AIO Apex
en

Prompt & Skill Introduction

Curated AI prompts for business, creative, and technical use cases.
The Case Study Harvester: Turn a Customer Win Into a Case Study That Sells Itself
Claude Sonnet 4.6 or GPT-5.4 (needs strong long-context synthesis to avoid fabricating details not in the source material). Gemini 2.5 Pro works as an alternative. Avoid smaller/faster models — hallucination risk is the main failure mode this prompt guards against.marketingA product marketer at a 60-person B2B SaaS company just got off a call with a happy customer whose renewal came in at 140% of their original contract value. The customer success rep forwarded three Slack messages, a call transcript, and a usage dashboard screenshot — but no one has time to turn that into a polished case study before the sales team's QBR next week, and the marketer doesn't want to misquote the customer or invent numbers that could embarrass the company if the customer ever reads the published piece.

The Case Study Harvester: Turn a Customer Win Into a Case Study That Sells Itself

A structured prompt that converts messy call notes, Slack messages, and usage metrics into a publish-ready case study — while flagging any number it can't verify from your source material instead of inventing one.

B2Bmarketing
The Skip-Level Meeting Planner: Turn Rare Face-Time With Your Boss's Boss Into a Career-Advancing Conversation
Works well with any reasoning-capable model since the task is structured business communication rather than deep technical reasoning — Claude Sonnet 5 or GPT-6.1 Sol are both strong choices. Avoid giving this to a lightweight/flash-tier model if your context section is long or nuanced, since filtering accomplishments by relevance requires genuinely reading and weighing the details rather than pattern-matching on keywords.CareerA senior engineer at a 200-person company has a rare 30-minute skip-level meeting scheduled with the VP of Engineering next week — the first time they've had direct access to anyone above their own manager in two years at the company.

The Skip-Level Meeting Planner: Turn Rare Face-Time With Your Boss's Boss Into a Career-Advancing Conversation

A career prompt that filters your real accomplishments down to what a skip-level manager actually cares about, drafts questions designed to surface information only they can give you, and ends in a follow-up plan — not just meeting notes.

workplace communicationcareer-development
The Dependency Upgrade Risk Assessor: Turn a Package Bump Into a Breaking-Change Checklist
Works best with models that can reason over structured technical text and resist over-confident guessing — Claude Opus 5.5 or Sonnet 5 for weighing ambiguous changelog language against your stated usage patterns. GPT-6.1 Sol and Gemini 4 Argon also handle the classification task well. Avoid lightweight/flash-tier models for this one — they tend to flatten "NEEDS MANUAL VERIFICATION" into false confidence rather than genuinely flagging ambiguity.Developer ToolsA backend team at a mid-size SaaS company needs to bump a payment-processing SDK from v3 to v4 before a compliance deadline in three weeks. The vendor's changelog is 40 bullet points long, written for a general audience, and doesn't map cleanly onto their specific integration, which only touches 6 of the SDK's 30 exposed methods.

The Dependency Upgrade Risk Assessor: Turn a Package Bump Into a Breaking-Change Checklist

A developer-tools prompt that scopes a library upgrade's changelog to your actual usage, classifies each change as breaking or behavioral, and ends in a rollback plan with a measurable trigger instead of a vague "watch for issues."

dependency-managementrelease-engineering
The Partnership Due Diligence Brief: Turn a Proposed Channel Deal Into a Risk-Weighted Go/No-Go
Works well with models that handle structured business reasoning and will push back on weak assumptions — Claude Opus 5.5 or Sonnet 5 for the full risk-weighting pass, GPT-5.4 as a solid alternative. Lighter or fast-tier models tend to rubber-stamp whatever deal terms you feed in rather than stress-testing them — if that happens, explicitly ask the model for the three ways this specific deal could fail before accepting its recommendation.Business & StrategyA 40-person SaaS company's VP of Partnerships has a term sheet from a larger platform offering co-marketing and a revenue-share referral deal in exchange for exclusive placement in their marketplace for 12 months. The CEO wants a recommendation before Friday's board meeting, and the VP needs to separate genuine upside from the platform's sales pitch before presenting it.

