Ep 85 - The Simple, Four-Part Framework You Can Use To Get Better Results From ChatGPT

Summary

---

Podcast Summary: The ChatGPT Experiment - Ep 85

Title: The Simple, Four-Part Framework You Can Use To Get Better Results From ChatGPT
Host: Cary Weston
Release Date: September 9, 2025

---

Main Theme & Purpose

In episode 85 of The ChatGPT Experiment, Cary Weston tackles one of the most common frustrations faced by new and experienced ChatGPT users: not getting the results you want. Cary introduces his practical, four-part framework—a strategy he personally uses—to transform ChatGPT from a “guessing machine” into a reliable thinking partner. Designed for curious beginners and everyday professionals, this episode breaks down the art of prompting into manageable steps to help listeners get clearer, more actionable, and relevant responses from ChatGPT.

---

Key Discussion Points & Insights

1. The Core Problem with ChatGPT Usage

• Many users express that ChatGPT’s answers often feel generic, vague, or miss the mark.
• Cary shares feedback from listeners:

  “I keep trying different prompts that I find online, but ChatGPT still feels like it’s guessing at what I want. I waste more time fixing its responses than if I’d just written it myself.” (04:08)

• Core mistake: Treating ChatGPT like a search engine or taskmaster, rather than as a conversation partner that needs onboarding.

2. Introducing the Four-Part Framework

• Cary’s sticky-note formula transforms results from ChatGPT:
  1. What are we doing? — State the exact task clearly.
  2. Why are we doing it? — Provide context or the objective.
  3. What does success look like? — Define a successful result/output.
  4. Do you have any questions for me? — Invite clarifying questions for better collaboration.
• Quote:

  “This four-part framework is going to help you a lot... it’s something that I use all the time and it gives me tremendous value.” (01:03)

3. Framework Breakdown & Practical Examples

1. What are we doing?

• Avoid making ChatGPT guess; clearly state the specific task.
• Examples:
  • “I’m putting together a short overview of our company for a new team member.”
  • “I need help reviewing a customer transcript and pulling out key insights we could use in training.”
    (07:10)
• Quote:

  “When you don’t clearly state what you’re trying to do, just like a person, ChatGPT sometimes has to guess. And when it guesses, you know what you get? Generic, confusing, or completely off track irrelevant answers.” (06:31)

2. Why are we doing it?

• Context helps ChatGPT make better decisions about tone, focus, and format.
• Examples:
  • “I'm writing this company overview so I can reuse it in future marketing materials.”
  • “I’m reviewing this transcript because I want to improve how we train new members.”
    (08:07)
• Quote:

  “Context is what turns a generic response into a useful one.” (08:07)

3. What does success look like?

• Define the desired output type, style, or format.
• Examples:
  • Success is “a table that compares the last three years of data, organized by region.”
  • “A first draft response that sounds like it came from a real person, not a robot.”
    (09:58)
• Quote:

  “You stop getting technically correct answers and you start getting formats and outputs the way you want it.” (09:59)

4. Do you have any questions for me?

• This prompts ChatGPT to clarify gaps and avoid misunderstandings.
• “You catch the misunderstandings before they happen, and your first response becomes your final response.” (10:37)

4. Practical Application: The Before-and-After Scenario

• Scenario: Creating training materials for a customer support team struggling to answer common software questions.
• Typical Prompt:
  • “Help me create training materials for customer support questions and reporting for our software company.”
    Result: Generic, rambling answers, not tailored to the real problem.
• Four-Part Prompt:
  1. State task: “I need to create a training document that helps our customer support team handle the most common questions we get about our software’s reporting features.”
  2. Give context: “I’m creating this because our new team members are struggling... it’s creating longer response times and some frustrated customers. I want something they can reference quickly during calls.”
  3. Define success: “A one-page reference sheet with top 5 reporting questions, clear answers, troubleshooting tips, formatted for quick scanning.”
  4. Ask for questions: “Do you have any questions for me to help you do your best work?”
• Outcome: More targeted, actionable, and usable output with far less back-and-forth.
• Quote:

  “Those are the four parts right in place. What are we doing? Why are we doing it? What does success look like? And do you have any questions for me?” (13:13)

