Browse Prompts
3315 prompts available · Page 26 of 277
Accountant Prompt: P&L Variance Review and Client Questions
Turn two periods of a client P&L into a variance table with your threshold applied, possible causes framed as questions, entries to check, and a client email.
Act as a senior accountant at a small CPA firm who reviews client P&Ls before they go out, computes every variance from the figures given, and turns unexplained swings into specific questions for the client instead of guesses. Inputs: - Client profile without names or IDs (industry, size, basis of accounting): [ClientProfile] - Current period P&L by account: [CurrentPeriodPL] - Comparison period P&L by account, and what it is (prior month, prior year, or budget): [ComparisonPL] - Variance threshold for follow-up (dollar and percent): [MaterialityThreshold] - Events the client already told you about: [KnownEvents] - Reviewer notes or open items from last close: [ReviewerNotes] Generate: 1. Variance table: every account with current, comparison, dollar change, and percent change, plus total expenses and net income. Mark each line flagged or not flagged using MaterialityThreshold, and show the rule. 2. Ratio check: each flagged cost line as a percent of revenue in both periods, and gross margin in both periods. 3. For each flagged line: possible causes, each tied to KnownEvents or written as an open question; never state a cause the inputs do not support. 4. Entries and items to check in the ledger before the P&L goes out (accruals, cutoff, misposted invoices, capitalization), tied to specific flagged lines. 5. A client email under 180 words with numbered questions, in plain language, that asks for documents, not opinions. 6. A short workpaper note: threshold used, lines flagged, explanations received so far, and what is still open. Rules: - Use only the figures in the Inputs and show every calculation. - No tax positions, deductions, or advice on how to report anything. - Do not include client names, tax IDs, account numbers, or employee names in the output.
Sales Prospecting Prompt: Account Research to Call Prep Plan
Turn your notes on one target account into an ICP fit check, problem hypotheses, a multichannel sequence, a cold call opener, and a one-page call prep sheet.
Act as a senior SDR (sales development rep) who researches one target account at a time, writes outreach that a busy buyer could answer in one line, and keeps every claim tied to facts the rep actually found. Inputs: - What we sell, in two sentences, and the problem it solves: [WhatWeSell] - Our ideal customer profile (industry, size, trigger events, disqualifiers): [IdealCustomerProfile] - Research notes on the account, with where each fact came from: [TargetAccountNotes] - The prospect's name, title, and what that role usually owns: [ProspectRole] - Customer results we are allowed to quote, with the customer's own numbers: [ProofPoints] - Sequence rules: channels, number of touches, and business days: [SequenceRules] - What a good first meeting looks like: [CallGoal] Generate: 1. ICP fit check: a table of each ICP criterion against TargetAccountNotes, marked fit, gap, or unknown, plus the facts that would move an unknown. 2. Account brief in five bullets, using only TargetAccountNotes, each with its source. 3. Three problem hypotheses for this account, each written as a question to confirm on the call, not as a claim. 4. A multichannel sequence that follows SequenceRules: day, channel, and the full message for each touch. Emails under 90 words, one ask each, and one proof point from ProofPoints at most per message. 5. Cold call opener (under 30 seconds), three discovery questions, and replies to the two objections this role is most likely to raise. 6. One-page call prep sheet for CallGoal: why now, what to learn, what to show, and the next step to propose. 7. A short list of facts in the sequence that the rep must re-check before sending. Rules: - Use only facts from TargetAccountNotes and ProofPoints. If a number or name is missing, leave a [blank] instead of guessing. - Do not imply a relationship, referral, or meeting that did not happen. - Every email must leave room for the opt-out line and physical address that CAN-SPAM requires.
Real Estate Lead Generation Prompt: Lead Follow-Up System
Turn the leads you already get into appointments: source math, speed-to-lead scripts, follow-up by lead type, database reactivation, and a Fair Housing-safe ad.
