A prompt is just the message you type to an AI chatbot. It’s the actual words you use when you ask it for something.
Here’s what most people never figure out. “Help me with my budget” and a prompt that gives your real income, forces a reality check, and asks for a specific format are both just a sentence or two. But they get you completely different answers. The gap between those two prompts is this whole page.
Every prompt below does more than ask a question. Each one uses a specific technique on purpose, most often a persona (telling the AI to act as a specific kind of expert before it answers, like “act as a hiring manager”).
Sometimes the technique is forcing the AI to ask you a question first, or making it show its reasoning step by step. Each prompt comes with a short note explaining why that technique is the right one for that situation.
Pew Research found that 64% of US teens have used an AI chatbot. A follow-up Pew survey found that only about a quarter of teens feel really confident using one.
That’s the gap this page is here to close. Most teens have never actually been taught any of this, because there isn’t really anywhere else that teaches it.
For prompts specific to homework and studying, see TeenLearner’s guide to AI prompts for school, organized subject by subject. This list covers everything outside the classroom.
One more thing before you start. Every prompt below has blanks in brackets, like [amount] or [job]. Those aren’t meant to be typed in literally. Swap each one out for your own real details before you send it.
The One Trick Most People Never Learn: Ask AI to Fix Your Own Prompt
Before you get to the categories, here’s the single most useful move in this whole article. Almost nobody uses it.
If a prompt isn’t working, don’t keep rewording it yourself. Paste it back in and ask the AI to fix it:
“Here’s the prompt I just used: [paste it]. The answer I got back was too [vague/generic/short/whatever it was]. Rewrite my prompt so it would actually get me what I wanted, then explain what you changed and why.”
This is called meta-prompting. It’s a real technique, used by people who prompt AI for a living, not a gimmick.
The AI is genuinely good at spotting why its own answer let you down, because it can see exactly where your instructions left room to be vague.
Two or three rounds of this usually gets you a prompt that’s sharper than anything on this page. More than three rounds and you’re usually just adding length, not quality. Stop there.
Money and Your First Paycheck
A generic budgeting prompt gets a generic budget. These prompts force the AI to ask questions, check its own math against reality, and show its reasoning instead of just handing you numbers.
Building a budget that accounts for what you actually spend:
“I earn [amount] every [week/month]. Ask me three questions about my spending habits before you build anything. Then give me a budget as a simple table by category, and flag any category that looks unrealistic for my income.”
Understanding a paycheck that came in lower than expected:
“Act as a payroll specialist reviewing my pay stub. I work in [state], my paycheck shows [gross amount], and I received [net amount]. Walk through every line that could explain the gap, including FICA, federal withholding, and state withholding. Tell me plainly which ones are mandatory and which ones I might be able to adjust on my W-4.”
This is the most common paycheck surprise for first-time workers. The number in your head is hours times wage, and the number in your account never matches it. A specific persona helps here because a generic answer tends to gloss over which deductions are unavoidable. A payroll specialist won’t.
Choosing between two saving methods instead of guessing:
“Compare the 50/30/20 rule against a simple three-jar system for someone earning [amount] a month with few fixed expenses. Tell me which one actually fits a teenager’s spending pattern and explain why.”
Setting a savings goal that survives contact with reality:
“I want [amount] saved by [date] for [goal]. Calculate exactly what I need to set aside each week. Then tell me honestly whether that number is realistic for typical teen income, not just whether the math works.”
Side Hustles and Making Extra Money
These prompts are built to catch a weak idea before you spend a whole weekend on it.
Finding the failure points before you commit:
“Act as a skeptical investor who has watched a hundred teen side hustles fail. I’m [age], and here’s my situation: I want to start [idea] for [specific customers], my relevant skills or experience are [list], and my budget is [amount, or “none”]. Give me the three most likely reasons this fails within two months, ranked by likelihood. Skip the generic risks every business has. I want the ones specific to my situation.”
The skeptical-investor persona matters more here than in most prompts on this page. AI’s default instinct is to encourage your idea. A persona built around doubt is one of the few reliable ways to push past that. But the persona only works if it has real details to be skeptical about. Without your actual skills, budget, and time, “the three reasons this fails” is just a generic list any business could get.
