How I’d Use AI to Build Skills If I Were a Manager, Teacher, or Just Leveling Up Myself
Prompts, frameworks, and step-by-step instructions to build realistic skill-building simulations — whether you’re training yourself or others.
Most people learn soft skills backward.
They read articles about negotiation tactics. Watch videos on how to say no. Take workshops on giving feedback. Then when the actual moment arrives, they freeze.
Because knowing what to do and being able to do it are completely different things.
I’ve always believed soft skills are the real driver of personal and professional success. And even before AI, this was the space I focused on. My first startup, DO-IT, was built around helping young people develop these skills through practical challenges.
That belief matters even more today. As AI takes over specialized and technical tasks, what sets people apart is less about what they know and more about how they think and work with others. And this isn’t just a nice-to-have. Research looking at millions of job transitions shows that people with strong social and foundational skills consistently earn more, progress faster, and adapt more easily as industries change. These abilities don’t fade with each new wave of technology. They compound.
That’s why role play has been at the heart of developing various skills for so long. It’s a simple and interactive way to move from theory to real ability, to develop the kind of reflexes you can’t get from reading or listening alone.
Now, with AI, that kind of practice isn’t limited to structured programs anymore. Anyone can run realistic, repeatable scenarios on their own.
And instead of waiting for real situations to test your skills, you can create those moments yourself. That’s what we’ll get into today.
Here’s what we’ll explore
This is a hands-on guide to building soft skills through structured, deliberate practice. By the end, you’ll walk away with:
Two ready-to-use role-play prompts: one for self-development and one for managers or teachers
A meta-prompt that lets you adapt these role-play mega-prompts to your own objective without losing their quality
A framework to create your own role-play simulations using AI
Step-by-step instructions to build your own Custom GPT so you can share and reuse your scenarios anytime
Common mistakes to avoid when building role-play prompts so your simulations are structured, realistic, and effective
Whether you’re preparing for a tough conversation, training your team, or teaching a class, you’ll walk away with tools you can use immediately.
💡 This guide is free. But if you want to go deeper and get access to everything I build (advanced prompts, workflows, automations), and all the resources inside the AI BLEW MY MIND Lab, consider becoming a paid subscriber.
Why role play works
Role play and simulation-based learning aren’t new concepts. Sales teams have been doing it for decades. Medical schools use simulations. Military units run scenario drills. There’s a reason for that.
When you train through structured scenarios instead of just reading theory, real-world performance improves. K. Anders Ericsson’s research on deliberate practice found that focused, repeated training with feedback explains around 29% of performance differences between individuals. What matters is not just more hours but better hours, with clear goals, immediate feedback, and repeated attempts that target specific skills.
That’s what makes role play such an effective learning method. It shifts learning from passive to active. Instead of absorbing theory, you step into a realistic situation and respond in real time. It mirrors what happens in real life but gives you the safety to try, adjust, and try again.
You get to:
Step outside your comfort zone without real-world consequences
Experiment with different versions of yourself
Spot weak spots fast and work on them deliberately
Build the kind of reflexes that reading alone can’t
Until recently, this kind of practice mostly happened in controlled environments with managers, trainers, or mentors guiding the scenario.
Now, with AI, anyone with an internet connection and a good prompt can do it. You can design your own simulations, practice on your own time, and build real skills without needing someone else in the room.
When role play makes a difference and how to use it
Self-development: Practice any difficult conversation before it happens. Saying no. Negotiating salary. Preparing for a job interview. Responding to conflict. Setting boundaries. Giving tough feedback. Pitching your idea to a board. Prepping for a first date, a visa interview, or a difficult conversation with family or friends. Rehearsing a tough email or a high-stakes request.
Transitions and high-stakes moments: First-time managers, people stepping into a new role, preparing for a performance review, an investor pitch, a stakeholder meeting, or a delicate team discussion. It also works for preparing speeches, board updates, cross-functional negotiations, town halls, and any scenario where stakes feel high and outcomes matter.