5. Final Advice & Encouragement

• Cary emphasizes framing your prompts as an ongoing conversation.
• Having the four-part framework as a mental checklist quickly becomes second nature.
• Quote:

  “It’s not about the perfect prompt, it’s really talking to it like you would talk to a person.” (15:18)

---

Notable Quotes & Memorable Moments

• “This isn’t a computer software. This is a conversation.” (05:14)
• “ChatGPT...is a new team member who needs proper onboarding.” (05:25)
• “I literally, by the way, have a post it note on my computer that has these four steps.” (13:42)
• “The more you become comfortable in giving it context and details, and quite frankly, asking it if it has any questions, the better off you’re going to be.” (15:18)

---

Important Segments & Timestamps

• Problem Introduction & Listener Frustration: [04:08 – 05:14]
• Unveiling the Four-Part Framework: [06:10 – 06:31]
• Detailed Step-By-Step Breakdown: [06:31 – 10:37]
• Applying the Framework - Worked Example: [11:52 – 13:13]
• Summary of Key Steps and Mindset: [13:42 – 15:18]

---

Resources & Next Steps

• Free workshop for marketing professionals and business owners on using ChatGPT as a marketing partner. Details and registration on chatgptExperiment.com (14:20)

---

Summary

This episode provides a simple yet powerful structure to dramatically improve the quality of your ChatGPT interactions—applicable to beginners and seasoned users alike. Cary’s conversational, practical approach helps demystify prompt engineering and turns it into a relationship-building tool for more productive AI conversations. The four-part framework (task, context, definition of success, and clarifying questions) is a handy, repeatable tool for anyone looking to boost the impact of their AI outcomes.

Takeaway:
Treat ChatGPT like a thoughtful new teammate instead of a magic box, and you’ll consistently get better results. Stay curious!

---

Transcript

[A]

Back to school is better. With family freedom from T Mobile, we'll pay off four phones up to $3200 and give you four free phones, all on America's largest 5G network. Visit your local T Mobile location or learn more@t mobile.com familyfreedom. Up to $800 per line via virtual prepaid card typically takes 15 days. Free phones via 24 monthly bill credits with finance agreement eg Apple iPhone 16128 gigabyte 820099 eligible trade in eg iPhone 11 Pro for well qualified credits end and balance due. If you pay off earlier, cancel contact T Mobile.

[B]

Ever notice how ads always pop up at the worst moments when the killer's identity is about to be revealed during that perfect meditation flow on Amazon Music, we believe in keeping you in the moment. That's why we've got millions of ad free podcast episodes so you can stay completely immersed in every story, every reveal, every breath. Download the Amazon Music app and start listening to your favorite podcasts. Ad free included with product.

[C]

Hey, it's Carrie. Hey. This week I'm going to help you solve one of the biggest frustrations, biggest problems, biggest feedback I get from listeners, and that's ChatGPT is not giving me what I want. I'm not getting the detail, the value, the clarity, the relevance that I want. Okay, so I've got a four part framework. It's the same one that's on the sticky note on my computer. It's something that I use all the time and it gives me tremendous value. So if you've ever used ChatGPT and you're thinking, geez, I'm just not getting what I want, this four part framework is going to be for you. I'm going to describe what the four parts are. I'm going to walk through a little bit more details of each. Then I'm going to give you an example, putting it all together. And then when I'm done, I have a free workshop for folks that are interested in using ChatGPT for marketing purposes. So if you are a business owner and you're looking to utilize ChatGPT to help you market better, if you're a marketing person, you're looking to use Chat GPT for the stuff you do, I've got a free workshop. I'll tell you about it at the end. So a short episode this week, so stay with me. If that interests you, then I'll see you on the other side of the music. Hey there. Welcome to the ChatGPT experiment podcast designed to Help you get more out of ChatGPT and understand how to get a nugget for your personal or professional needs. My name is Kerry Weston. I'm your host. Glad you're here. Episode 85, four part formula. And as I mentioned at the beginning, I've got a workshop that I want to tell you about. All right, so big news today, by the way. Literally, as I am going to record here. 20 minutes ago, my daughter comes home from her driver's test with a piece of paper in her hand. So she is now a licensed driver. My second in the house. My daddy taxi duties, my free Uber duties are dwindling before my eyes. Very proud, very exciting. Do you have any kids that have gotten a license recently? It's one of those bittersweet things, isn't it? You start worrying about all the sounds and sirens again, but you free up a little bit more of your time. Another chapter. Another chapter in our house. Hope you're well. Hey, listen. This episode was born because of multiple conversations and trainings that I have, group and individual, where people are saying, I'm using it. I'm listening, but I'm not getting what I want out of it. So this framework is going to help you a lot. By the way, before I jump in, chatgpt, experiment.com is the site for the show. If you haven't been there, you're going to find some resources, you're going to find some articles, you're going to find some guides. There's free guides on the homepage. You can reach out and contact me. I'd love to talk to you, answer questions. I got trainings of you. I had a few trainings this week, actually had somebody come back for the third training we've been building on. It's been awesome to see the progress that they make with the sessions that we have. One little request before we jump to our theme for this week is if you find any value in the show, I'd certainly appreciate if you would. Subscribing to the podcast and leaving a rating and review it helps others find the show and bring them into our community.

