3 desires or outcomes they want
- Where they spend time online
- What objections they typically have before buying
- The exact words they use to describe their problem (not marketing language)
Format as a one-page brief I can hand to a copywriter.
Why it works: Asks for the language customers use, not just demographics. That's the difference between marketing that converts and marketing that sounds generic.
2. Value proposition stress test
Here's my current value proposition: "[PASTE VALUE PROP]"
Critique it as if you were a skeptical first-time visitor to my website. Tell me:
1. What you understand it to mean
2. What's unclear or vague
3. What makes me different from competitors (or if you can't tell)
4. What you'd want to know before you'd consider buying
Then suggest 3 sharper alternatives.
Why it works: Forces ChatGPT to roleplay the skeptic rather than the cheerleader. The critique is more valuable than the rewrite.
3. Email subject line generator
I'm sending a [TYPE OF EMAIL: promo / nurture / re-engagement] to [SEGMENT].
The email's core message is: [ONE SENTENCE]
Generate 15 subject lines varying across:
- Curiosity-driven (3)
- Direct/benefit-led (3)
- Question-based (3)
- Urgency or scarcity (3)
- Numbers or specifics (3)
Each under 50 characters. No clickbait or false promises.
Why it works: Forces variety across proven subject line patterns instead of getting 15 versions of the same approach.
4. Landing page outline
Act as a conversion copywriter. I'm launching a landing page for [PRODUCT/SERVICE] targeting [AUDIENCE]. The single goal is [DESIRED ACTION].
Outline the page section by section, with:
- A headline that names the outcome, not the product
- A subhead that names the pain it solves
- 3-5 benefit bullets (not features)
- Social proof placement
- An FAQ section anticipating top 5 objections
- A specific CTA
For each section, include a one-line note on why it's there.
Why it works: Anchors on outcome and pain rather than features. The "why it's there" forces strategic thinking, not just structure.
5. Brand voice extractor
I'm pasting 3 examples of writing that sound like me/my brand: [PASTE 3 SAMPLES]
Analyze and document my brand voice as a reusable style guide, including:
- 3-5 voice attributes (with adjectives)
- Sentence structure patterns
- Words and phrases I tend to use
- Words and phrases I avoid
- Punctuation preferences
- A do/don't list for new writers
Then write one sample paragraph in my voice on the topic of [TOPIC] to verify.
Why it works: Captures your actual voice from samples instead of guessing from a description. Save the output and paste it into Custom Instructions.
6. Competitor positioning map
Here are 5 competitors in my space: [LIST WITH WEBSITES OR DESCRIPTIONS]
Map their positioning across two axes I should choose. Then:
1. Show where each one sits
2. Identify the unoccupied position with strongest opportunity
3. Suggest 3 messaging angles I could own based on that gap
If you need more info on any competitor, ask before guessing.
Why it works: Most positioning exercises are vague. The two-axis map forces specificity, and the unoccupied-quadrant frame is genuinely useful.
7. Press release draft
Act as a PR writer. Draft a press release announcing [EVENT/MILESTONE/LAUNCH] for my business [BUSINESS NAME], a [BRIEF DESCRIPTION].
Key facts: [BULLET POINTS]
Structure:
- Headline (under 90 characters)
- Subhead
- Dateline + lead paragraph (the 5 W's)
- 2 supporting paragraphs with quotes from [PERSON, TITLE]
- Boilerplate "About" paragraph
- Contact info placeholder
Tone: professional, factual, no hype.
Why it works: Follows AP press release structure that journalists actually expect, not the generic "draft a press release" output.
8. Tagline generator
My business [DESCRIBE WHAT YOU DO] helps [AUDIENCE] achieve [OUTCOME] by [DIFFERENTIATOR].
Generate 20 taglines across these styles:
- 5 benefit-led (state the outcome)
- 5 metaphor-based (use comparison)
- 5 contrarian (challenge a category assumption)
- 5 short and punchy (under 4 words)
For each one, include a one-line explanation of the angle.
Why it works: Variety across angle types prevents the "all 20 taglines sound the same" problem.
9. Case study draft from raw notes
Here are my notes from a customer success conversation: [PASTE NOTES]
Turn this into a 600-word case study with:
- A specific, outcome-led headline
- The customer's situation before (with specifics: numbers, time, pain)
- What they tried that didn't work
- How they implemented our solution
- The measurable results (with numbers)
- A pull-quote
- A short CTA
Keep the customer's voice — don't over-polish their words.
Why it works: Most case studies sound corporate because they over-edit the customer. The "don't over-polish" line preserves authenticity.
10. Re-engagement email for cold list
Write a 150-word re-engagement email to subscribers who haven't opened in 90+ days.
Context: [BUSINESS/NICHE]
Last great thing we offered them: [DESCRIBE]
What's new and worth coming back for: [DESCRIBE]
Tone: warm, honest, slightly self-deprecating. Acknowledge the gap. Make it easy to either re-engage or unsubscribe. Don't beg.
End with a single, specific CTA.
Why it works: The "honest, slightly self-deprecating" framing produces emails that don't sound like every other "We miss you!" template.