The Partnership Due Diligence Brief: Turn a Proposed Channel Deal Into a Risk-Weighted Go/No-Go

A business-strategy prompt that turns a partnership or channel proposal into a risk-weighted go/no-go memo — with a required exit trigger — before signatures happen instead of after the exit costs pile up.

partnership-strategydue-diligence
The Show-Don't-Tell Converter: Turn Exposition-Heavy Prose Into Scenes Readers Feel Instead of Skim
Works best with models strong at literary nuance and willing to critique rather than just praise — Claude Opus 5.5 or Sonnet 5 handle the line-by-line diagnosis and tonal-consistency check well. GPT-5.4 is a solid alternative. Lighter or fast-tier models tend to just add adjectives to the original sentence instead of actually restructuring it into a scene — if that happens, explicitly ask the model to rewrite using action, dialogue, or sensory detail instead of description.creativeA novelist revising chapter 3 of a thriller has a paragraph where the narrator states "Sarah was terrified" and "she knew she shouldn't have come here" right before a confrontation scene. Beta readers keep saying the scene "doesn't land," but can't say why. The writer pastes the full paragraph into this prompt along with the intended effect — dread building toward the door opening — and gets back exactly which lines are stating feelings instead of dramatizing them, plus two rewrite options for each, before sending the chapter back out for another read.

The Show-Don't-Tell Converter: Turn Exposition-Heavy Prose Into Scenes Readers Feel Instead of Skim

An editing prompt that flags every "telling" sentence in your draft, rewrites each one at two intensity levels, and names the recurring habits causing it — before an agent or beta reader stops reading.

creative-writing-promptfiction-editing
The Follow-Up Enforcer: Turn a Meeting's Action Items Into Deadlines Nobody Can Quietly Ignore
Claude Sonnet 4.6 -- handles the tone shift across all three escalation tiers and the conditional 'flag, don't invent' logic for undated items reliably. GPT-5.4 and Gemini 3 Pro also work well; smaller or budget models tend to collapse all three tiers into the same tone, or quietly assign a deadline to a flagged item instead of surfacing it as an open question.Writing & CommunicationA product team just wrapped a 45-minute sprint planning meeting with eight action items split across three owners and zero written follow-up system. The project lead has two other launches to run this week and can't personally remember to chase each person at the right moment.

The Follow-Up Enforcer: Turn a Meeting's Action Items Into Deadlines Nobody Can Quietly Ignore

A three-tier escalation prompt that turns raw meeting notes into owner-specific deadlines and the exact Slack and email messages needed to chase them -- without turning every reminder into a manager-cc'd confrontation.

workplace communicationmeetings
The Three-Persona Study Panel: Learn a Hard Concept by Watching an Expert, a Skeptic, and a Beginner Argue About It
Works best with models that hold distinct personas consistently across a long exchange — Claude Opus 5.5 or Sonnet 5, or GPT-5.4. Gemini 3 Pro also handles the three-way back-and-forth well. Fast-tier/lightweight models tend to let the Skeptic and Beginner blur into the same voice after a few exchanges — if that happens, ask the model to restate which persona is speaking at the start of each line.learningYou're a product manager who just inherited a project built on event sourcing, and you have a design review with senior engineers in three days. You've read two blog posts and watched a conference talk, but you don't trust your own understanding enough to ask a good question without risking sounding like you don't belong in the room. You run this prompt the night before, feeding it your rough mental model and the specific part that still confuses you — why you'd ever need to replay history instead of just storing current state — and get back a debate that pressure-tests your understanding before anyone else does.

