Curated AI prompts for business, creative, and technical use cases.
Claude Sonnet 5 (works well with GPT-5.4 and Gemini 2.5 Pro too — any strong reasoning model handles cross-domain analogy well)learningA product manager who has spent 8 years mastering restaurant kitchen operations gets assigned to work with a data engineering team and needs to understand how database indexing works well enough to have credible technical conversations in sprint planning — without taking a 6-week course first.
Generic explanations don't stick because they're written for nobody in particular. This prompt maps a new concept onto a domain you already understand deeply, sub-concept by sub-concept.
prompt-engineeringlearning
Claude Sonnet 5 (also works well with GPT-5.4 and Gemini 3 Pro — any frontier model with strong reasoning handles the multi-step negotiation logic)finance-negotiationYou're a mid-level individual contributor who was just told in a 15-minute call that your role is being eliminated in a company-wide restructuring. HR emails you a severance agreement the same afternoon with a 5-business-day deadline to sign, and you have never negotiated anything like this before.
Most laid-off employees sign their first severance offer without countering it. This prompt builds a complete negotiation plan — priority asks, a ready-to-send script, and anticipated objections — from the offer you actually received.
CareerNegotiation
Claude Opus 4.8 or GPT-5 (works well with any frontier model with strong creative writing ability)creativeYou are a corporate trainer who needs new hires to understand a technical concept (e.g., zero-trust security, API rate limiting, why cloud costs scale non-linearly) before their first week ends, but the official documentation puts people to sleep and a straight lecture format gets zero retention in a 15-minute onboarding slot.
Straight explanations bounce off people who have heard the jargon before and tuned out. This prompt turns any dense topic into a dialogue between two characters who would never normally discuss it — and the friction makes the concept stick.
onboardingcreative-writing
Claude Opus 4.8 or GPT-5Startups & BusinessYour company entered a cost-reduction cycle and your VP asked you to find 20-25% savings in your department budget. You have 48 hours before the leadership review, a spreadsheet with 30 line items, and no formal framework for distinguishing safe cuts from dangerous ones.
Most budget cuts are made on raw numbers, not impact analysis. This prompt gives you a CFO-level framework that models what each cut actually costs long-term — before you eliminate anything you will regret.
strategystartup
Claude Sonnet 4.6Software & AppsYou have 45 minutes before your weekly newsletter is due. Your notes are a mess of half-formed bullets from five browser tabs and a Notion doc. The prompt takes your raw material and turns it into a complete, voice-consistent issue — opening hook, commentary on each item, and a closing that ties everything together.
A structured AI prompt that transforms your raw weekly notes into a complete, voice-consistent newsletter issue — opening hook, item commentary, and closing — in under 10 minutes.
productivitywriting
Claude Opus 4.8 or GPT-5Artificial IntelligenceYou have a performance review or promotion conversation coming up in the next 4-8 weeks. You have a mental list of things you have accomplished this year, but no structured way to present them as a case for advancement — and no idea how to frame your future value in the language your organization uses to make level decisions.
Most capable people lose promotions not because they underperform but because they cannot articulate their value in the language organizations use to make advancement decisions. This prompt builds a complete, manager-ready promotion case document from your own notes in a single pass.
Careerperformance-review
Claude Sonnet 4.6 or GPT-5.5 (any frontier model with long context)Developer ToolsYour team just decided to switch your mobile API from REST to GraphQL after a performance review, and the lead architect is leaving for three weeks tomorrow. You need a complete decision record committed to the repo before institutional memory walks out the door.
Most engineering teams make dozens of significant technical decisions every year and document almost none of them. This prompt generates a complete Architecture Decision Record — the format senior engineers and tech leads use to capture why a decision was made, what was considered, and what it means for future work.
software-designcoding
Claude Opus 4.7 or GPT-5.5 (works with any frontier model)Data AnalysisYou ran a 200-person customer satisfaction survey after your product launch and have a spreadsheet of responses — NPS scores, feature ratings, and 80 open-ended comments. You need a structured analysis ready for a board meeting in two hours.
Most survey results die in a spreadsheet. This prompt turns raw responses into a structured executive report — with statistical patterns, sentiment analysis, key themes, and prioritized recommendations your team can act on.
prompt-engineeringdata analysis
Claude Sonnet 4.6 / GPT-5.4 (also works well with Gemini 2.5 Pro)Startups & BusinessYour primary project management tool renewal lands next month. The account manager just emailed about 'updated pricing for FY27.' You have a call in two days, no prep done, and the contract is $60,000/year. Fill in the brackets and run this prompt — you will have a full negotiation brief before the call.
Walk into any vendor renewal call with a complete strategy: leverage points, BATNA, opening script, and a three-tier negotiation ladder — all generated from a 2-minute prompt.
Claude Sonnet 4.6 / GPT-5.4marketingYou are launching a paid campaign for a new product next week and need five distinct ad copy variants across Google and Meta — each testing a different creative angle — so your media buyer can run structured A/B tests and optimize spend based on what is actually being measured, not gut feel.
Most ad campaigns waste budget because variants have no clear test hypothesis. This prompt generates five platform-ready ad copy variants for Google, Meta, or LinkedIn — each testing a different variable, with a built-in hypothesis so your media buyer knows exactly what to optimize.
google-adsmarketing
Claude Sonnet 4.6CareerYou have your mid-year or annual performance review due in 48 hours. You have a messy Notion doc full of project names, a few Jira ticket summaries, and a vague memory of "fixing a lot of bugs." You need to turn this into 8–12 polished bullets that make your work legible to an HR calibration committee who will never read the PRs.
Stop staring at a blank performance review form. This prompt takes your rough notes about what you did this year and rewrites them as quantified, impact-first bullets using the XYZ formula — the format managers and HR actually remember.
productivityCareer
Claude Sonnet 4.6 / GPT-5 / Gemini 3.5 FlashlearningYou have pulled together 4–7 articles, papers, or reports on a topic before writing a brief, making a decision, or presenting to stakeholders — and you need to understand the actual state of evidence across them, not just what each article says on its own.
Stop reading sources one-by-one and ending up holding whichever conclusion you read last. This prompt maps the actual state of evidence across multiple sources — agreements, genuine disputes, unique contributions, and open gaps.