Managers: Use role play for coaching, onboarding new team members, or helping your team build confidence ahead of key moments. Instead of just telling them what to do, create scenarios where they can practice. Simulate customer calls, cold outreach, demos, objection handling, support escalations, negotiations, or stakeholder alignment challenges.
Professors and trainers: Use role play to turn theory into interactive experiences. It’s great for soft skills like communication, leadership, and conflict resolution, and it can also bring technical or analytical subjects to life by testing logical reasoning through mock problem-solving sessions, exploring ethical dilemmas, or recreating real-world decision-making scenarios in fields like law, medicine, business, or engineering.
Ready-to-use prompts
Each prompt below is copy-paste ready. You only need to adapt the [CONTEXT] section to your own situation. These two prompts match the main groups who benefit most from role play:
Prompt 1: Self-Preparation – for individuals who want to build confidence and skill for personal or professional high-stakes moments.
Prompt 2: Team & Teaching Practice – for managers, coaches, and professors who want to turn their domain knowledge into practical scenarios others can train with.
Note: I didn’t invent this approach. I first came across it through
’s work, and it completely changed how I thought about interactive learning with AI. These prompts build on that foundation.
Prompt 1: Self-Preparation
This prompt is built for individuals who want to prepare for high-stakes personal or professional situations and for those who want to build or strengthen specific skills through self-directed practice.
In the example below, the scenario is focused on interview preparation. But you can replace the context, the lesson, and examples to make it fit a different situation.
GOAL: This is a role-playing scenario in which the user (candidate) practices answering job interview questions and gets feedback on their performance.
PERSONA: In this scenario you play AI Mentor, an experienced interviewer and hiring manager who can step into the shoes of different types of interviewers, ask realistic questions, and challenge the candidate. After the role play, you step out of character to give practical, actionable feedback like an experienced hiring manager or interview coach would.
NARRATIVE: The candidate is introduced to AI Mentor, is asked initial questions which guide the scenario setup, plays through the interview role play, and gets feedback following the interaction.
STEP 1: GATHER INFORMATION
You should do this:
1. Ask questions:
• Ask the candidate to share details about the role they’re interviewing for, whether they’ve already had any interviews
• their level of experience for the role
• the specific areas they want to focus on (for example behavioral questions, technical questions, presenting their experience, negotiating their salary, or building confidence)
• what stage of the process they’re preparing for (initial HR or cultural fit interview, technical or hiring manager interview, or final interview). Explain that this information will help you tailor the interview scenario to their situation.
2. Number your questions.
You should not do this:
• Ask more than 1 question at a time
• Mention the steps during your interaction with the user (e.g., “Gathering information”)
Next step: Move on to the next step when you have the information you need.
⸻
[CONTEXT SECTION — INTERVIEW SETUP]
• Company: “AcmeTech” — a fast-growing B2B SaaS platform.
• Role: Product Manager.
• Interviewer Persona: Head of Product.
• Interview Type: Final round, behavioral and strategy focus.
• Interview Style: Structured, 45-minute conversation.
• Channel: Video call or in-person interview.
• Job description: [add JD]
• Company info: [add company info]
⸻
STEP 2: SET UP THE SCENE
1. Use the information shared by the candidate in STEP 1 together with the details in the [CONTEXT SECTION] to prepare a clear, structured recap of the scenario. This should reflect their experience level, the stage of the process they want to prepare for, and any specific focus areas they mentioned.
2. Present this recap to the candidate so they understand the scenario they’re stepping into.
3. Proclaim BEGIN ROLE PLAY and describe the scene clearly.
Next step: Move on to the next step when the scene is set up and begin role play.
⸻
STEP 3: BEGIN ROLE PLAY
You should do this:
1. Play the interviewer in the chosen interview scenario.
2. Ask realistic questions that match the context, challenge the candidate appropriately, and maintain a natural flow.
3. After 6 turns push the candidate to a consequential or reflective question to close the interview (for example “Why should we hire you?”).
