47 battle-tested prompts and straight-talk tips for getting more out of ChatGPT, Claude, Gemini, and every other AI — whether you're just starting out or already live in the thing.
The prompts in this kit are strong on their own — but these six habits are what separate people who get gold from people who get generic. Learn these and you'll write better prompts than the ones in any pack, including this one.
Who you are, what you're really after, what you've tried. AI can't read your mind — the detail you leave out is the detail that would've made the answer fit.
"Act as a sharp editor… give me a bulleted list." Telling it who to be and what shape to return changes the output more than any clever wording.
The first answer is a draft. "Too generic — make it punchier," "shorter," "now the opposite view." The magic is in round two, not round one.
End with "ask me any questions that would make your answer better before you respond." This one line fixes most bad outputs before they happen.
Paste an example of what "good" looks like — your writing, a format you like, a sample you admire. AI matches patterns; give it one to match.
AI can be confidently wrong. For anything that matters — facts, numbers, quotes, legal or medical stuff — check it. It's a brilliant assistant, not a source of truth.
They're not all the same. Each has something it's genuinely best at — and one habit that unlocks it. Match the tool to the task and everything gets easier.
The all-rounder. Brainstorming, writing, images, voice — the deepest feature set and the one most people already have open.
Fill in its memory and custom instructions once — who you are, how you like answers — and stop re-explaining yourself every chat. Use Projects to keep related work together.
Living inside Google — Gmail, Docs, Drive, Search — plus standout image and video generation.
Reach for it when the task touches your Google stuff or needs fresh info from the web. Give it permission to pull from your Workspace and it'll draft off your actual documents.
Careful reasoning, long documents, code, and writing with nuance. The thoughtful partner when you want depth over speed.
Paste in the whole long document or messy draft and ask for analysis — it handles length well. Use Projects to give it standing context it remembers across chats.
Answering with sources. It's a search engine that writes — current facts with citations you can actually click.
Use it when you need up-to-date, checkable facts, not open-ended chat. Always follow the source links — the citations are the whole point.
Working inside Microsoft 365 — Word, Excel, Outlook, Windows. It acts on the files and email you already have.
If your work lives in Office, let it. Ask it to summarize a Word doc, build an Excel formula, or draft a reply right inside Outlook.
Real-time takes and what's happening right now on X. A looser, more casual, sometimes edgier voice.
Reach for it for live trends and current conversation. Expect personality over polish — and double-check anything factual.
Strong reasoning and coding for free. Open, capable, and easy on the wallet.
A solid no-cost option for thinking-heavy tasks. Check your workplace's data rules before using it for anything sensitive or confidential.
Power users don't marry one AI. Run the same prompt through two and compare — you'll spot blind spots, catch errors, and learn which tool wins which kind of task.
The AI world moves fast, so exact model names and features shift — but what each platform is fundamentally built for tends to hold. Landscape as of 2026; when in doubt, check the tool's own site.
Most planning asks "how will this work?" This asks "how will this fail?" — which surfaces the risks your optimism is hiding.
It's one year from now and [my decision/project] has failed badly. You're a sharp outside observer looking back. Walk backward from the failure and tell me the most likely reasons it went wrong — ranked from most to least probable. For each, name the earliest warning sign I'd have seen.
"Most likely: you launched to an audience that wasn't warmed up — the early sign was low replies to your pre-launch emails."
Forces the strongest version of each position — including the one you disagree with — then judges between them. The cure for arguing against a strawman.
I'm deciding whether to [decision]. Build the strongest possible case FOR it, then the strongest case AGAINST — argue each as if it's the position you genuinely hold. Then step back and tell me which case is actually weaker, and the exact point where it breaks down.
Two persuasive arguments, then: "The 'quit now' case is weaker — it assumes your income replaces itself in 3 months, and nothing supports that."
Pulls you out of the heat of right now by forcing three time horizons. Brilliant for decisions where your feelings are loud.
Help me think through [decision] across three timeframes. How will I likely feel about this in 10 minutes? In 10 months? In 10 years? Be honest about where my current emotion is distorting the picture.
"In 10 minutes: relief. In 10 months: the income gap. In 10 years: this won't register — so decide on the pattern, not this incident."
Hunts for the assumption you didn't know you were making — the one so baked in you can't see it.
Here's my reasoning for [decision/plan]: [explain it]. Don't critique the logic. Instead, find the assumption underneath it that I haven't questioned — the thing I'm treating as obviously true that might not be. What am I not seeing because I'm too close to it?
"You're assuming your audience buys on features. But your own testimonials talk about feelings. That's the crack."
