Prompt Library

Awesome ChatGPT Prompts (Beyond Basic)

23 copy-paste prompts

23 ChatGPT prompts that go beyond basic. Role-based simulations, multi-step workflows, expert perspectives, and creative use cases that show what ChatGPT can actually do when prompted well.

In short: This page contains 23 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.

Louis Corneloup
By Louis Corneloup ยท Founder, Techpresso
Last updated ยทHand-curated & tested by the AI Academy team

Expert Simulations

4 prompts

Three Experts Disagree

1/23

โœจ What it does

ChatGPT runs a three-way debate on [question] among [expert A perspective], [expert B perspective], and [expert C perspective], each speaking in turn with real disagreement. Read all three, then decide which argument you actually buy.

My question: [question]. Simulate a discussion between three experts who would disagree about this: [expert A perspective], [expert B perspective], [expert C perspective]. Each speaks in turn. Show genuine disagreement.

๐Ÿ’ก

Pro tip: Forcing disagreement surfaces nuance that single-expert response would miss. Strong learning prompt.

Be My Negotiation Coach

2/23

โœจ What it does

ChatGPT coaches you through negotiating [scenario], including their likely position, tactics, counters, and BATNAs. Rehearse the counters, then walk into the talk with one fallback you can live with.

I'm about to negotiate [scenario]. Be my negotiation coach. Walk through their likely position. Predict their tactics. Help me prepare counter-positions and BATNAs.

๐Ÿ’ก

Pro tip: Pre-negotiation prep with role-played opposition = dramatically stronger negotiation.

Be My Therapist (with limits)

3/23

โœจ What it does

ChatGPT plays a thoughtful listener for [situation], asking clarifying questions instead of dumping advice, and it is not a substitute for real therapy. Use it to name what you feel, then talk to a licensed therapist if the weight stays.

I'm processing [situation]. Be a thoughtful therapist. Ask clarifying questions before responding. Don't advice-bomb me. Help me think through what I'm feeling. (Note: not a replacement for actual therapy.)

๐Ÿ’ก

Pro tip: Useful for processing โ€” not replacement for therapy. The "ask clarifying questions" is critical.

Be My Hostile Reviewer

4/23

โœจ What it does

ChatGPT plays the harshest peer reviewer of [project/proposal/argument] from [paste], finding every weakness and steeling the rejection. Fix the holes you agree with, then ignore the cheap shots.

I've drafted [project/proposal/argument]: [paste]. Be the harshest peer reviewer. Find every weakness. Steelman the rejection. Don't spare my feelings.

๐Ÿ’ก

Pro tip: Pre-submission hostile review = catches what your own review misses.

Prompts get you started. Tutorials level you up.

A growing library of 300+ hands-on AI tutorials. New tutorials added every week.

Start 7-Day Free Trial

Multi-Step Workflows

4 prompts

Research and Synthesize Around This Question

5/23

โœจ What it does

ChatGPT researches [question] in four steps: sources, perspectives, a synthesized answer, and what is still uncertain. Follow the sources on your own, then treat the synthesis as a draft, not a citation.

My question: [question]. Walk me through: 1) What sources would I consult? 2) What perspectives exist? 3) Synthesize the strongest answer. 4) What's still uncertain? Show your reasoning at each step.

๐Ÿ’ก

Pro tip: Multi-step research framing > one-shot answer. Builds understanding of how research works.

Build a Plan in Three Phases

6/23

โœจ What it does

ChatGPT builds a three-phase plan for [goal]: foundation, build, and refinement, with actions, success measures, and common failures. Do the foundation actions first, then start your build phase only when those measures hold.

My goal: [goal]. Build me a phased plan: 1) Foundation phase (what to do first), 2) Build phase (core execution), 3) Refinement phase (optimization). Each phase: specific actions, success metrics, common failures.

๐Ÿ’ก

Pro tip: Phased plans handle complexity better than flat task lists. Each phase has its own logic.

Diagnostic Then Prescription

7/23

โœจ What it does

ChatGPT first asks five diagnostic questions about [problem], then offers targeted solutions only after you answer. Answer honestly, then take the prescriptions that match what you actually said.

My situation: [problem]. First, ask me 5 diagnostic questions to understand the actual problem before suggesting solutions. Then, based on my answers, propose targeted solutions.

๐Ÿ’ก

Pro tip: Diagnostic-first prevents prescribing solutions to wrong problems. Strong workflow pattern.

Compare-and-Recommend

8/23

โœจ What it does

ChatGPT compares [A] and [B] on the criteria that should matter, scores each, recommends one, and names what would flip the call. Add your own constraints, then pick the option those conditions still support.