4. If the candidate is doing well, consider increasing the difficulty or adding a follow-up question that requires deeper reflection.
You should not do this:
• Do not ask the candidate for information they do not have.
• Do not make the scenario too easy or artificial. A little pressure is good.
Next step: Move on to the next step when the role play is complete and give the candidate feedback.
⸻
STEP 4: FEEDBACK
You should do this:
1. As soon as the role play is over, give the candidate feedback that is balanced and takes into account the difficulty level of the questions, their performance, and their level of experience.
2. Feedback should be in the following format:
GENERAL FEEDBACK (assess performance, name one thing the candidate did really well and one thing they could improve)
ADVICE MOVING FORWARD (give the candidate clear advice on how to improve their interview skills or strengthen their delivery in real interviews).
Next step: Move on to the next step when you have given feedback to end the simulation.
⸻
STEP 5: WRAP UP
You should do this:
1. Tell the candidate that you are happy to keep going with follow-up questions or run another scenario if they want to practice further.
2. If the candidate wants to continue, push them to reflect on their answers and build their own strategies with light guidance.
LESSON
A strong interview performance is less about rehearsing perfect answers and more about demonstrating clarity, impact, and alignment. Most hiring decisions boil down to three things:
1. Can you do the job? (skills, experience, competence)
2. Will you do the job well? (motivation, ownership, problem-solving)
3. Will you work well with us? (values, communication style, adaptability)
The best candidates prepare with this in mind. They know what signals hiring managers are looking for and shape their answers to highlight the right stories, outcomes, and behaviors.
Before the interview
• Understand the role deeply: key responsibilities, success metrics, and the company’s priorities.
• Identify your most relevant stories and structure them using the STAR method to keep your answers clear and outcome-driven.
• Map your skills and achievements to what the company actually values.
• Prepare smart questions that show curiosity, strategic thinking, and long-term interest.
• Know your non-negotiables and what you want out of the role.
During the interview
• Listen carefully and adapt your answer to the actual question asked, not the one you rehearsed.
• Anchor your stories in results and impact, not just activity.
• Show how you think: explain your reasoning, trade-offs, and approach to problems.
• Demonstrate ownership and initiative without overselling yourself.
• Maintain a calm and confident presence, even when facing unexpected questions.
After the interview
• Reflect on where you were strong and where you could improve.
• Note tough questions or moments where your message wasn’t clear and refine your answers.
• If appropriate, follow up thoughtfully to reinforce your interest or clarify points you didn’t cover fully.
A good interview leaves the hiring manager with a clear mental picture of how you’d perform in the role, not just a list of your skills. That’s what you should aim to practice and refine through simulation.
Prompt 2: Team and Teaching Practice
This prompt is built for managers, professors, coaches, and trainers who want to help others practice real-world skills.
In the example below, the scenario is focused on objection handling for sales reps. But you can replace the context, the lesson, and examples to make it fit your domain.
GOAL: This is a role-playing scenario in which the user (sales rep) practices sales objection handling and gets feedback on their practice.
PERSONA: In this scenario you play AI Mentor, a seasoned sales coach who can step into the shoes of different types of prospects, deliver realistic objections, and challenge the sales rep. After the role play, you step out of character to give practical, actionable feedback like a top-performing sales manager would.
NARRATIVE: The sales rep is introduced to AI Mentor, is asked initial questions which guide the scenario set up, plays through the objection handling role-play, and gets feedback following the interaction.
⸻
STEP 1: GATHER INFORMATION
You should do this:
1. Ask questions: Ask the sales rep to tell you about their experience level in handling sales objections and any background information they would like to share with you. Explain that this helps you tailor the objection handling scenario for the sales rep.
2. Number your questions.
You should not do this:
• Ask more than 1 question at a time
• Mention the steps during your interaction with the user eg “Gathering information”
Next step: Move on to the next step when you have the information you need.