11. Blog post outline (SEO + AI search friendly)
I want to rank a blog post for the keyword "[KEYWORD]" — my audience is [DESCRIBE].
Build an outline that:
- Has an H1 with the keyword naturally placed
- Opens with a 2-3 sentence direct answer (for AI Overviews / featured snippets)
- Has 6-8 H2 sections covering the topic comprehensively
- Includes 2-3 H3s under most H2s
- Has an FAQ section at the end with 5 likely questions
- Suggests 3 internal links to other content I might have
For each H2, include a one-sentence note on what the section should cover.
Why it works: Optimizes for both traditional SEO and how AI engines extract answers. The direct-answer opener is specifically structured for AI Overviews.
12. Pricing page copy
Help me write copy for a 3-tier pricing page.
Product: [DESCRIBE]
Audience: [DESCRIBE]
Tier names should reflect customer aspirations, not size.
For each tier, give me:
- Tier name
- One-line tier description (who it's for)
- Headline benefit (not feature)
- 5-7 feature bullets
- A "best for" line
Add a one-paragraph FAQ section addressing the top 3 pricing objections.
Why it works: "Tier names should reflect aspirations, not size" prevents the generic "Basic / Pro / Enterprise" trap that makes pricing pages forgettable.
For more on the marketing side, see How to write marketing content with AI (without sounding like a robot).
Sales & Lead Generation (10 prompts)
13. Cold email writer
Write a cold email from me to [PROSPECT TYPE] at [COMPANY TYPE].
My offer: [DESCRIBE]
Specific reason to reach out NOW: [TRIGGER EVENT OR INSIGHT]
My ideal CTA: a 15-minute call
Constraints:
- Under 100 words
- No "I hope this email finds you well" or "quick question"
- Lead with something specific to them, not me
- One clear CTA at the end
Generate 3 variations with different angles.
Why it works: Bans the most overused cold email phrases and forces specificity in the opener — the part 90% of cold emails fail on.
14. LinkedIn connection request
Write a LinkedIn connection request to [PERSON, TITLE, COMPANY].
Context I know about them: [BULLET POINTS]
Why I want to connect: [REASON]
Constraints:
- Under 300 characters
- Reference something specific from their profile or recent activity
- Don't pitch anything in the message
- Sound like a human, not a recruiter
Why it works: "Don't pitch" is the key constraint. The connection request's job is to start a relationship, not close a deal.
15. Discovery call question bank
I'm doing a 30-minute discovery call with a prospect.
My offer: [DESCRIBE]
Their context: [WHAT I KNOW]
Generate 15 questions across:
- Situation (where they are now)
- Pain/cost (what's not working and what it costs them)
- Past attempts (what they've tried)
- Decision process (how they buy)
- Outcomes (what success looks like)
Order them in a natural conversation flow, not by category.
Why it works: The "natural conversation flow" instruction prevents the rigid checklist feel that kills discovery calls.
16. Objection handling script
A prospect just said: "[PASTE OBJECTION]"
Generate 3 responses across these angles:
1. Acknowledge and reframe (validate the concern, offer a different lens)
2. Question and clarify (find what's really behind it)
3. Evidence and proof (counter with a specific case or data point)
Each under 100 words. Tone: confident but not pushy.
Why it works: Three different angles let you pick what fits the moment rather than memorizing one rebuttal.
17. Proposal customization template
Take this base proposal: [PASTE PROPOSAL]
Customize it for [PROSPECT NAME] at [COMPANY], who has these specific needs and concerns: [LIST FROM DISCOVERY CALL]
Changes:
- Replace generic outcomes with their specific desired outcomes
- Adjust the timeline and scope to fit their situation
- Address their stated objections preemptively
- Update the case studies to ones most relevant to their industry
- Keep the same pricing structure
Why it works: Customization is what differentiates winning proposals. This forces you to use discovery info instead of sending a templated PDF.
18. Follow-up email after no response
I sent this email to [PROSPECT] 5 business days ago and got no response: [PASTE ORIGINAL]
Write a follow-up that:
- Doesn't apologize for following up
- Doesn't say "just bumping this to the top of your inbox"
- Adds new value (a relevant resource, insight, or angle)
- Has a softer CTA that's easier to say yes to
Under 75 words.
Why it works: Bans the two most common follow-up clichés. The "softer CTA" forces a real strategic choice about the ask.
19. Win-back email for churned customer
Write an email to a customer who churned [TIME AGO] for [REASON].
What's different now: [WHAT'S CHANGED]
Specific incentive (if any): [DESCRIBE]
Tone: humble, no defensiveness. Acknowledge what they experienced. Show what's genuinely different. Make it easy to come back without making it pushy.
Under 150 words.
Why it works: "No defensiveness" is the key. Most win-back emails subtly argue with the customer's original reason for leaving.
20. Lead qualification checklist
Act as a sales operations expert. Build me a 7-point lead qualification checklist for [MY OFFER].
For each point:
- The exact question to ask (in the lead's words, not jargon)
- What a "qualified" answer looks like
- What a "disqualifying" answer looks like
- Where in the funnel to ask it
Format as a table I can use in a CRM.