Getting a real plan instead of general advice:
“I’m [age] and live in [city or area]. Here’s what I’m actually good at or have done before: [your skills or experience]. I have [X] hours a week to work on this and no budget. Build me a plan to land one paying customer for [idea] within two weeks, starting from zero audience. Give me the exact first three actions, not broad strategy.”
A plan built without knowing who’s actually doing the work is just generic startup advice with your idea’s name pasted in. The three first actions for a kid who’s great on camera and the three first actions for a kid who’s great at spreadsheets should not look the same.
Checking your price against the market you’re actually in:
“I’m charging [amount] for [service]. My experience level: [how long you’ve done this, or any relevant background]. Here’s what others nearby charge: [list]. Tell me if I’m underpricing based on my experience level, and what that price signals to a customer before I even talk to them.”
Writing one description that works across three formats:
“Here’s what I actually do: [describe your business or idea in a sentence or two]. Write three versions of a one-sentence description of it: one for a text to a friend, one for a flyer, one for an Instagram bio. Keep the actual meaning identical across all three.”
Career Exploration
It’s completely normal to have no idea what you want to do yet. Almost everyone feels this way at some point, and career counselors will tell you the same thing.
These prompts aren’t built to hand you an answer. They’re built to help you figure out what you actually think.
Getting past the highlight reel of a career:
“I’m interested in [career]. Skip the exciting version. Tell me the three parts of this job that surprise people who only saw the highlight reel, and how those parts show up on a normal week.”
Getting AI to interview you instead of handing you a list:
“I don’t know what I want to do yet, and that’s fine. Instead of giving me career suggestions, ask me questions about what I actually enjoy doing, what I’m good at, and what I hate doing, one question at a time. Only after five or six questions, suggest two or three directions based on what I told you, and explain your reasoning for each.”
A generic “best careers for someone who likes X” list is the least useful thing AI can give you here. It skips the part where it actually learns anything about you. Forcing it to ask first fixes that.
Ranking skills by how long they’ll actually matter:
“Act as a career strategist who tracks how AI is changing [field]. I’m currently [your age or grade in school], and I want to work in this field within five to ten years. Rank the three skills I should build now by how likely they are to still matter once AI has changed the field further, and explain the ranking.”
Comparing two paths on the same three measures:
“Compare [career A] and [career B] using only three things: daily work, income ceiling, and how hard each one is to automate. Give me a short table, not paragraphs.”
Interview Practice
This is the category where a persona genuinely earns its place. Real hiring managers roleplay mock interviews with each other for practice, and that’s exactly what you’re asking AI to do here.
A mock interview only works if it feels like a real one, which means one question at a time and specific feedback, not a wall of generic tips.
Running a full interview with feedback at the end:
“Act as a hiring manager interviewing candidates for a [job] role at a [type of business]. Here’s my background: I’m [age], and this is the experience or activities I can actually talk about: [list, or “this would be my first job” if that’s true]. Ask me one question at a time and wait for my answer before asking the next one. After five questions, rank my answers from strongest to weakest and explain exactly what made the difference between the best and worst one.”
Without knowing anything about you, the hiring-manager persona can only ask generic questions and give generic feedback. A real hiring manager works from your actual background, even a thin one, so give it something real to work with.
Fixing the answer almost everyone struggles with:
“Here’s my answer to ‘tell me about yourself’: [paste answer]. Act as the hiring manager listening to this live. Tell me the exact sentence where you’d stop paying attention, not general feedback, and tell me what a stronger opening sentence would sound like instead.”
“Tell me about yourself” is the single most dreaded interview question for first-time job seekers, because it’s the most open-ended one. Getting feedback on the exact sentence that loses a listener is more useful than any general tip about confidence.
Preparing two versions of a hard answer:
“I’m worried about being asked [specific question] in an interview for [job]. Give me two honest answers, one direct and one softer, and tell me which fits an interview better.”
College Essays and Applications
The value here is entirely in what the AI does not write. Admissions readers can spot a generic, AI-smoothed essay, and using AI to write your essay for you is exactly the kind of shortcut TeenLearner’s guide on using AI for school without cheating warns against.