Flips the question. Instead of "how do I succeed?" it asks "what would guarantee failure?" — then you just avoid those things.
I want to [goal]. Don't tell me how to succeed. Instead, list everything I could do that would reliably sabotage this — the surest ways to fail. Be specific and a little ruthless. Then I'll just avoid them.
"Surest ways to kill it: show up inconsistently, only appear to sell, make signing up feel like a commitment not a gift."
Turns a fuzzy "I can't decide" into a weighted scoring grid based on your criteria — so the decision becomes visible instead of a feeling in your gut.
I'm choosing between [options]. Help me build a decision matrix. First ask me what factors matter and how much each weighs. Then score each option and show the totals — and flag any place the "winning" number doesn't match my gut, because that gap is information.
A grid on your weights, plus: "Option A wins on points, but you hesitated on 'will I enjoy this daily' — that's worth more than the math."
Stops AI writing sounding like AI. Feed it a sample of you first, then it matches your tone instead of defaulting to corporate mush.
Here are three things I've written: [paste samples]. Study my voice — rhythm, word choice, how formal or casual I am. Now rewrite this in my voice: [paste the new text]. Match how I actually sound, not how AI thinks I should.
Writing that sounds like you on a good day — because it learned your actual patterns first.
Cuts the flab. Tightens your draft without gutting the meaning, and flags where you're being vague.
Edit this ruthlessly: [paste text]. Cut it by about 30% without losing meaning. Flag anything vague, any word doing no work, and any sentence I could say in half the space. Show me the tighter version, then a short note on what you cut and why.
A leaner draft that says the same thing in fewer words — and a list of the crutch phrases you lean on.
Fixes weak openings. Give it your first line and get stronger ones that actually make people keep reading.
Here's the opening of my [email / post / article]: [paste it]. It's not landing. Give me 5 stronger first lines that would make someone stop scrolling and keep reading — across different angles (curiosity, a bold claim, a question, a story, tension). Tell me which you'd bet on and why.
Five openings across different angles and a recommendation — so you pick, not agonize.
Generates a batch of subject lines (or titles, or headlines) across different emotional angles, so you can test instead of guess.
My [email / post / video] is about [topic] for [audience]. Give me 10 subject-line options across different angles — curiosity, urgency, clear benefit, a number, a question, a little contrarian. Keep them under [X] characters. Then tell me your top 3 and why.
Ten angles to choose or test from, plus a shortlist you can actually A/B.
Turns your messy brain-dump of bullets into a structured first draft you can shape — beating the blank page every time.
Here are my rough notes for a [email / doc / post]: [paste the messy bullets]. Turn this into a clear, well-structured first draft. Keep my points and my intent — don't invent facts — just give it shape, flow, and a logical order. Mark anything that needs a detail I didn't give you.
A real draft built from your own points — the hardest 60% done, ready for you to make it yours.
Rewrites the same message in several tones so you can pick the one that fits the moment — warm, firm, formal, or casual.
Rewrite this message in four tones so I can choose: warm and friendly, firm and direct, polished and formal, and light and casual. Here's the message: [paste it]. Keep the core point identical in each — just change the temperature.
The same message at four temperatures — so you match the tone to the reader, not your mood.
Takes your chaotic task dump and sorts it by impact vs. effort — so you start the week on what matters, not what's loudest.
Here's everything on my plate this week: [dump it all]. Sort it into a simple grid by impact and effort. Tell me the 2–3 high-impact, low-effort things to do first, what to schedule, what to delegate or drop, and what's just noise dressed up as urgent.
The 2–3 things that actually move the needle first — and permission to drop the noise.
Turns messy notes or a transcript into decisions, owners, and deadlines — so nothing important dies in your notebook.
Here are my notes / the transcript from a meeting: [paste]. Pull out: the decisions we made, the action items with an owner and a due date for each, any open questions we didn't resolve, and a two-line summary I can send the group. Flag anything that had no clear owner.
A clean list of who's doing what by when — plus a summary ready to paste into the follow-up.
Forces a real choice when everything feels essential. No "it depends" — a ranked answer you can act on.
Here's my list: [the tasks/goals]. If I could only do 3 of these this week and had to let the rest wait, which 3 — and why those over the others? Be decisive. Then tell me the one I'm probably avoiding because it's hard, not because it's unimportant.
Three things, ranked, with reasons — and a callout of the hard one you keep skipping.
Turns a vague task in your head into a clear handoff — context, the actual ask, and what "done well" looks like.
I need to hand off this task: [describe it roughly], to [who]. Help me write a brief that sets them up to succeed: the context and why it matters, exactly what I'm asking for, what "good" looks like, the deadline, and what to check with me versus decide themselves.