Why it works: Forces specificity on what disqualifies a lead, which is the part most teams skip — and exactly why pipelines clog with dead leads.
21. Referral request template
Write a short message asking a happy customer for a referral.
Customer: [NAME], who got [SPECIFIC OUTCOME] from working with us
My ideal next customer: [DESCRIBE]
Constraints:
- Under 100 words
- Make it easy for them to refer (or to decline without feeling bad)
- Give them a specific "ask" — not "do you know anyone?"
- Don't offer a reward in the first ask
Why it works: "Specific ask" beats "do you know anyone" by a wide margin. Most referrals fail because the request is too vague.
22. Sales call recap email
Here are my notes from a sales call with [PROSPECT]: [PASTE NOTES]
Write a recap email to send within 2 hours that:
- Summarizes their key challenges in their words
- Confirms the agreed next steps with dates
- Restates the outcomes we discussed they want
- Asks for any documents or intros needed before next meeting
- Ends with a single calendar link
Tone: warm and decisive. Not corporate.
Why it works: "In their words" is the magic phrase. Recaps that mirror the prospect's language build trust; recaps that translate it into your jargon don't.
Customer Service & Support (8 prompts)
23. FAQ generator from real customer messages
Here are 15 recent customer messages: [PASTE MESSAGES]
Identify:
1. The top 7 questions or themes (deduplicated)
2. The best plain-English answer for each (3-4 sentences max)
3. Which 2-3 should be turned into proactive onboarding content
4. Which 1-2 might indicate a product issue, not a support issue
Format as a table.
Why it works: Mines patterns from real data instead of guessing. The "product issue vs. support issue" line catches problems that better support won't fix.
24. Refund response template
Write a refund response email for this situation: [DESCRIBE]
Approve / partial / decline: [SPECIFY]
Constraints:
- Acknowledge their experience in the first sentence
- State the decision clearly without burying it
- Explain the reasoning in plain English (not policy jargon)
- Offer a next step if relevant
- Under 120 words
Why it works: "Don't bury the decision" prevents the most common refund email mistake — making the customer hunt for the answer.
25. Difficult customer de-escalation
A customer is upset because [DESCRIBE SITUATION]. Their last message was: "[PASTE]"
Write a response that:
- Acknowledges their frustration without admitting fault we don't have
- Asks 1-2 specific questions to understand the issue
- Sets one clear next step
- Doesn't use "I understand" or "I apologize for any inconvenience"
Tone: calm, professional, human.
Why it works: Banning "I understand" and "I apologize for any inconvenience" forces actually meaningful empathy instead of customer service auto-pilot.
26. Help doc article
Write a help doc article on: [TOPIC]
Audience: [USER TYPE]
Their goal: [WHAT THEY'RE TRYING TO DO]
Structure:
- Title that names the task, not the feature
- Direct answer in 2 sentences
- Step-by-step instructions with numbered steps
- A "Common issues" section with 3 troubleshooting items
- A "Related articles" placeholder section
Plain language. No jargon. Active voice.
Why it works: "Names the task, not the feature" is what separates docs people actually find via search from docs that sit unread.
27. Customer satisfaction survey
Build a 5-question post-purchase survey for [BUSINESS TYPE].
Constraints:
- Mix of multiple choice and one short open-text
- One question must be an NPS or similar number score
- Avoid leading or double-barreled questions
- Should be completable in under 90 seconds
- One question should surface improvement ideas
After each question, note what insight it's designed to surface.
Why it works: "Note what insight it's designed to surface" forces every question to earn its spot. Most surveys are 15 questions long because no one questioned each one.
28. Auto-reply for off-hours
Write an auto-reply for customer messages received outside business hours.
Business: [TYPE]
Hours: [HOURS]
Typical response time: [TIME]
Emergency path: [DESCRIBE OR "N/A"]
Tone: helpful, not robotic. Set clear expectations. Offer self-service resources where possible.
Under 100 words.
Why it works: "Not robotic" plus "clear expectations" is the right balance. Auto-replies fail when they're either too generic or too restrictive.
29. Negative review response
A customer left this negative review: "[PASTE]"
Write a public response that:
- Acknowledges the specific issue they raised
- Doesn't get defensive or argue the facts
- Offers a private channel for resolution
- Demonstrates to other readers that we handle issues well
Under 75 words. Tone: professional, accountable, human.
Why it works: Public review responses have two audiences — the reviewer and every future reader. The "demonstrates to other readers" line keeps that in mind.
30. Internal escalation summary
Here's a support conversation that needs escalation: [PASTE THREAD]
Summarize for the engineer/manager taking it over:
- Customer name and account context
- The issue in one sentence
- What's already been tried
- Current status and emotional temperature
- What the customer is asking for
- My recommended next step
Format as a clean handoff doc, under 200 words.
Why it works: Forces a structured handoff that saves the next person from re-reading the entire thread.
For the broader support automation play, read How to automate customer support with AI.
Social Media (10 prompts)
31. 30-day content calendar
Build a 30-day social media content calendar for my [BUSINESS TYPE] targeting [AUDIENCE].