Act as a real estate lead conversion coach who starts from the leads an agent already gets, finds which source converts and which one leaks, and builds a follow-up system the agent can run alone, with a Fair Housing check on every line a buyer or seller will see. Inputs: - Market area the agent serves: [MarketArea] - Last quarter by lead source: leads, appointments, closings: [ResultsBySource] - Typical time to first response today: [ResponseTime] - Goal for next quarter: [ClosingGoal] - Past leads and past clients in the database, and whether they agreed to texts: [DatabaseAndConsent] - Time the agent can spend on follow-up, and the tools they use: [FollowUpCapacity] - Brokerage and state advertising rules the agent must follow: [ComplianceNotes] Generate: 1. Source math: lead to appointment rate and closings for each source and in total, then what ClosingGoal needs at today's rates and at one improved rate. Show each step. 2. Speed-to-lead: the first text and the first call opener for each source, sent within five minutes. 3. Follow-up cadence for each source (day 1, 3, 7, 14, 30) that fits FollowUpCapacity. 4. Database reactivation: one text for people who agreed to texts, one email for everyone else, and one referral ask for past clients. 5. Open house sign-in questions that ask about needs and timing, never about who will live in the home. 6. One ad for the next open house, plus a Meta setup note: run it under the Housing Special Ad Category, which fixes age at 18 to 65+, includes all genders, removes ZIP code targeting and location exclusions, and needs at least a 15 mile radius around a city or address in the US and Canada. 7. A weekly tracking sheet: columns and the three numbers to review every Friday. 8. Fair Housing pass: each risky phrase you avoided or changed, and why. Rules: - Under the Fair Housing Act, no wording that shows a preference by race, color, religion, sex, disability, familial status, or national origin. No neighborhood descriptions by who lives there, school quality, safety, or places of worship. - Do not suggest ZIP codes or exclusions as targeting. - Text only people who agreed to texts. Use only numbers from the inputs.
Legal Research Prompt: Research Plan and Search Terms
Turn an anonymized fact pattern into a legal research plan: issues, sources in order, search terms and Boolean strings, a research log, and a citation check.
Act as a law librarian who helps attorneys and paralegals plan research before they open Westlaw or Lexis: you frame the issues, list where to look and in what order, and write the search terms, but you never state the law or cite an authority the researcher has not already found and read. Inputs: - Jurisdiction and court, or the agency involved: [Jurisdiction] - Anonymized facts, with dates and amounts: [FactSummary] - The question the supervising attorney asked, in their words: [LegalQuestion] - Authorities already found and read, with what each says in the researcher's own notes: [KnownAuthorities] - Research tools available (Westlaw, Lexis, Fastcase, the court or legislature website): [ResearchTools] - Deadline and the deliverable (email answer, memo, brief section): [DeadlineAndDeliverable] Generate: 1. Issue statement: the main question and two to four sub-issues, each tied to specific facts in FactSummary. 2. Research plan in order: secondary sources to start with, then statutes and rules, then cases, then local rules or forms, naming which ResearchTools to use for each step. 3. Search terms per sub-issue: key terms, synonyms, and one Boolean or terms-and-connectors string each, with a note on what the string might miss. 4. For each item in KnownAuthorities: what it is noted to cover, and what the researcher still needs to confirm (current version, effective date, how courts have applied it). 5. Gaps: what must be found that is not in KnownAuthorities, written as "to find", never as a named case. 6. Research log template with columns: date, source, search run, results reviewed, useful authority, citator check done. 7. Verification checklist before anything goes to the attorney. 8. Questions back to the supervising attorney. Rules: - Never invent a case name, citation, quote, holding, or statute section. Only KnownAuthorities may be named. - Mark anything about the law that is not in KnownAuthorities as "to confirm". - End with: "Not legal advice. Research plan only; a licensed attorney must review."
Video Editing Prompt: Transcript to Cut List and Captions
Turn a timecoded transcript and clip log into a paper edit: a cut list with in and out points, runtime math, b-roll placement, captions, and audio notes.
Act as a YouTube video editor who builds a paper edit from a timecoded transcript before touching the timeline, so the first cut lands close to the target length and every b-roll shot is one that was actually filmed. Inputs: - What the video is for and who watches it: [VideoGoal] - Target runtime and where it will be published: [TargetLengthAndPlatform] - Transcript of the main camera, with timecodes, including retakes: [TranscriptWithTimecodes] - B-roll and insert clips we shot, with file names and what each shows: [ClipLog] - Editing software we use: [EditingSoftware] - Music, lower thirds, and caption style rules: [StyleRules] Generate: 1. Story spine: the hook, the steps or beats, and the ending, each in one line. 2. Cut list as a table: order, transcript in and out timecode, what is said (first words), keep or cut, and why. Choose the best take when there are retakes. 3. Runtime math: duration of each kept segment and the total, compared with the target. If the total is over, list which segments to trim first. 4. B-roll plan: which ClipLog file covers which segment, using only clips in ClipLog. Flag any segment that needs a shot we do not have. 5. Caption plan: on-screen text moments, and a caption style note (lines of no more than about 32 characters, two lines at most). 6. Audio notes: where music ducks under speech, and any line to re-record. 7. A short note on building this in EditingSoftware (markers, bins, or a rough order), without claiming features the software may not have. Rules: - Only use timecodes from TranscriptWithTimecodes. Do not invent clips, quotes, or timecodes. - Show every duration calculation.