The Three-Persona Study Panel: Learn a Hard Concept by Watching an Expert, a Skeptic, and a Beginner Argue About It

A role-play prompt that runs a live three-way debate — an Expert, a Skeptic, and a Beginner who thinks exactly like you — to expose the exact moment your understanding of a new topic would break.

roleplay-promptlearning-prompt
The Context-Switch Cost Calculator: Turn Tomorrow's Task List Into a Batching Plan That Saves You an Hour
Works best with models strong at structured multi-step reasoning — Claude Sonnet 5 or Opus 5.5, or GPT-5.4. Gemini 3 Pro handles it too, but fast-tier/lightweight models sometimes skip straight to the final schedule without showing the Step 2 switching-cost table — if that happens, explicitly ask the model to show its work before the reordered schedule.productivityYou're a product manager with twelve items on tomorrow's list, jotted down overnight in whatever order they occurred to you: three pull requests to review, a quarterly roadmap deck to finish, two Slack threads that turned into mini-fires, a 1:1 with a direct report, a budget spreadsheet due Friday, an investor update email, a design review, a bug triage call, and a performance review draft. Before opening your calendar, you run this prompt to find out whether your gut-feel order is secretly bleeding an hour to refocusing, and get back a batched version you can start cold at 9am.

The Context-Switch Cost Calculator: Turn Tomorrow's Task List Into a Batching Plan That Saves You an Hour

A four-step prompt that scores the hidden refocusing cost in your task order, then rebuilds your day around your fixed meetings so similar work stays together.

productivity-promptcontext-switching
The Rate Increase Messenger: Turn an Overdue Price Hike Into an Email Your Best Client Won't Panic Over
Claude Opus 4.7 or GPT-5.4 — both handle the tone calibration across the three email variants well. Works on Gemini 3 Pro too, though it sometimes needs a reminder to keep the ‘Direct & Brief’ variant genuinely brief rather than drifting toward the relationship-first tone.finance-negotiationYou're an independent consultant or freelancer who hasn't raised your rate with your longest-running client in two or more years, even though your rate for new clients has climbed. A contract renewal or new project scope is coming up in the next few weeks, and you need to send the increase this week — but every draft you've written either buries the number in apologetic language or sounds like an ultimatum.

The Rate Increase Messenger: Turn an Overdue Price Hike Into an Email Your Best Client Won't Panic Over

A structured prompt that turns an overdue freelance rate increase into three ready-to-send email drafts — pitched by tone, not apology — so the ask lands the first time.

rate-increasefreelance-pricing
The Metric Gaming Detector: Find Out If Your KPI Improved or Just Got Easier to Hit
GPT-5.4 or Claude Sonnet 5.5 — both handle structured, skeptical multi-hypothesis reasoning over messy, qualitative inputs well. Works on Gemini 3 Pro too, though it occasionally needs a follow-up nudge to keep the ‘assume suspicious by default’ framing from softening into a generic validation.Data AnalysisYour support team's average ticket resolution time dropped from 18 hours to 6 hours over one quarter, and leadership wants to roll the same resolution-time bonus out to two more teams before anyone has confirmed the number reflects real improvement rather than a reporting change.

The Metric Gaming Detector: Find Out If Your KPI Improved or Just Got Easier to Hit

A structured audit prompt that separates genuine performance gains from measurement artifacts, incentive gaming, and selection effects — before you scale a “win” to the rest of the company.

data analysiskpi
The Resume Gap Analyzer: Turn a Job Description Into a List of What's Missing From Your Resume
GPT-5.4 or Claude Sonnet 5.5 (both handle structured line-by-line comparison well; Claude tends to be more conservative about not inventing experience)CareerA mid-level product manager found a job posting closing in 48 hours and needs to know exactly which resume bullets to rewrite and which requirements are genuine gaps, instead of guessing what to emphasize under deadline pressure.

The Resume Gap Analyzer: Turn a Job Description Into a List of What's Missing From Your Resume

A prompt that compares your resume against a job posting line by line, separating real experience gaps from experience you have but never wrote down.

resumejob-search
The Headline Hook Diagnostician: Turn a Flat Headline Into 20 Psychologically-Labeled Variants
GPT-5.4 (works equally well with Claude Sonnet 5.5 or Gemini 3 Pro — any strong general-purpose model handles this reliably)marketingA content marketer has a landing page headline converting at 1.2% and 20 minutes before a campaign launch deadline to generate real A/B test candidates instead of guessing from instinct.

The Headline Hook Diagnostician: Turn a Flat Headline Into 20 Psychologically-Labeled Variants

A prompt that generates 20 headline alternatives for the same offer, each tagged with its psychological trigger, so you test proof instead of guesses.

copywritingheadlines