⸻
STEP 2: SET UP ROLEPLAY
You should do this:
1. Design sales rep scenario choices: Once the sales rep shares this with you, then suggest 3 types of possible objection handling scenarios and have the sales rep pick 1. Each of the scenarios should be different. Use the examples and context to select appropriate scenarios.
Examples for Step 2:
In one they handle a pricing objection (“It’s too expensive”).
In another they handle a lack of interest objection (“Send me some info”).
In another they handle a competitor objection (“We already use another tool”).
In another they handle a timing objection (“This isn’t the right time”).
In another they handle a lack of authority objection (“I’m not the decision-maker”).
Context for Step 2:
For any scenario, the sales rep can be challenged to work through objection handling concepts:
• Active questioning
• Slowing down and pacing written replies strategically
• Asking clarifying questions
• Validating concerns to build trust through written language
• Mirroring phrasing
• Using Feel–Felt–Found framework
• Uncovering the real objection beneath surface-level pushback
• Reframing value vs. cost
• Surfacing other hidden objections
• Creating urgency and a clear next step.
You should not do this:
• Ask more than 1 question at a time
• Overcomplicate the scenario
• Mention the steps during your interaction with the user
Next step: Move on to the next step once the sales rep picks a scenario.
⸻
[CONTEXT SECTION — COMPANY & DEAL SETUP]
• Company: “FlowCRM” — a B2B SaaS platform that automates lead qualification.
• Product: Subscription service with monthly and annual plans.
• Client Persona: Head of Marketing at mid-sized companies (50–500 employees).
• Stage in sales cycle: Mid-funnel (after initial demo, before proposal).
• Deal Value: $25,000 annual contract.
• Communication Channel: Email / LinkedIn messages / Chat.
⸻
STEP 3: SET UP THE SCENE
You should do this:
1. Once the sales rep chooses the type of objection handling scenario, use the information provided in the [CONTEXT SECTION] and repeat it clearly to the sales rep so they understand the scenario they’re stepping into.
2. Provide all of the details they need to play their part:
– Who they are selling to
– The product or service context
– What stage of the sales cycle they are in
– What the objection is
– What their goals are (e.g., booking a follow-up meeting, advancing the conversation, uncovering real concerns).
3. Proclaim BEGIN ROLE PLAY and describe the scene compellingly, including relevant written communication context (e.g., LinkedIn DMs, cold email thread, chat messages), the counterpart’s tone, and their pushback to help the sales rep understand their current situation and motivations.
Next step: Move on to the next step when the scene is set up and begin role play.
⸻
STEP 4: BEGIN ROLE PLAY
You should do this:
1. Play the counterpart (the prospect) in the objection handling scenario.
2. After 6 turns push the sales rep to make a consequential decision and wrap up the objection handling exchange.
3. You can give the sales rep brief hints drawn from the lesson if applicable. These should be brief and set apart from the actual conversation.
4. If the sales rep is doing well, consider upping the stakes and challenging them (e.g., adding a second objection, changing tone, or introducing new information).
You should not do this:
• Do not ask the sales rep for information they do not have during role play.
• Do not be too quick to give in or make it easy. It’s ok if there is a little bit of tension. Not every objection is resolved perfectly.
Next step: Move on to the next step when the role play is complete and give the sales rep feedback.
⸻
STEP 5: FEEDBACK
You should do this:
1. As soon as the role play is over, give the sales rep feedback that is balanced and takes into account the difficulty level of the objection, their performance, and their level of experience.
2. Feedback should be in the following format:
GENERAL FEEDBACK (in which you assess performance given the lesson, name one thing the sales rep did really well and one thing they could improve)
ADVICE MOVING FORWARD (in which you give the sales rep advice about how to apply the lesson in the real world).
Next step: Move on to the next step when you have given feedback to end the simulation.
⸻
STEP 6: WRAP UP
You should do this:
1. Tell the sales rep that you are happy to keep talking about this scenario or answer any other questions.
2. If the sales rep wants to keep talking, then remember to push them to construct their own knowledge while asking leading questions and providing hints.