A brief clear enough that they can run with it — and you get back what you actually needed.
Breaks a big, fuzzy goal into phases, dependencies, and — most importantly — the first concrete step you can take today.
I want to [the goal/project] by [when]. Break it into clear phases with the key milestones in each. Show me what depends on what, where the risky or slow parts are, and the single first action I could take in the next hour to get moving.
Phases, dependencies, and one first action small enough to actually start now.
Figures out why you're avoiding something, then hands you a 10-minute starting move small enough to actually do.
I've been avoiding [the task] for a while. Ask me a few questions to figure out what's really behind the avoidance — is it unclear, boring, scary, too big, or something else? Once we find it, give me a 10-minute starting move designed for that specific reason.
The real reason named ("it's not lazy — it's unclear"), and a first step sized to beat it.
You explain a concept to the AI; it catches every place you're fuzzy or wrong. The fastest way to find the holes in what you "know."
I'm going to explain [concept] to you in my own words: [your explanation]. Act as an expert checking my understanding. Catch every place I'm vague, oversimplifying, or plain wrong. Then give me the one correction that would most improve my grasp of it.
The exact spots where your understanding is thinner than you thought — surfaced by explaining, not cramming.
Turns "I want to learn X" into a sequenced path that fits your time — so you're not just watching random videos.
I want to learn [skill], going from [current level] to [goal]. I have about [hours] a week. Build me a sequenced learning plan — what to learn in what order, a small project at each stage to make it stick, and how I'll know I'm ready to move on.
An ordered path with checkpoints and small projects — a route, not a pile.
Instead of handing you answers, it asks the questions that lead you there — so the understanding is actually yours.
Be my Socratic tutor for [topic]. Don't give me answers. Ask me one question at a time that leads me toward understanding it myself. If I'm stuck, give me a hint, not the answer. Keep going until I clearly get it.
Understanding you arrived at yourself — which is the kind that sticks.
Explains a hard thing using something you already understand — the shortcut your brain needs to make it click.
Explain [hard concept] to me using an analogy from something I already know well: [cooking / sports / driving / gardening / whatever]. Then give me a second analogy from a totally different area, so I can see the idea from two angles.
The idea mapped onto something you already get — the "ohhh" moment on demand.
Turns material into an active practice session with feedback — because testing yourself beats re-reading every time.
Here's what I'm studying: [paste material or name the topic]. Quiz me with [10] questions, one at a time, mixing easy and hard. After each answer, tell me if I'm right and why. At the end, show me my weak spots and what to review first.
Active recall with instant feedback and a map of exactly what you don't know yet.
Finds where buyers will hesitate on your product or price — before they hesitate with their wallets.
Here's my offer: [product, price, who it's for, the promise]. Play a skeptical version of my ideal customer. Where would you hesitate, doubt, or click away — and why? List the top objections in order, and for each, the one thing that would ease it.
The objections ranked, each with a fix — so you patch the leaks before launch.
Pulls themes, objections, and the exact words your customers use out of a pile of reviews or feedback — gold for your copy.
Here's a batch of customer [reviews / survey answers / support messages]: [paste]. Find the recurring themes, the biggest complaints, what they love most, and the exact phrases they use to describe the problem and the payoff. Pull out lines I could use word-for-word in my marketing.
The themes, the objections, and your customers' actual language handed back to use.
Writes a first message that's personal, specific, and human — the opposite of the copy-paste pitch everyone deletes.
I want to reach out to [who, and what you know about them] about [the ask]. Write a short, genuinely personal first message — specific to them, respectful of their time, and easy to say yes to. No fake flattery, no "hope this finds you well," no paragraph about me before I've earned it.
A message that sounds like one human noticing another — the kind people actually answer.
Analyzes a competitor to find the positioning gaps you could own — the spaces they're leaving wide open.
Here's a competitor: [name / describe them / paste their page]. And here's me: [my thing]. Break down how they position themselves — who they speak to, what they emphasize, what they ignore. Then find the gaps: the audiences, needs, or angles they're leaving open that I could own.
The open lane they're ignoring — where you can be first instead of second.
Gets you ready for a rate, raise, or deal conversation — your goal, their likely goal, your walk-away, and three ways to open.
I'm negotiating [the situation — a rate, raise, contract, price]. Here's the context: [details]. Help me prep: what I should aim for, what the other side likely wants, my walk-away point, the strongest case for my number, and three ways to open the conversation with different levels of assertiveness.
Your number, their angle, your walk-away, and three opening lines ready to go.
Lets you practice a tough talk before you have it — and role-plays the other person's likely pushback so nothing catches you cold.