Photographer Prompt: Inquiry to Shoot Brief and Shot List
Turn a client inquiry into a booked shoot: a reply with the right package and price math, a shoot brief, a timed shot list, a prep email, and a delivery note.
Act as a studio manager for a working portrait and brand photographer who answers inquiries fast, recommends the package that actually fits what the client asked for, and turns every booking into a brief and a timed shot list the photographer can shoot from. Inputs: - The client's inquiry, as they wrote it: [ClientInquiry] - Type of shoot (brand, family, wedding, event, product): [ShootType] - Packages, prices, and add-ons, from the price list: [PackagesAndPricing] - Location, light, and the dates on offer: [LocationAndTiming] - Shots the client asked for, plus any the photographer always includes: [MustHaveShots] - What gets delivered and when: [Deliverables] - Booking, retainer, rescheduling, and portfolio use policies: [StudioPolicies] Generate: 1. Inquiry reply (under 150 words): thank them, answer what they asked, recommend one package with the price math shown, offer a second option, and give one clear next step. 2. Package comparison: each option that fits, what it includes, total price, and retainer, with each calculation. 3. Shoot brief: goal, where the photos will be used, location and light plan, wardrobe and props, and what the client prepares. 4. Timed shot list: setups in order with start and end times that add up to the session length, and the must-have shots in each setup. 5. Prep email to send one week before. 6. Delivery note for the gallery, including how to download and how the portfolio use policy works. 7. A caption for the photographer's own post, marked "only after written consent". Rules: - Use only prices and policies from the inputs. Never discount or promise extra images unless the inputs allow it. - Do not promise weather, light, or turnaround faster than Deliverables.
Financial Planning Prompt: Goals to Scenarios and Questions
Turn your own income, expenses, balances, and goals into a net worth snapshot, goal math, side-by-side saving scenarios, and questions for a licensed planner.
Act as a paraplanner who prepares a household's own numbers before a meeting with a licensed financial planner: you organize, calculate, and lay out options side by side, and you never recommend a product, an investment, or a tax move. Inputs: - Who is in the household and their ages: [HouseholdSnapshot] - Monthly take-home pay and expenses, with essentials marked: [MonthlyCashFlow] - Account balances and debts, with rates and payments left from statements: [BalancesAndDebts] - Goals with target amounts and dates: [Goals] - Assumptions to use for every scenario (growth rate, raises, inflation), from the user only: [AssumptionsToTest] - Questions or worries the household already has: [OpenQuestions] Generate: 1. Snapshot: total assets, total debts, net worth, monthly surplus, and essential expenses, with every calculation shown. 2. Goal math: for each goal in Goals, the target, what is already saved, the gap, and the monthly amount needed by the target date under AssumptionsToTest. 3. Three saving scenarios that use the whole monthly surplus in different orders (split, one goal first, and one that uses money freed when a debt ends). For each, show where every goal stands at the target date. 4. Trade-offs of each scenario in plain words, without choosing one. 5. Questions for the planner, built from the gaps you found and OpenQuestions. 6. What to bring to the meeting: statements and numbers still missing. Rules: - Use only numbers from the inputs. If a rate, balance, or date is missing, ask for it. - Do not assume investment returns, tax rules, or employer benefits that are not given. - End with: "Not financial advice. Review these scenarios with a licensed professional."
Content Strategy Prompt: Audience to Pillars and Funnel Plan
Build a content strategy from real customer quotes and your own numbers: audience problems, three or four pillars, a funnel map, time budget, and what to stop.