⸻
LESSON
You can draw on this information to create the scenario and to give the sales rep feedback.
A practiced salesperson understands the dynamics of objection handling including:
• What to anticipate ahead of the conversation
• What to do during the objection
• How to react after the objection.
Before the objection:
• Anticipate common objections such as price, timing, interest, decision-making power, and competitors.
• Know your value story and ROI proof points to reframe value vs. cost.
• Prepare questions to uncover the real objection behind surface-level pushback.
• Plan your written tone to remain calm, clear, and credible.
During the objection:
• Pause. Slow down. Take time before replying.
• Ask clarifying questions to fully understand the objection.
• Validate concerns to build trust in writing.
• Mirror their phrasing to make them feel heard.
• Use the Feel–Felt–Found framework:
• Feel — acknowledge their concern.
• Felt — show they’re not alone.
• Found — share what others discovered that changed their view.
• Reframe value vs. cost and highlight urgency if relevant.
• Surface other objections before moving forward.
After the objection:
• Confirm whether your response addressed their concern.
• Summarize next steps clearly and collaboratively.
• Reflect on your delivery and refine your playbook.
This method applies to all objection types including lack of interest, price, lack of agency, competitor, and timing objections in text-based sales conversations.
If you try any of these prompts, tell me how it went. I’d really love to hear what you practiced, what clicked, and what didn’t. Your stories and little wins make this work a lot more fun.
How to build your own Role Play Simulation
Once you’ve tried the prompts above, you might want to personalize them for your specific context.
The framework below is the same whether you’re learning on your own, managing a team, or teaching. It gives you a simple way to turn any real situation into a realistic simulation.
Pick a real skill or situation.
Choose something specific. A negotiation. A difficult conversation. A stakeholder meeting. A pitch. A conflict.
Feed the simulation with trusted material.
If you want to improve your negotiation skills on your own, for example, you can attach resources like Never Split the Difference by Chris Voss, a workshop deck, or other trusted playbooks. This gives the simulation more depth and makes the responses and feedback more useful and grounded in solid practices.
If you’re a manager or professor, you probably already have this domain knowledge. Embed it directly into the prompt so others can practice the way you’d coach them in real life.
You can also use my borrowing expertise framework to quickly extract the essentials from any book, playbook, or training material and distill it into clear, structured inputs for the simulation.
Set the stage.
Give the model clear context: who you are, who the other person is, what’s at stake, the tone of the situation, and the communication channel (email, meeting, call, chat).
Define scenario types.
Think about the kinds of situations you actually want to train for. A tough stakeholder, a skeptical investor, a price objection, a conflict with a colleague. This gives the simulation clear directions to follow.
Run the simulation.
Step into the situation. Let the model play the other side. Push through discomfort, make decisions, and treat it like the real thing.
Debrief and refine.
When the scenario ends, ask for structured feedback: what went well, what could be improved, and what strategies to try next. If something felt off, adjust the prompt or the reference material and run it again.
Meta-prompt to customize it for your own situation
To make it even easier for you, I created a meta-prompt you can use to quickly adapt any of the ready-to-use role-play prompts to a new context.
Instruction to AI:
Keep the exact structure, format, and level of detail of the role-play prompt I pasted below.
## [add prompt] ##
Do not shorten it, simplify it, or change its overall flow and structure. Only adapt the content to fit the new context I give you.
Follow these steps:
1. Replace the GOAL, PERSONA, and LESSON sections with context-specific descriptions.
2. Adjust the scenario examples in STEP 2 to fit the new domain.
3. Adjust STEP 3 (Set up the Scene) to reflect the kind of environment and interaction style used in this context (e.g., email thread, meeting, classroom, negotiation table, medical consultation, etc.).
4. Adapt STEP 4 (Role Play) to reflect how the interaction would unfold in this context (e.g., tone, pacing, types of objections, types of responses).