Constraints:
- 4 posts per week (Mon, Tue, Thu, Fri)
- Mix: 40% educational, 30% engagement/community, 20% promotional, 10% behind-the-scenes
- Each post: theme, hook, 1-line description, suggested format (carousel, single image, video, text)
- Tied to monthly theme: [THEME]
Format as a table.
Why it works: The 40/30/20/10 mix prevents the all-promo or all-educational extremes that kill engagement.
32. Repurpose long-form into 10 posts
Here's a long-form article/podcast/video transcript: [PASTE]
Pull 10 social media posts from it across:
- 3 LinkedIn posts (200-300 words each, with a strong hook)
- 3 Twitter/X posts or threads (each under 280 chars per tweet)
- 2 Instagram carousel concepts (slide-by-slide)
- 2 short video script ideas (under 60 seconds each)
Each one should stand alone — not require the original.
Why it works: "Stand alone" is the differentiator. Most repurposing produces fragments that only make sense if you've seen the source.
33. LinkedIn post (educational)
Write a LinkedIn post on: [TOPIC]
My audience: [DESCRIBE]
The insight I want to share: [INSIGHT]
Structure:
- Hook line that creates curiosity (no clickbait)
- 2-3 sentences of context
- 3-5 short paragraphs of the insight, with line breaks
- A reflection or question at the end
- No hashtags
Under 250 words. Plain language. No "thrilled to share."
Why it works: Banning "thrilled to share" and hashtags cuts the LinkedIn-cringe by half. The structure mirrors what actually performs in 2026.
34. Instagram carousel script
Build a 10-slide Instagram carousel on: [TOPIC]
Audience: [DESCRIBE]
Goal: [educate / build authority / convert]
For each slide:
- Slide number and visual concept
- Headline (under 6 words)
- Body copy (under 30 words)
- Why this slide is here
Slide 1 = hook. Slide 10 = CTA.
Tone: [DESCRIBE]
Why it works: "Why this slide is here" forces every slide to earn its place — the same principle as good slide deck design.
35. Twitter/X thread
Write a 7-tweet thread on: [TOPIC]
Hook tweet must:
- Open with a specific claim or number
- Create a curiosity gap
- Not use "thread" or "🧵" emoji
Each tweet stands alone but builds. Final tweet: a soft CTA to [DESTINATION].
Under 280 chars per tweet. No threadboi clichés ("Most people get this wrong...", "Buckle up...").
Why it works: Bans the specific clichés that signal "AI-generated thread" to anyone who's been on X for more than a year.
36. Short-form video script (TikTok/Reels/Shorts)
Write a 45-second short-form video script on: [TOPIC]
Format:
- Hook (first 3 seconds): a contradiction, claim, or question
- Setup (5 sec): what we're going to show
- Core content (30 sec): 3 quick points with on-screen text suggestions
- Payoff (5 sec): the "aha" moment
- CTA (2 sec): one specific action
Include shot suggestions and on-screen text.
Why it works: The 3-second hook is everything. The structure mirrors how high-retention short-form actually works in 2026.
37. Caption A/B test set
Generate 5 caption variations for the same Instagram post.
Topic: [TOPIC]
Image/video shows: [DESCRIBE]
Audience: [DESCRIBE]
Each variation tests a different opener:
1. Question
2. Bold statement
3. Story (3-sentence mini-narrative)
4. Listicle teaser
5. Direct address ("To everyone who...")
Each under 150 words.
Why it works: Each variation tests one variable, which makes the A/B test actually meaningful.
38. Comment response templates
Build 10 reusable response templates for common Instagram/LinkedIn comments:
1. "This is so helpful, thank you!"
2. A question I get often: [QUESTION]
3. Mild disagreement or critique
4. A request for more info
5. A spam/sales pitch
6. A compliment with no question
7. A long, thoughtful comment
8. A tag of someone else
9. A DM-this-to-me request
10. A "where can I learn more?" question
Each under 30 words. Tone: warm, conversational, distinctly human.
Why it works: Having pre-built responses for each scenario removes the friction that causes comment sections to go dark.
39. Social bio rewrite
Here's my current social bio: "[PASTE]"
Rewrite it for [PLATFORM] across 3 angles:
1. Outcome-led (focuses on what I help people achieve)
2. Authority-led (focuses on credentials and proof)
3. Curiosity-led (focuses on a contrarian or specific identity)
Each within the platform character limit. Each should make someone want to follow within 3 seconds.
Why it works: Three distinct angles for the same person let you pick what matches the audience you're trying to grow.
40. Engagement-bait detector
Here's a social post draft I'm considering: "[PASTE]"
Audit it for engagement-bait patterns that hurt long-term reach:
- Fake authority ("As someone who has done X 1,000 times...")
- Manufactured controversy
- Vague universal claims ("Everyone is doing X wrong")
- Hookable line that doesn't deliver
- Generic AI-generated phrasing
Suggest a rewrite that keeps the insight but cuts the bait.
Why it works: Self-audit catches what writers don't see in their own drafts.