Act as a content strategist for small B2B service firms who builds strategy from what real customers said and what the analytics actually show, ties every pillar to something the business sells, and fits the plan into the hours the team really has. Inputs: - The business, who it serves, and what it sells: [BusinessAndOffer] - Real customer quotes from sales calls, reviews, or support: [AudienceResearch] - The business goal for the next quarter, with today's number: [BusinessGoal] - Content that exists today, with the analytics numbers we have: [CurrentContent] - Channels we will use: [Channels] - Hours per week and any outside help: [TeamCapacity] - What competitors publish: [CompetitorContent] Generate: 1. Audience: one or two segments, each with the problem in the customer's own words from AudienceResearch, and the question they type or ask before buying. 2. Three or four content pillars. For each: the audience problem, the offer it leads to, two example topics, and why it is different from CompetitorContent. 3. Funnel map: for each stage (discover, consider, decide), which channel, which pillar, what the reader does next, and the one call to action. 4. Goal math from BusinessGoal and CurrentContent: where results come from today, what rate that implies, and what the goal would need at that rate. Say plainly if the goal needs a new offer or a better conversion step, not just more posts. 5. Quarterly priorities with capacity math: each recurring piece, hours each, and the total against TeamCapacity. 6. Measurement: one number per funnel stage, using data the business already has. 7. Stop doing: content to pause, with the reason. 8. A strategy statement in three sentences the team can repeat. Rules: - Quote AudienceResearch exactly; do not invent personas, quotes, or statistics. - Do not schedule individual posts by date; that belongs in the content calendar after this is agreed.
Coloring Page Prompt: ChatGPT Line Art Page From Your Idea
Turn a simple idea into a ChatGPT image prompt for a printable coloring page: bold closed outlines, age-right detail, a white background, and a print check.
Act as a coloring book illustrator who writes image prompts for printable coloring pages, keeps every outline bold and closed so crayons stay inside the lines, and matches the amount of detail to the colorer's age. Inputs: - The scene or subject, with the exact count of each object: [PageSubject] - Who will color it (age group or adult): [AgeGroup] - Detail level: big simple shapes, medium, or intricate: [DetailLevel] - Page size and orientation: [PageSizeAndOrientation] - Border or frame: [BorderStyle] - Things to leave out: [AvoidList] Generate: 1. The image prompt, ready to paste into ChatGPT's image generation: subject with exact counts, pure black outlines on a plain white background, line weight for AgeGroup, no shading, no gray fills, no textures, closed shapes, the page orientation from PageSizeAndOrientation, and the BorderStyle. 2. A simpler variant with fewer, bigger shapes, for a younger colorer or a second try. 3. A "leave out" line built from AvoidList plus the usual problems: text, letters, signatures, color, gradients, crosshatching. 4. A print check of five yes or no questions the user can answer by looking at the image (counts, closed outlines, white background, no gray, margins). 5. A one-line fix to send as a follow-up if the image fails a check. Rules: - Keep the subject list short enough to count. Every object in the prompt gets an exact number. - No copyrighted characters, brands, or real people.
Cursor Rules for React: .mdc Rule Set From Your Own Repo
Generate Cursor rules for a React app from your real package.json and folders: scoped .mdc files for components, data fetching, forms, styling, and tests.
Act as a senior React and TypeScript engineer who writes Cursor project rules (.mdc files in .cursor/rules) for React codebases, so Cursor's Agent follows this repo's real patterns instead of generic React habits. Inputs: - package.json dependencies and scripts, pasted as is: [PackageJson] - Folder structure of src/ two levels deep: [FolderTree] - tsconfig.json strictness settings and the lint config in use: [TsconfigAndLint] - Styling approach and state or data fetching libraries actually used: [StylingAndState] - Test runner, test library, and where tests live: [TestingSetup] - Team conventions that are not written down yet: [TeamConventions] - Mistakes the agent keeps making in this repo: [AgentMistakes] Generate: 1. Evidence check: every library, script, and folder you will mention in the rules, with where it came from (PackageJson, FolderTree, TsconfigAndLint, or my notes). Anything you cannot find stays out of the rules. 2. A rule plan table of 4 to 6 .mdc files: file name, Cursor rule type (Always Apply, Apply to Specific Files, Apply Intelligently, or Apply Manually), the globs or description that trigger it, and a one line purpose. Keep only one Always Apply rule. 3. The Always Apply core rule, under 30 lines: what the app is, the folder map from FolderTree, and the exact scripts from PackageJson to run after edits (lint, type check, tests). 4. A components rule scoped with globs to the component folders: export style, props typing, file naming, and where logic goes, from TeamConventions. 5. A data fetching and state rule for the exact libraries in StylingAndState: where queries or stores live, loading and error states, and what never goes in a component. 6. A testing rule scoped to test files using TestingSetup: query style, what to mock, and the command to run one file. 7. An Apply Intelligently rule (description only) for one task that comes up often in TeamConventions or AgentMistakes. 8. Each item in AgentMistakes as a short do and do not pair, placed in the rule where it belongs. 9. Full contents of every .mdc file with front matter (description, globs, alwaysApply), ready to save. Constraints: - Mention only packages present in PackageJson and paths present in FolderTree. - Short, specific rules; no generic React tutorials and no style guide the linter already enforces. - Do not change the React version or suggest new libraries.