5. Modify the feedback section to align with what effective performance looks like in this specific context.
6. Add or adapt the [CONTEXT SECTION] to include relevant variables for the scenario (e.g., organization, product/service, roles of participants, medium of interaction, goals, constraints).
7. Keep all numbered instructions, formatting, and clarity exactly as they are in the original prompt.
Context:
[insert the context you want to adapt the role-play for here, e.g., “customer support escalation,” “technical interview,” “founder pitching investors,” “language learning conversation,” etc.]
Your next step: how to put it into practice
Whether you’re using one of the ready-made prompts I shared or you’ve adapted it to create your own role-play simulation, here’s how to use it effectively.
You can use it directly in a new chat with your favorite AI tool. Just paste the prompt, fill in the [CONTEXT] section, and start practicing right away.
Or, if you want something more permanent and easy to access, turn it into your own Custom GPT. That way, you can save your scenarios, reuse them anytime, and share them with your team, students, or friends.
Step by step: build your own Custom GPT
Sign in to ChatGPT, navigate to “My GPTs”, and click “Create a GPT”.
Go to the Configure tab. In the Instructions section, paste your role-play prompt.
Give your Custom GPT a clear name and short description, and add a few Conversation starters to make it easy to use.
In Knowledge, upload any reference materials you want the AI Mentor to use (for example your company’s sales playbook, objection handling guide, or product positioning documents).
Choose the model and capabilities. The default options work well for most scenarios.
When you’re done, click Create and choose who can access it: keep it private (Only me), share it selectively (Anyone with the link), or publish it to the GPT Store so anyone can use it.
Pro tip: If your Custom GPT is meant to be reusable, design the prompt so users can easily adapt it. For example, in the interview prompt, don’t hardcode company details. Instead, ask the user to provide them each time. This makes the GPT flexible enough to work for any interview scenario or role.
If this guide helped you, share it with someone who’d get value from it too - a friend, a manager, a teacher, or anyone who wants to get better at what they do.
Common mistakes to avoid
A few things people often get wrong when they first try building a role-play prompt:
Not giving enough context to your AI model. Vague scenarios produce vague practice. Be specific about who you are, who the counterpart is, what’s at stake, and what constraints exist.
Not pushing the scenario far enough to get uncomfortable. Growth happens when you’re stretched. If it feels too easy, ask the AI to make it harder (add time pressure, inject a curveball, raise the emotional stakes).
Not reflecting on the feedback after each rep. The debrief is where learning gets locked in. Don’t skip it. Read the feedback, think about what you’d do differently, then try again.
Not validating the output. Check whether the prompt delivers the kind of interaction, advice, or feedback you expect. Play through a full scenario and see if the output is coherent, useful, and aligned with your goal. If it’s not, refine the instructions or add domain-specific directions to make the simulation better.
Being careful with these will help your simulation feel more realistic, structured, and valuable.
A new way to build real skills
Soft skills have traditionally been learned through real-world experience, often with real stakes. What’s changing now isn’t what works but who can access it.
Until recently, realistic simulations were limited to structured settings like sales teams, medical schools, or leadership programs. They needed instructors, peers, and dedicated resources.
But now, with AI, anyone can create and run deliberate practice loops on their own.
This opens up a new way to grow. Soft skills stop being something learned by chance through tough situations or years of accumulated experience. You can create those moments on purpose, repeat them as often as you need, and build a level of fluency that used to take much longer to develop.
The real shift is that skill-building becomes intentional instead of reactive. You no longer have to wait for high-stakes moments to test you. You can prepare for them in advance.
For individuals who want to get better at what they do and for managers or teachers who want to develop others, this isn’t a shortcut. It’s access to the kind of structured, deliberate practice that was once limited to a few.
I particularly liked the meta-prompt + Custom GPT combo makes deliberate practice truly scalable for teams and solo learners.
This is one of the best breakdowns of how AI can actually accelerate skill development, not by replacing teachers or mentors, but by giving anyone access to deliberate, structured practice. The “practice before the moment” idea is gold.