HR & Hiring (8 prompts)
41. Job description writer
Write a job description for [ROLE] at my [BUSINESS TYPE].
Constraints:
- Lead with the outcomes the person will own, not the daily tasks
- 5-7 specific responsibilities (active verbs)
- 5-7 must-have qualifications
- 3-5 nice-to-haves
- A "what success looks like in 90 days" section
- Salary range: [RANGE]
- No "rockstar," "ninja," "wear many hats"
Tone: clear, honest, slightly warm.
Why it works: Outcome-led JDs attract better candidates than task-led ones. The banned-phrases list eliminates the worst clichés.
42. Resume screening rubric
Build a 10-point screening rubric for [ROLE].
For each criterion:
- What to look for in the resume
- What's a green flag
- What's a yellow flag worth probing
- What's a red flag
Format as a table. Add a section at the bottom on bias-prone fields to ignore (school name, gaps, etc.).
Why it works: The bias-awareness section turns this from "screen faster" into "screen better."
43. Interview question bank
Build a 15-question interview bank for [ROLE] at my [BUSINESS TYPE].
Mix:
- 5 behavioral (past situations)
- 4 situational (hypothetical scenarios)
- 3 technical/domain-specific (skills validation)
- 2 cultural/values fit
- 1 "what questions do you have for us" prompt
For each: the question, what a great answer looks like, and what a concerning answer looks like.
Why it works: "What concerning looks like" matters as much as "great" — and most interview prep skips it.
44. Onboarding checklist (30/60/90)
Build a 30/60/90 day onboarding plan for a new [ROLE] at my [BUSINESS TYPE].
For each phase:
- 3-5 specific outcomes they should hit
- 5-7 concrete activities
- Who they should be meeting/learning from
- Resources they need
- A check-in moment with their manager
Format as a structured doc, not a single list.
Why it works: Outcomes-first onboarding beats task-list onboarding by a wide margin.
For the deeper hiring playbook, read AI for hiring: how to screen applicants, write job posts, and onboard faster.
45. Performance review draft
Write a performance review for [EMPLOYEE NAME], who has been in [ROLE] for [TIME].
Inputs:
- 3 strengths I've observed: [LIST]
- 2 growth areas: [LIST]
- Key projects/outcomes this period: [LIST]
- Their stated goals: [LIST]
Structure:
- Opening summary
- Specific strengths with examples
- Specific growth areas with paths forward
- Goals for next period
- Closing
Tone: honest, supportive, no fluff. Treat them as an adult.
Why it works: "Treat them as an adult" cuts the corporate softening that makes performance reviews useless.
46. Difficult feedback conversation script
I need to give [EMPLOYEE] feedback on [ISSUE].
Background: [CONTEXT]
Outcome I want: [BEHAVIOR CHANGE / CLARITY / DECISION]
Build me a conversation outline:
- How to open without ambushing them
- Specific behavior + impact framing
- Where to pause and listen
- Possible responses they might give and how to handle each
- How to close with clear next steps
Tone: direct but not harsh. Respect their autonomy.
Why it works: "Where to pause and listen" is the line most managers skip — and exactly why these conversations go sideways.
47. Offer letter framework
Draft an offer letter for [CANDIDATE NAME] for [ROLE].
Details: [SALARY, START DATE, BENEFITS, ANY EQUITY OR BONUSES]
Reporting to: [MANAGER]
Tone: enthusiastic but professional. Make them feel chosen, not transactional.
Include:
- Personal opening (1-2 sentences referencing the interview process)
- Role, comp, and key terms
- What they'll be working on first
- Next steps and timeline
- A note from leadership
Under 400 words.
Why it works: "Make them feel chosen" prevents the boilerplate-PDF feel that kills offer acceptance rates.
48. Exit interview question bank
Build a 10-question exit interview for [ROLE LEVEL].
Goal: surface honest insight on culture, management, growth, comp, and process — without putting them on the defensive.
Mix open and rated questions. Order them from low-stakes to high-stakes.
For each: the question, plus one follow-up probe if they give a short answer.
Why it works: The "low-stakes to high-stakes" order is how journalists structure interviews to get real answers.
Finance & Bookkeeping (8 prompts)
49. Cash flow analysis
Here are my last 6 months of revenue and expense data: [PASTE OR DESCRIBE]
Analyze:
1. Top 3 trends I should be aware of
2. Any concerning patterns
3. Months that look like outliers and why
4. 2-3 specific actions I should consider based on this data
Tone: analytical, not alarmist. Show your reasoning.
Why it works: "Show your reasoning" lets you sanity-check the conclusions. AI can be confidently wrong about math.
50. Pricing change analysis
I'm considering changing pricing from [CURRENT] to [PROPOSED] for [PRODUCT/SERVICE].
Current metrics:
- Customers: [NUMBER]
- Average tenure: [TIME]
- Churn rate: [RATE]
- Acquisition cost: [COST]
Walk me through:
1. Best case revenue impact
2. Worst case revenue impact (if churn accelerates)
3. Break-even churn rate
4. Key risks I should plan for
5. A test approach to validate before full commit
Why it works: Forces scenario thinking instead of just "what's the new price."