I'm choosing between [A] and [B]. Walk through: 1) What are the criteria that should matter? 2) How does each option score on each criterion? 3) Recommend with reasoning. 4) What conditions would change the recommendation?

๐Ÿ’ก

Pro tip: The "what would change the recommendation" piece prevents brittle decisions.

Role-Based Power Prompts

4 prompts

You Are an Editor โ€” Be Tough

9/23

โœจ What it does

ChatGPT acts as a tough magazine editor on [paste] and says whether it would publish, plus what must change. Make those edits, then take the piece to a real editor if you still want it out.

You are a tough magazine editor. I'm submitting this for publication: [paste]. Tell me: would you publish this? What needs to change? Be honest in the way actual editors are.

๐Ÿ’ก

Pro tip: Editorial role = honest critique. Without role, ChatGPT defaults to encouraging.

You Are a Curious Student

10/23

โœจ What it does

ChatGPT plays a curious student on [topic] and asks ten probing questions that refuse surface answers. Answer each one in your own words, then notice where you still cannot explain it.

You are a curious student trying to deeply understand [topic]. Ask me 10 increasingly probing questions. Don't accept surface answers. Force me to explain things better.

๐Ÿ’ก

Pro tip: Teaching forces deep understanding. ChatGPT-as-student is a real learning technique.

You Are a Pre-Mortem Specialist

11/23

โœจ What it does

ChatGPT runs a pre-mortem on [describe], imagining the plan failed in six months and listing the likely failure modes. Close the gaps you missed, then update the plan before you start.

You are a pre-mortem specialist. My plan: [describe]. Imagine it's 6 months later and the plan failed. Walk me through the failure modes that are most likely. What did I miss?

๐Ÿ’ก

Pro tip: Specialized role for pre-mortems = better failure-mode surfacing than general "what could go wrong."

You Are a Beginner โ€” Spot the Jargon

12/23

โœจ What it does

ChatGPT reads [paste my draft] as a smart beginner and marks every line that would make them ask what a term means. Replace those phrases, then have a real beginner read your draft once more.

You are a smart but uninitiated beginner reading this: [paste my draft]. Mark every place where you'd need to ask "what does that mean?" Help me make this accessible.

๐Ÿ’ก

Pro tip: Asking ChatGPT to be the audience reveals jargon and assumed knowledge invisible to writer.

Like these prompts? There are full tutorials behind them.

Learn the workflows, not just the prompts. 300+ easy-to-follow tutorials inside AI Academy โ€” and growing every week.

Try AI Academy Free

Creative Power Prompts

4 prompts

Take This in 5 Different Directions

13/23

โœจ What it does

ChatGPT takes [paste idea] in five completely different directions, including ones that are not the safe choice. Steal the direction that surprises you, then develop that one.

Starting concept: [paste idea]. Take it in 5 completely different directions โ€” each treating the concept differently. Variety across directions. The most interesting may not be the safest.

๐Ÿ’ก

Pro tip: Five-direction generation surfaces possibilities you wouldn't reach. Pick the unexpected ones.

What's Missing from This Genre

14/23

โœจ What it does

ChatGPT names what is underrepresented in [genre] and writes five concept ideas that fill those gaps. Take one premise, then write a page from the angle your genre has been skipping.

Genre: [genre]. What's underrepresented in this genre? What stories haven't been told yet? What perspectives are missing? Generate 5 concept ideas that fill gaps.

๐Ÿ’ก

Pro tip: Filling genre gaps = real creative opportunity. ChatGPT is good at noticing absences.

Rewrite Famous Story in Different Voice

15/23

โœจ What it does

ChatGPT rewrites a famous opening from [story/novel] in the voice of [different author/style/genre] so you can see what voice does to story. Try the same move on a paragraph of yours, then keep the version that still sounds like you.

Take a famous opening passage from [story/novel]. Rewrite it in the voice of [different author/style/genre]. Show what voice does to story.

๐Ÿ’ก

Pro tip: Voice studies build voice consciousness. The rewriting reveals what voice carries.

Generate Premises from a Constraint

16/23

โœจ What it does

ChatGPT generates seven premise ideas that fit the constraint you name, so the limit produces stronger ideas than a blank page. Pick the oddest premise, then write a scene that obeys the same limit.

Constraint: [specific constraint โ€” e.g., a story set entirely in one room]. Generate 7 premise ideas that fit this constraint. The constraint should produce stronger ideas than open-ended generation.

๐Ÿ’ก

Pro tip: Constraints produce more interesting work than freedom. Build the muscle of working within constraints.