I need to have a hard conversation with [who] about [what]. First help me plan how to open it and what I most need to say. Then role-play as them — react the way they realistically might, including pushback — so I can practice. Afterward, tell me where I got defensive or unclear.
A real rehearsal with realistic pushback — so the actual conversation is your second take, not your first.
Takes what you want to say angrily and rewrites it so it lands — without burning the bridge you'll need tomorrow.
Here's what I want to say, unfiltered: [the angry version]. Rewrite it so it gets the real point across firmly but professionally — no passive aggression, no groveling. Keep my backbone, lose the stuff I'd regret. Give me one version I could actually send.
The same firm point, minus the shrapnel — a message that wins the issue without losing the person.
Helps you say no — clearly and kindly — with three levels of firmness so you can match the situation.
I need to say no to [the request / person] but I'm bad at it — I over-explain or cave. Give me three versions: a gentle no, a firm no, and a hard no with no wiggle room. Keep all three kind but clear, and don't over-apologize for having a limit.
Three clean nos at different strengths — pick the one that fits and send it.
Real accountability without groveling or excuses — an apology that repairs instead of one that makes it about you.
I messed up: [what happened] with [who]. Help me apologize well — owning it clearly, no "sorry you felt that way," no pile of excuses, and no over-grovelling that makes them comfort me. Just genuine accountability and, if it fits, how I'll make it right.
A clean apology that owns the thing and offers repair — the kind that actually mends it.
Turns "this isn't working" into feedback that's honest, specific, and actually usable — respecting the person while telling the truth.
I need to give [who] feedback about [the issue]. Help me make it honest and specific without crushing them — anchored to what actually happened, focused on the behavior not the person, and ending with a clear ask for what to do differently. Skip the fake compliment sandwich.
Specific, kind, actionable feedback with a clear next step — heard, not resented.
Lists every hidden assumption baked into a plan or claim — so you can check the foundation before you build on it.
Here's a plan / claim: [paste it]. List every assumption it quietly depends on to be true — including the ones so obvious nobody states them. For each, tell me how risky it is if it turns out to be wrong, and how I could quickly check it.
Every load-bearing assumption listed and rated — so you test the risky ones first.
Not a generic summary — it tells you only what would actually change what you do. Signal, not noise.
Here's a long [document / report / thread]: [paste]. I need to decide [the decision]. Don't summarize everything — tell me only the parts that would actually change my decision, what they imply, and anything that contradicts itself. Skip the rest.
Only the parts that move your decision — the 10% that actually matters.
Attacks your work the way a tough critic would — so you find the weaknesses before someone else does.
Here's my [plan / draft / argument / pitch]: [paste]. Tear it apart as if you were a smart skeptic trying to poke holes and get it rejected. Give me the strongest objections, the weakest links, and the questions I'd struggle to answer — so I can fix them first.
The hardest objections in advance — with time to answer them.
Scores your options against the criteria you name — not the generic factors a review site would use.
Compare [option A] and [option B] — but only on the things I care about: [your criteria]. Score each on those, be honest about the tradeoffs, and tell me which one wins for someone with my specific priorities, not the average buyer.
A comparison scored on your priorities — a real answer for your situation.
A clear explanation that respects your intelligence without assuming you already know the jargon. No dumbing down, no gatekeeping.
Explain [topic] to me like I'm smart but completely new to it. Skip the jargon, or define it the first time you use it. Give me the core idea, why it matters, and one example — clear enough to actually get, without talking down to me.
The real concept, made clear — respected, not patronized.
Takes one concept and spins it into 20 variations across different angles — so you have options to react to instead of a blank page.
Here's one idea: [the concept]. Give me 20 variations on it across different angles — some safe, some weird, some bold, some tiny. Don't filter for quality yet; I want range. Then star the 3 you'd chase if it were you.
20 directions to react to — and reacting is far easier than inventing.
Generates names for your product, project, or brand — plus why each one works, so you can judge instead of just vibe.
I need a name for [what it is, who it's for, the feeling I want]. Give me 15 options across a few styles — plain and clear, playful, evocative, made-up. For each, one line on why it works and what it signals. Then flag any that might be hard to spell, say, or find online.
Names with reasoning attached — so you can pick on logic, not just gut.
A thoughtful end-of-week check-in that asks you the right questions based on what actually happened — journaling with a guide.
Here's how my week went: [a quick brain-dump]. Be a thoughtful reflection partner. Ask me 3–4 questions that would help me actually learn something from it — not generic prompts, but ones that fit what I just told you. Go one at a time.
Questions tuned to your actual week — reflection that goes somewhere.