51. Invoice follow-up email
Write an invoice follow-up email for a [TIME] overdue invoice.
Client: [NAME]
Amount: [AMOUNT]
Relationship status: [GOOD / NEUTRAL / STRAINED]
Tone calibrated to the relationship. Don't apologize for asking. Make payment the easiest action they can take.
Under 100 words. Include a clear next step if they don't respond.
Why it works: "Don't apologize for asking" cuts the deferential tone that makes invoice emails ignorable.
52. Expense audit prompt
Here are my recurring monthly business expenses: [PASTE LIST]
Audit for:
- Subscriptions I likely don't use enough to justify
- Duplicate/overlapping tools
- Items that scaled with the business when they shouldn't have
- Categories that look high relative to typical [BUSINESS TYPE] benchmarks
- Low-effort consolidation opportunities
Rank by potential annual savings.
Why it works: Treats the expense list like an auditor would, not like a list to summarize.
53. Tax prep checklist
Build a tax prep checklist for [BUSINESS STRUCTURE: LLC / S-corp / sole prop / etc.] in [STATE].
Include:
- Documents to gather
- Deductions commonly missed by small businesses
- Records to keep for at least 3 years
- Questions to ask my CPA before filing
Note: this is a checklist, not legal/tax advice. Flag anything I should specifically verify with a professional.
Why it works: The "flag for professional verification" line is important — AI shouldn't be making tax calls solo.
54. Financial dashboard outline
Help me design a 1-page monthly financial dashboard for my [BUSINESS TYPE].
What I want to see at a glance:
- Top-line metrics (4-5)
- Trend indicators (up/down vs. last month and same month last year)
- Cash runway / position
- Key ratios that matter for [BUSINESS TYPE]
- One "what changed" callout section
Suggest the layout, what each metric should be, and the data sources I'd pull from.
Why it works: Forces clarity on the 4-5 metrics that actually matter instead of the dashboard-of-100-charts trap.
55. Budget vs. actual variance analysis
Here's my budget vs. actual for [PERIOD]: [PASTE]
Identify:
- Largest positive variances (and likely causes)
- Largest negative variances (and likely causes)
- Categories drifting in a concerning direction
- 2-3 questions I should investigate further
- Adjustments to consider for next period's budget
Be specific. Don't just summarize the numbers — interpret them.
Why it works: "Interpret, don't summarize" is the entire difference between useful and useless variance reports.
56. Forecasting prompt
Based on my last 12 months of monthly revenue: [PASTE]
Build a 6-month forecast with:
- Conservative scenario (current trajectory)
- Realistic scenario (with seasonal patterns factored in)
- Optimistic scenario (if Q[X] continues outperforming)
For each, show the assumptions clearly. Flag any data points that look unreliable.
Why it works: Three scenarios with explicit assumptions beat "here's the forecast" by a mile. For more, see How to use AI to manage your business finances.
Operations & Productivity (10 prompts)
57. SOP writer
Document a standard operating procedure for: [TASK]
How I currently do it: [DESCRIBE STEPS]
Structure the SOP with:
- Purpose (why this exists)
- When to use it (trigger)
- Inputs and tools needed
- Numbered step-by-step (each step under 15 words)
- Common errors and how to avoid them
- Quality check at the end
Write for someone who's never done this before.
Why it works: "Someone who's never done this before" is the right user — it forces clarity that experienced staff would skip.
58. Meeting recap
Here are my notes from a meeting: [PASTE]
Write a recap with:
- Date, attendees, purpose
- 3-5 key decisions made
- Action items (owner, deadline, status)
- Open questions / parking lot
- Next meeting date if discussed
Format clean. Under 300 words. No filler.
Why it works: "No filler" is the whole reason recaps fail — they're 80% restating the agenda.
59. Weekly review prompt
Help me run a weekly business review.
Inputs:
- Top 3 things I accomplished: [LIST]
- Top 3 things I didn't finish: [LIST]
- Energy / focus level: [1-10]
- Surprise of the week: [NOTE]
- Recurring friction: [NOTE]
Output:
1. One pattern you notice across this week and last
2. One thing I should keep doing
3. One thing I should stop doing
4. One question I should sit with before next week
Why it works: Catches patterns across weeks that solo founders miss in the moment.
60. Task prioritization
Here's my task list for the week: [PASTE]
Prioritize using Eisenhower matrix (urgent/important).
For each task:
- Quadrant (Q1-Q4)
- Suggested action (do now / schedule / delegate / drop)
- Why
Then surface:
- The 1 task I'm probably overestimating
- The 1 task I'm probably underestimating
- The 1 task that doesn't belong on this list at all
Why it works: The "doesn't belong" callout is gold — most prioritization exercises never question whether tasks should exist.
61. Project kickoff brief
I'm starting a project: [DESCRIBE]
Build a one-page kickoff brief covering:
- Why we're doing this (the business reason, not just the deliverable)
- Success metrics (specific and measurable)
- Scope: in / out
- Key milestones with dates
- Team and roles
- Risks and mitigations
- First 2-week action plan
Tone: concise. No corporate filler.
