March 5, 2025

ikayaniaamirshahzad@gmail.com

AI-powered practice – by Doan Winkel and Mamie Rheingold


“If I’m interested in interviewing single moms and I’m practicing those interviews with a 19-year-old dude, he can’t actually play that character,” says , who leads the entrepreneurship program at John Carroll University. For years, Doan relied on traditional role-playing exercises to teach customer interviewing — pairing students to simulate conversations with potential customers for their business ideas. The results? Unconvincing and ineffective. “Either you’re practicing with a friend, so it doesn’t feel real, or you’re paired with someone random, and it’s just super-awkward.”

Doan has replaced these clunky classroom exercises with custom GPTs designed to simulate authentic customer interactions. Students can practice interviewing personas tailored to their target audience, receive immediate feedback, and refine their approach — without the social discomfort. The impact: “When it actually comes time to do the real interview, they feel more confident in their skills,” Doan observes.

Take a closer look at how he designed and implemented a custom GPT with a feedback loop to learn with students.

Let’s take a tour of Doan’s Customer Interviewer GPT.

I created the custom GPT where basically I’m saying: you’re going to play this role, ask the students what kind of person they want you to play, and then conduct the role-play with the student. After, you’re going to offer them developmental feedback.

I find it works well because I give it explicit step-by-step, concise instructions.

Role: You are an experienced entrepreneurship mentor. You have been coaching college students in entrepreneurship for over 20 years, specializing in customer interviewing techniques from the Lean Startup methodology. Your goal is to help students master the skill of conducting effective customer interviews to validate business ideas.

Objective: Guide students through the process of practicing customer interviews, ensuring they learn to ask insightful questions that reveal deep customer insights. Provide detailed, developmental feedback on their performance.

Next, it’s prompting the student to give it the information it needs to set up the role it’s supposed to play.

Step-by-step Instructions:

Role Identification:
The student will type "HELLO", then you will begin by asking the student questions to identify the type of customer you will portray in this role-play scenario. Use the following questions to clarify your role:

What product or service is the student focusing on?

Who is the typical customer for this product/service?

What specific problem does this product/service aim to solve for the customer?

Confirm your role to the student based on their responses.

Initiating the Interview:

Once the role is established, let the student know they can start the interview.

Remain in character throughout the interview, responding as the identified customer.

I’m priming the Custom GPT for providing developmental feedback to the student by identifying the areas the student should cover in the interview.

Interview Focus Areas:
Ensure the student covers the following key areas during the interview:

Problem Identification:
Ask questions to understand the problems the customer is experiencing.

Problem Timing:
Inquire about the last time the customer experienced this problem.

Emotional Impact:
Explore how the customer feels when they encounter this problem.

Current Solutions:
Discuss what solutions the customer is currently using to address the problem.

Dissatisfaction with Current Solutions:
Identify what the customer dislikes about their current solutions.

Source of Current Solutions:
Find out where the customer heard about their current solutions.

The feedback is where the power is. If we don’t include this, then what’s the point?

Something I didn’t do in these that another professor was telling me about is, before it gives the developmental feedback, it can ask the student to reflect on their experience. So the student self-assesses and then it gives the feedback.

Post-Interview Feedback:

After the interview concludes, provide comprehensive, developmental feedback to the student. Focus on:

How well they understood and conveyed the customer's problems and emotions.

The effectiveness of their questions in uncovering useful insights.

Areas where they can improve their questioning technique or interview structure.

Specific examples of strong and weak points during the interview.

Encouragement and Improvement:

Encourage the student by acknowledging their efforts and progress.

Offer practical tips for improving future customer interviews, such as active listening, asking open-ended questions, and probing deeper into customer responses.

Try it yourself

  • Sharpen and iterate on your prompt.

    A lot of trial and error goes into creating these. I start by using Ruben Hassid’s Prompt Maker custom GPT, which essentially turns bad prompts into good prompts. I cheat because I don’t want to put the time and effort into writing a good prompt. I tweak it a lot, and then I’ll have some colleagues kick the tires on it and see what it looks like. I’m always iterating on it.

  • Give human feedback on the conversations students have.

    I’ve been learning a lot from about grading the chats. They don’t just send me a chat, but they have to annotate it and they have to go through and assess where they could have done better or what they don’t agree with.

    So they’re learning by having the experience, getting the feedback, and then having to go through and reflect on the experience they just had.

    Read more: Stop Grading Essays, Start Grading Chats by Mike Kentz

Share a chat in just a few clicks

GPTs are a way for anyone to create a tailored version of ChatGPT to be more helpful with a specific task—and then share that creation with others. It’s a powerful tool in teaching in learning as you can train a GPT to be an on the ready tutor or simulate learning experiences that go beyond what’s possible in the classroom.

Here’s our cheat sheet to create your own GPT.

is the Chair and Director of the Entrepreneurship Center at John Carroll University. He believes “as an employee of the university, it’s my job to help students get a job” and is rigorously focused on helping them build real-world skills. He’s built other GPTs, including Employment Interviewer, Sales Interviewer, Negotiate with Investors, and Solving Product Design Failures Role-Play. You can connect with him on LinkedIn and his Substack, Set The World On Fire.

Leave a comment

Subscribe to receive profiles from innovators in education like Doan in your inbox.



Source link

Leave a Comment