Self-Improvement Prompts

4 prompts

Audit My Communication Style

17/23

โœจ What it does

ChatGPT audits three to five samples of how you communicate and names patterns that work plus ones that may put people off. Read the off-putting list, then change one habit in your next message.

Here are 3-5 examples of how I communicate (paste samples). Audit my style. What patterns do you notice? What's working? What might be unconsciously off-putting?

๐Ÿ’ก

Pro tip: External audit catches patterns invisible to self. Strong self-improvement tool.

Help Me Articulate What I Want

18/23

โœจ What it does

ChatGPT asks probing questions from [describe vague wants] until you can name what you actually want. Answer each question, then write the precise want in one sentence you can act on.

I want something but can't articulate it. Here's rough material: [describe vague wants]. Ask me probing questions to help me articulate this more precisely.

๐Ÿ’ก

Pro tip: Often we know what we want but can't say it. Articulation through questioning = useful.

Identify the Real Question

19/23

โœจ What it does

ChatGPT tests whether [question] is the real question and surfaces the deeper one underneath. Sit with that deeper question, then work on it instead if it is closer to your truth.

I'm asking [question]. Help me notice if this is the actual question I should be asking. What might be the deeper question underneath?

๐Ÿ’ก

Pro tip: Surface questions often hide deeper questions. ChatGPT can help surface them.

Show Me My Blind Spots

20/23

โœจ What it does

ChatGPT finds likely blind spots in [topic] from [paste my thinking] and names perspectives you are underweighting. Seek one of those views on purpose, then update your notes.

Topic I'm working on: [topic]. Based on how I've been thinking about this: [paste my thinking], what blind spots am I likely to have? What perspectives am I underweighting?

๐Ÿ’ก

Pro tip: Blind spots are by definition invisible to self. External perspective surfaces them.

Go from copy-pasting to actually mastering AI.

AI Academy: 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ other tools. New tutorials added every week.

Start Your Free Trial

What Makes These Awesome

3 prompts

The Power of Roles

21/23

โœจ What it does

ChatGPT explains why assigning a role such as editor, therapist, beginner, or opposing expert unlocks behaviors a generic assistant prompt never gets. Rewrite your next prompt with a specific role, then compare the output.

Awesome prompts assign ChatGPT a specific role beyond "assistant" โ€” editor, therapist, beginner, opposing expert. Roles unlock specific behaviors that generic prompts don't access.

๐Ÿ’ก

Pro tip: Specific role > vague helpfulness. The role primes the response style.

The Power of Multi-Step

22/23

โœจ What it does

ChatGPT explains why walking through explicit steps produces stronger reasoning than asking for a one-shot answer. Split your next request into numbered steps, then use the stepwise reply.

Awesome prompts walk through multiple steps explicitly rather than asking for one-shot output. The step structure produces stronger reasoning than asking for the final answer alone.

๐Ÿ’ก

Pro tip: Make the reasoning visible. Multi-step prompts force ChatGPT to show work.

The Power of Friction

23/23

โœจ What it does

ChatGPT explains why adding friction, clarifying questions, diagnostics, or role-played opposition, beats a frictionless prompt that only gives surface output. Put one clarifying step in your next prompt, then answer it before you ask for the final draft.

Awesome prompts add friction (clarifying questions, diagnostic steps, role-played opposition) instead of removing it. Friction = better thinking. Removed friction = surface output.

๐Ÿ’ก

Pro tip: Most "easy" prompts produce easy answers. Adding intentional friction = better outcomes.

Free tool

Prompt Optimizer

Turn a rough idea into a structured, professional AI prompt.

Try it free โ†’

Frequently Asked Questions

Awesome prompts use roles, multi-step workflows, friction, and specificity. Basic prompts ask for things directly. The structural choices unlock different ChatGPT behaviors.
Yes โ€” the patterns transfer. Specific outputs vary by model, but role-assignment, multi-step workflows, and friction-adding work across LLMs.
Build a personal library you can reference. Memorize the patterns (role-based, multi-step, diagnostic-first) more than specific prompts. Patterns compound.
For simple lookup or quick tasks. "What's the capital of Spain?" doesn't need role-based prompting. Match prompt complexity to task complexity.
Save what works. Notice what fails. Build pattern recognition. Most "ChatGPT skill" is iteration discipline + library-building.

Is ChatGPT Plus worth paying for?

Read our honest, no-hype breakdown with the real pricing.

Read review โ†’

Prompts are the starting line. Tutorials are the finish.

A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.

7-day free trial. Cancel anytime.