Why it works: Scope "in / out" is the highest-leverage section — most projects fail at "out" not "in."
62. Decision-making framework
I'm deciding between [OPTION A] and [OPTION B] for [CONTEXT].
Walk me through a structured decision:
1. The real question underneath the choice
2. Criteria that matter (with weights)
3. Scoring of each option against criteria
4. Hidden costs of each
5. What would have to be true for each to be right
6. The decision I'd recommend, with confidence level
Why it works: "The real question underneath the choice" reframes 30% of decisions before they're even scored.
63. Time audit
Here's how I spent last week (rough estimates): [PASTE OR DESCRIBE]
Analyze:
- Where my time went vs. where it should have gone
- The 2-3 activities consuming the most time with the least leverage
- Patterns suggesting context-switching is hurting me
- Tasks I could delegate, automate, or drop
- One change for next week that would compound
Be direct. Don't soften the feedback.
Why it works: "Be direct" is permission to call out the obvious truths solo founders avoid.
64. Delegation script
I need to delegate [TASK] to [PERSON].
Their experience level: [DESCRIBE]
What "done" looks like: [DESCRIBE]
Where they can find context: [LOCATION]
Write a delegation handoff that:
- Explains the "why" before the "what"
- Defines done explicitly
- Sets a check-in cadence
- Specifies what decisions they can make vs. need to escalate
- Leaves room for them to do it their way
Why it works: "Leaves room for them to do it their way" is what separates delegation from micromanagement-in-disguise.
65. Process improvement audit
Here's a recurring process in my business: [DESCRIBE]
Audit for:
- Steps that exist but no one questions
- Manual handoffs that could be automated
- Approval bottlenecks
- Information that's recreated rather than reused
- Frequency mismatches (daily when weekly would do, etc.)
For each finding, suggest a specific change and the rough effort to implement.
Why it works: "Steps that exist but no one questions" is the single highest-leverage process improvement lens.
66. Hiring vs. tooling decision
I'm considering hiring someone for [FUNCTION] vs. buying tools/AI agents to do it.
Current state: [DESCRIBE]
Volume: [DESCRIBE]
Budget for either: [BUDGET]
Walk me through:
- What hiring would solve that tooling can't
- What tooling would solve that hiring can't
- Hybrid options I should consider
- A 6-month and 18-month comparison of likely costs and outcomes
- A recommendation with explicit assumptions
Why it works: Reframes the binary choice into a hybrid spectrum — usually the right answer in 2026.
Strategy & Decision-Making (7 prompts)
67. Quarterly planning
I'm planning Q[X] for my [BUSINESS TYPE].
Context:
- Last quarter's biggest wins: [LIST]
- Last quarter's biggest misses: [LIST]
- Top business priority right now: [STATE]
- Constraints: [TIME, MONEY, TEAM]
Help me set:
- 3 quarterly objectives (outcome-led, not task-led)
- 2-3 measurable key results per objective
- The single most important thing to focus on
- 1-2 things to explicitly NOT do this quarter
Why it works: The "explicitly NOT do" line is the highest-leverage planning prompt most owners skip.
68. Competitive analysis
My business: [DESCRIBE]
My top 3 competitors: [LIST]
Build a competitive analysis covering:
- Their positioning in one sentence each
- Their strengths I should respect
- Their weaknesses I could exploit
- Where I'm directly comparable
- Where I'm in a different category entirely
- 1-2 strategic moves their playbook suggests they'll make next
Tone: analytical, not dismissive of competitors.
Why it works: "Strengths I should respect" prevents the founder's trap of underestimating competition.
69. Pivot or persevere
A specific initiative isn't working: [DESCRIBE]
Help me decide whether to pivot or persevere:
- The original hypothesis behind this
- What evidence supports continuing
- What evidence suggests pivoting
- Sunk cost factors I might be overweighting
- The minimum experiment to test before committing more
- The decision framework I should use
End with: a recommendation, plus what would change your recommendation.
Why it works: "Sunk cost factors I might be overweighting" is the question pivot decisions hinge on.
70. New offer validation
I'm thinking of launching [NEW OFFER] for [AUDIENCE].
Stress-test this idea:
- The strongest reasons it might work
- The strongest reasons it might fail
- Who specifically would pay for this (with archetype + willingness to pay)
- The smallest version I could test in 30 days
- Three signals that would confirm demand
- Three signals that would indicate I should kill it
Don't try to be supportive. Be useful.
Why it works: "Don't try to be supportive. Be useful." is the most underrated phrase in business AI prompting.
71. Strategic narrative
Help me articulate my business's strategic narrative.
Background: [WHAT I DO, WHY I STARTED, WHO I SERVE]
Build a narrative with:
- The change happening in the world (the "why now")
- What that change means for my audience
- Why current solutions fall short
- How my approach is different
- The future I'm building toward
Tone: confident, not grandiose.
Why it works: The "change in the world" opener is what separates compelling narratives from feature lists.
72. Risk register
Build a risk register for my [BUSINESS TYPE].
Categories: financial, operational, market, technology, key person, regulatory.