These prompts are built so the essay stays yours. AI works here as an editor and a sharp question-asker, never as the author.
Brainstorming before any writing starts:
“Before we start, ask me five questions about experiences I might not think to mention in a college essay. Do not suggest a topic until I’ve answered all five.”
Getting interviewed instead of handed a topic:
“Interview me about a specific challenge I’ve faced. Ask one follow-up question at a time based on my previous answer, the way a real interviewer would, not from a fixed list.”
Finding exactly where a draft goes generic:
“Act as an admissions reader who has read ten thousand essays this cycle and is tired of generic openings. Read this draft and point to the exact paragraph where my voice starts sounding like every other applicant. Explain specifically what makes it generic, not just that it is: [paste essay].”
A reader-fatigue persona gets a more honest, specific answer than asking a plain assistant, which tends to go easy on your own writing.
Practicing Hard Conversations
AI should never write the message you actually send, and it should never hand you a script to memorize. A conversation that sounds rehearsed is often exactly what makes a parent or manager suspicious in the first place. What actually helps is rehearsing the conversation itself and getting clear on what you want to say, in your own words, before it happens.
Rehearsing a request by role-playing the other person:
“Play the role of my [parent/manager/teacher]. Here’s what’s actually relevant about them or how they usually react: [context, like “gets defensive about money” or “always asks to see a plan first”]. Ask me the questions they would realistically ask if I brought up [specific request]. Stay in character until I say stop.”
Getting to a clear point without being handed the words:
“I need to talk to someone about [situation]. Don’t write anything for me. Ask me questions until I can state how I feel, why I feel that way, and what I actually want, in two sentences I’d be comfortable saying out loud.”
That structure, how you feel, why, and what you want, is the same one counselors actually recommend for a hard conversation. It works better as questions to answer than as a script to read.
Pressure-testing your reasoning before a big ask:
“Here’s my reasoning for asking for [specific thing]: [explain]. Poke holes in it the way someone who wants to say no would. Be critical, not encouraging.”
Making Better Decisions
The strongest decision prompts resist giving you an answer. That’s the point. Forcing the AI to reason out loud, step by step, instead of jumping straight to a verdict, produces better thinking than just asking it what to do.
Comparing two options with no recommendation attached:
“Compare [option A] and [option B] using only pros, cons, and the consequences one year from now. Here’s what actually matters most to me when I’m deciding things like this: [your priorities, like money, free time, or what a specific person thinks]. No recommendation. Format it as two columns.”
Getting better questions instead of a verdict:
“I’m deciding between [option A] and [option B]. Don’t answer for me. Walk through what each choice actually depends on, step by step, then ask me five questions that would reveal what I care about most here.”
Catching an impulsive decision before it happens:
“I’m about to decide [decision] in the next hour. Ask me the one question most likely to change my mind if I’m about to make a mistake.”
Learning Something New
Not everything worth understanding is for a grade. These prompts are built for curiosity, not homework.
Getting a real explanation with a built-in check:
“Explain [topic] to a [your age or grade] who knows nothing about it, the way a genuinely good teacher would: one clear analogy, no jargon. Then ask me a question that would reveal whether I actually understood it or just followed along.”
Skipping the basics you already know:
“I know the basics of [topic]. Skip that. Tell me the one thing about this that most people still get wrong even after they’ve learned the fundamentals.”
Building a five-step understanding of something complex:
“I want to understand [topic] well enough to explain it to a friend. Give me the five things I’d need to know, ordered by how much each one depends on the last.”
Creative Projects and Hobbies
AI works best here as a collaborator that fills a specific gap, not one that replaces the idea.
Getting unstuck at the exact point you’re stuck:
“I’m stuck on [project] at this exact point: [describe]. Give me three directions, ranked from safest to most unexpected. Not three random ideas.”
Getting feedback that names a real weakness:
“Here’s what I made: [description]. Tell me the single weakest part and exactly why it’s weak. If you can’t find a real weakness, say so instead of inventing one.”
Finding a project that exposes a gap you don’t know you have:
“I want to get better at [skill]. Give me one specific project I could finish this week that would expose a weakness I don’t already know about.”