Describe the person; get thoughtful, non-obvious gift ideas tuned to who they actually are — not the top-10 list everyone gives.
Help me find a gift for [who they are, what they're into, our relationship, the occasion, budget]. Skip the obvious stuff. Give me a range of thoughtful ideas — a few safe, a few creative, one splurge, one that's really personal to them — and a line on why each fits.
Ideas that fit the actual person — the "how did you know?" kind.
Turns a vague "I should plan that" into an actual itinerary — a trip, an event, a packed day — timed and realistic.
Help me plan [the trip / event / day]. Here's what I'm working with: [who, when, budget, must-dos, constraints]. Give me a realistic plan with timing, flag anything I'm forgetting, and keep it doable — build in breaks instead of cramming every minute.
A timed, realistic itinerary with the stuff you'd have forgotten flagged.
The most valuable prompt here, because it makes every other prompt better. Hand it a weak prompt; get a stronger one back, with the reasoning.
Here's a prompt I'm using: [paste your prompt]. Rewrite it to get a noticeably better result. Then tell me what you changed and why, so I learn to write better ones myself next time.
A stronger prompt and the lesson behind it — you level up, not just this one answer.
The one-line add-on that quietly fixes most bad outputs: make the AI gather what it needs before it answers.
I want help with [the task]. Before you answer, ask me any questions that would make your response sharper and more tailored to my situation. Once I've answered, then give me your best response.
A couple of sharp questions first — then an answer built on your real details.
The underlying structure of every great prompt. Learn to set who it is, what limits to respect, and how to shape the answer — and you can build your own for anything.
Act as [the role/expertise]. Your task: [what you want, with context]. Constraints: [length, tone, what to avoid, must-haves]. Format the answer as [list / table / short paragraphs / script].
A repeatable skeleton — fill the four blanks and you've got a strong prompt for anything.
When the AI nails something, don't lose it. Convert the great result into a reusable prompt you can run again with new inputs.
That last response was exactly what I wanted. Turn what just worked into a reusable prompt template I can use again for similar tasks — with clear [bracketed blanks] where I'd swap in new details next time. Then show me one example of it filled in.
A saved, reusable template — so the win becomes a tool, not a one-off.
No pack covers every situation. When you hit one these prompts don't, build your own with this. Fill the blanks and you'll get something useful almost every time.
Who should it be? "A sharp editor," "a skeptical investor," "a patient tutor." This sets the whole lens.
The specifics only you know. The more real detail, the less it has to guess.
The actual verb — analyze, rewrite, brainstorm, compare, plan. Be direct about what you want done.
The guardrails — length, tone, what to include, what to avoid. This is where "good" gets defined.
The shape of the answer — a list, a table, a script, three options. Ask and you'll get it.
Here's the Pre-Mortem filled in and the kind of answer it gives back — so you can feel the difference between a lazy ask and a sharp one.
It's one year from now and my paid newsletter launch has failed badly. You're a sharp outside observer looking back. Walk backward from the failure and tell me the most likely reasons — ranked — and the earliest warning sign I'd have seen for each.
1. You launched to a cold list. The most likely killer. Early sign: your free content wasn't getting replies or shares in the months before — engagement was quietly flat.
2. The paid tier wasn't clearly different from the free one. People couldn't articulate why to pay. Early sign: when you described the paid version, even you hedged.
3. You priced on what you needed, not what it was worth to them. Early sign: you picked the number in your head, not from talking to a single potential subscriber.
Watch the first one hardest — it's the one that sinks most launches, and the one you're least likely to admit is happening.
The setup is the tedious part. Do it once and every prompt after gets faster, sharper, and more you.
Use custom instructions or memory to tell your AI who you are, what you do, and how you like answers — concise, no fluff, your industry. Now every response starts tailored instead of generic.
Most tools have Projects or folders. Make one for work, one for a side hustle, one for personal — each keeps its own context so chats don't bleed together.
When a prompt nails it, save it (use #47). A note in your phone with your ten best beats hunting for the right words every time.
The people who get the most out of AI aren't the ones hoarding magic prompts. They're the ones who give it context, push back on the first answer, keep the good stuff, and always double-check what matters.
You've got the prompts. Now it's just reps. And remember — AI's the co-pilot; you're still the one flying. Use it as an extension of you, never a replacement for you.
Get the next drop → join the list
Part of Brained by Britt — smart stuff from an overactive brain.
A note: AI is a powerful assistant, not an infallible one. It can sound completely confident while being wrong, so verify anything that matters — facts, figures, quotes, and anything legal, medical, or financial — and don't paste in sensitive or confidential information you wouldn't want stored. Tools, models, and features change fast; details here reflect the landscape as of 2026.