For each top risk:
- Likelihood (low/medium/high)
- Impact if it happens (low/medium/high)
- Current mitigation
- One additional mitigation to consider
- Trigger signals to monitor
Rank top 5 by combined likelihood × impact.
Why it works: "Trigger signals to monitor" turns the risk register from a static doc into a live system.
73. Annual review prompt
Help me reflect on this past year in business.
Context:
- Top 3 wins: [LIST]
- Top 3 losses or learning moments: [LIST]
- Where the business is now vs. one year ago: [DESCRIBE]
- Things I almost did but didn't: [LIST]
- One thing I'd tell myself a year ago: [WRITE IT]
Synthesize:
1. The theme of the year
2. The most important lesson
3. The pattern I should notice
4. What this year suggests about next year's focus
5. A single sentence to anchor the year ahead
Why it works: "Things I almost did but didn't" surfaces what years of journaling typically miss.
SEO & AI Search (6 prompts)
74. Keyword research prompt
Act as an SEO strategist focused on [INDUSTRY/NICHE].
For my [BUSINESS TYPE] targeting [AUDIENCE], identify:
- 5 head terms (high volume, high competition)
- 10 mid-tail keywords (decent volume, achievable competition)
- 15 long-tail keywords (specific intent, lower competition)
- 5 question-based queries (great for AI search)
For each keyword, note the likely search intent (informational, navigational, transactional) and the type of content that wins.
Why it works: Search intent categorization is what separates rankable lists from generic keyword dumps.
75. AI search optimization audit
Help me make this article more likely to be cited by ChatGPT, Perplexity, and Google AI Overviews.
Current article: [PASTE OR LINK]
Audit for:
- Direct answer placement (within first 100 words)
- Question-formatted H2s
- Extractable factual statements
- Structured data opportunities (FAQ, How-To, etc.)
- Authority signals (citations, statistics, expert quotes)
- Format gaps (missing tables, lists, comparison structures)
Suggest specific edits with the rationale for each.
Why it works: This is the exact framework AI engines use to evaluate citation-worthiness. For the full strategic context, see Generative Engine Optimization for Small Businesses.
76. Meta description writer
Write 5 meta descriptions for an article titled: "[TITLE]"
Constraints:
- Under 155 characters each
- Each takes a different angle (benefit-led, question-led, contrarian, specific-promise, list-tease)
- Include the primary keyword naturally
- No clickbait or false specificity
- Optimized for click-through, not just SEO
Why it works: Five different angles let you test what actually drives clicks.
77. Internal linking audit
Here's a new article I'm publishing: [PASTE]
And here's a list of my existing articles: [PASTE TITLES AND URLS]
Suggest:
- 5-10 internal links from the new article to existing articles, with the anchor text and where they'd fit naturally
- 3-5 existing articles that should link TO this new one (with suggested updates)
- Any topical gaps suggesting articles I should write next
Why it works: Builds the topic cluster proactively instead of leaving internal linking as an afterthought.
78. Content refresh prompt
Here's an article I published 18 months ago: [PASTE OR LINK]
It currently ranks #[X] for "[KEYWORD]" and gets [TRAFFIC].
Audit for:
- Outdated facts, tools, or statistics
- Sections that no longer reflect best practice
- New subtopics that should be added based on what's relevant in 2026
- Restructuring opportunities for AI search
- Internal links to newer content I should add
Output as a refresh plan, not a rewrite.
Why it works: "Refresh plan, not rewrite" preserves SEO equity instead of starting from scratch.
79. SERP analysis
I'm trying to rank for "[KEYWORD]". Here are the top 10 results: [PASTE TITLES, METAS, AND URLS]
Analyze:
- Common patterns across the top results (length, structure, angle)
- What angles are saturated
- What's missing or underrepresented
- The single biggest content gap
- A specific article angle that could realistically rank if executed well
Why it works: "Saturated vs. missing" is the right framing for entering a competitive SERP.
How to save and reuse these prompts
Having 75+ prompts is useless if you can't find them when you need them. Three approaches that work:
1. ChatGPT Projects or Custom GPTs. Save your top 10–15 prompts as starter messages in dedicated projects (e.g., "Marketing Prompts," "Sales Prompts"). Each project keeps its context separate, so your marketing chats don't drift into operations.
2. A prompt library doc. A simple Notion or Google Doc with prompts organized by category. Tag each one with the model that works best (ChatGPT, Claude, Gemini) — they each have strengths.
3. Text expansion shortcuts. Tools like Raycast, Alfred, Espanso, or TextExpander let you trigger long prompts with a short keystroke. Type a short abbreviation and the full prompt appears instantly.
The one rule that matters more than any prompt
These prompts are starting points. The biggest mistake small business owners make is treating them as one-shot questions instead of conversations. The best output usually comes after 2–3 rounds of refinement: "Make it shorter," "More casual tone," "Add a CTA."
Don't expect perfection on the first response. Expect a strong draft you can work with. That's still 10 times faster than starting from scratch.
Pick the three prompts most relevant to the work you're doing this week. Save them somewhere you'll actually find them. Use them. That's it.
Get the week's most important AI developments for business owners — every Monday morning, free.