Using AI Safely: Habits Worth Building
These habits take less than a minute each, and they matter more the more you use AI.
Building a fact-check into every claim you use:
“Is [specific claim] actually true? Give me a one-line answer, then tell me exactly how to verify it myself in under five minutes.”
Catching an AI that’s just agreeing with you:
“Here’s my thinking on [topic]: [explain]. Find the weakest part of my reasoning. If you agree with everything I said, treat that as a warning sign, not a compliment.”
Knowing what never to type into a chatbot:
“List the specific types of personal information I should never type into an AI chat, and explain what could actually happen if I did.”
A Pew Research survey found that 12% of teens have used a chatbot for emotional support or advice. Common Sense Media’s 2026 survey found an even bigger share have used some kind of AI companion for casual conversation.
That use is real, and it has real limits. Common Sense Media’s own safety testing has found that leading chatbots, including the ones covered on this site, regularly miss warning signs of real distress.
A trusted adult, a school counselor, or a mental health professional will always know your actual situation in a way a chatbot cannot.
Choosing the Right AI Tool
Every prompt above works in more than one tool, but some tools fit certain categories better than others.
TeenLearner’s ChatGPT guide for teens covers the widest range of use cases and is a solid default for the money, career, and decision-making prompts. The Claude guide for teens is worth using specifically for prompts that need honest, critical feedback instead of encouragement. That matters most in the interview practice, essay feedback, and side-hustle reality-check prompts above, where you want the skeptical persona to actually hold up.
The Perplexity guide for teens is the better choice any time a prompt depends on current, real information, like checking market prices for a side hustle. Gemini and Copilot are solid general-purpose options if one of them is already built into your phone, laptop, or school account. Gemini Notebook is worth knowing about if you ever want to turn a set of interview notes or research into a study tool of its own.
Your Starting Point
Pick two or three categories that apply to your life right now: a first paycheck, an upcoming interview, a decision you’ve been avoiding. Try one prompt exactly as written before you customize it.
If the answer falls flat, don’t start over. Use the meta-prompting trick above and ask the AI to fix your prompt for you.
The pattern underneath every prompt on this page is the same one. Give real context, choose the technique that actually fixes what usually goes wrong, and build in a check on the answer instead of accepting the first one. That pattern works on every AI tool, not just the one you happen to be using today.
Frequently Asked Questions
What makes an AI prompt actually good, not just basic?
A basic prompt asks a single open question and accepts whatever comes back. A strong prompt gives specific context, picks the right technique for the problem, and builds in a check on the answer instead of accepting the first one.
That might mean a persona when you need honest or expert-level feedback, a forced question when you need the AI to learn about you first, or step-by-step reasoning when you’re working through a decision. Prompts that ask AI to ask you questions before it answers consistently produce sharper results than prompts that just ask for information directly.
Is it safe for teens to use AI for personal advice?
AI can help you organize your thoughts, practice a conversation, or think through a decision. It is not a substitute for a trusted adult, a school counselor, or a mental health professional, especially for anything emotionally serious.
Pew Research and Common Sense Media have both found that a real share of teens use AI for some kind of emotional support, and Common Sense Media’s research has also raised specific concerns about how AI chatbots handle serious mental health situations. Use AI to think, not to replace the people in your life.
Should I use ChatGPT, Claude, or another AI tool for these prompts?
Most prompts on this page work in any major AI tool. ChatGPT is a strong general choice for the money, career, and decision-making prompts, and Claude tends to perform better on prompts that need honest, critical feedback, like the interview and creative feedback prompts.
Perplexity is the better pick when a prompt needs current, real information rather than reasoning alone. TeenLearner’s guides to ChatGPT, Claude, Perplexity, Gemini, and Copilot go deeper on which tool fits which task.
This article is for informational purposes only. AI tools and their capabilities change frequently. The prompts and guidance here reflect what works as of September 2026 and may need adjusting as these tools update. AI is not a substitute for advice from a trusted adult, teacher, or professional.
Want more practical guides on AI, money, and the skills that actually matter? Subscribe to the TeenLearner newsletter and get one lesson delivered to your inbox each week.