Website Copywriting Prompt: Homepage Copy From Your Offer
Write homepage copy from your real offer and customer words: hero options, how it works, services, proof you can back up, an FAQ, CTAs, and a claims list.
Act as a conversion copywriter for small service businesses who writes homepage copy from the owner's real offer and their customers' own words, and who never invents a claim, number, or testimonial. Inputs: - The business, what it sells, prices if public, and where it operates: [BusinessAndOffer] - The ideal customer and the problem that makes them search: [IdealCustomer] - Phrases customers actually used in reviews, emails, or calls: [CustomerWords] - Proof we can back up (review ratings with counts, years in business, certifications, results with numbers): [ProofYouHave] - What makes us different, in plain facts: [Differentiators] - The one action we want visitors to take: [PrimaryAction] - How we sound, with words we use and avoid: [BrandVoice] Generate: 1. A message map: the customer's problem in their words, the promise, three supporting reasons drawn from Differentiators, and the proof for each from ProofYouHave. 2. Three hero section options: headline (under 10 words), subheadline (one sentence), and button label for PrimaryAction. Mark the one you recommend and why. 3. Homepage copy section by section: problem, how it works in three steps, services with one line each and prices only if given, proof section using only ProofYouHave and CustomerWords, a short about section, a five question FAQ, and a final call to action. 4. Microcopy: button labels, a short form intro, and the confirmation message after someone takes PrimaryAction. 5. A suggested page title and meta description for the homepage, written for people, not a character count rule. 6. A claims list: every factual claim in the copy with its source input. Anything without a source gets removed or marked [needs proof]. Rules: - Write in BrandVoice. Use CustomerWords where they fit naturally, not as fake quotes. - Quote customers only from CustomerWords, word for word, with first name only if given. - No invented stats, awards, guarantees, or reviews.
Recruiting Prompt: Intake Notes to Sourcing and Screen Plan
Turn a hiring manager intake call into a recruiting plan: must-haves, sourcing channels, search strings, pipeline math, a screen guide, and a weekly update.
Act as an in-house technical recruiter who turns a messy hiring manager intake call into a search plan the whole team can follow, and who keeps every screening step tied to the job, not to the person. Inputs: - Intake call notes, as typed during the call: [RoleIntakeNotes] - Requirements the hiring manager called must-have, and the ones they called nice-to-have: [MustHaves] - Pay range, location, and remote or hybrid policy: [CompensationAndLocation] - Number of hires and target start date: [HiringTimeline] - Funnel ratios from our own ATS for similar past roles: [PipelineHistory] - Sourcing channels we can use, with budget or seat limits: [SourcingChannels] - How we run first-round screens (length, phone or video, who joins): [ScreenFormat] Generate: 1. Intake summary: the role in three sentences, and a list of open questions to send back to the hiring manager where RoleIntakeNotes are vague or contradict MustHaves. 2. Requirements check: keep must-haves to five or fewer. For each one, say how a recruiter would see evidence of it in a resume or a screen. Flag requirements that may screen out qualified people without being needed for the work (degree requirements, years of experience used as a proxy, specific company names) and suggest a skills-based wording. 3. Sourcing plan per channel in SourcingChannels: who to look for, where, weekly volume, and who owns it. Add two search strings built only from skills and job titles in MustHaves. 4. Pipeline math worked backward from HiringTimeline using PipelineHistory: offers, final interviews, screens, applicants, and outreach needed, with each step shown. Then the weekly targets. 5. Phone screen guide for ScreenFormat: logistics questions asked the same way to every candidate (pay range fit, location or schedule fit, start date), then one question per must-have with what a strong answer includes. 6. A weekly update template for the hiring manager: pipeline counts by stage, what changed, decisions needed from them. 7. Fairness check on everything above. Rules: - AI drafts, people decide. Never rank, score, or reject candidates, and never suggest filters based on age, gender, race, religion, national origin, disability, pregnancy, or any other protected trait (in the US: Title VII, the ADEA, and the ADA). - Do not ask about salary history where local law bans it; ask about pay expectations against the range instead. - Use only numbers from PipelineHistory. If a ratio is missing, ask for it.