[D]

Okay.

[C]

All right, so let's get to using ChatGPT and the framework. And typically it goes like this. So you open ChatGPT and you type in what feels like a clear request and what comes back is technically an answer. But it's vague. It might be fluffy or it's just completely missing the point or feels a little sterile. And now you're stuck. You're rewording or you're re asking or you're just find too much to try and tweak. You just kind of give up and maybe do yourself. I had one podcast listener recently told me, I keep trying different prompts that I find online, but ChatGPT still feels like it's guessing at what I want. I waste more time fixing its responses than if I just written it myself. So if you feel this, this frustration is normal, you're not alone. Yeah, most people are making the same fundamental mistake and it's killing their productivity using tools like ChatGPT. And I want to go back to that line that a podcast listener gave with me. I keep trying different prompts that I find online.

[D]

Okay?

[C]

So the real problem is the magic is not finding the perfect prompt. This isn't a computer software. This is a conversation.

[D]

Okay?

[C]

And so instead of treating it like Google or a quick search or a taskmaster, we should be treating ChatGPT like a new team member who needs proper onboarding.

[D]

Right?

[C]

You've heard me say the amazing intern before. Well, this article is going to walk you through a simple four part framework that has helped me get better output. I know it'll help you to get more relevant, more helpful responses. And it works for everything, whether you're writing or planning or analyzing or problem solving. So I'm going to tell you exactly what the four parts are and then I'm going to walk through some additional details and then I'm going to put it all together for you in like, for instance. Okay, so the four part framework that changed everything for me is this. Part one is literally what are we doing?

[D]

Right.

[C]

I start by stating the task of the project clearly and directly. The second part is why are we doing it? I give it context. I help ChatGPT understand the bigger picture. Or goal number three, what does success look like? I want to describe what a great answer would include or what a great final product would look like.

[D]

Right.

[C]

Or what I'm hoping to walk away with. And then number four, do you have any questions for me? I ask ChatGPT to give me clarifying questions that it may have so it can get the information, help fill in the gaps, do its best work. So that's it. That's the four simple steps that has helped me turn ChatGPT basically from a guessing machine into a thinking partner. Okay, now let me dig into each one and then I'll put it together for you. So part one, what are we doing? We're going to stop making ChatGPT guess. Okay, so this is where you tell ChatGPT what the actual task is not what you hope it figures out, but what you are specifically working on. Okay, so most people skip this step or assume the task is obvious from the prompt. But when you don't clearly state what you're trying to do, just like a person, ChatGPT sometimes has to guess. And when it guesses, you know what you get? Generic, confusing, or completely off track Irrelevant answers. Here's what this first part sounds like. I'm putting together a short overview of our company for a new team member. Or I need help reviewing a customer transcript and pulling out key insights we could use in training. Or I'm drafting an email to reconnect with a former client and I want it to sound friendly and professional. Okay, so what changes when I do it this way is ChatGPT stops guessing and starts focusing. You've heard me say narrow the band. Perhaps before I'm narrowing the band. If it's focused, you start getting targeted responses. Okay, now that's not the whole prompt. That's part one.

[D]

Right.

[C]

So part two is called why are we doing It? Well, this part gives ChatGPT context, and context is what turns a generic response into a useful one. Okay, so when you tell it why you're doing the task, you help it understand your priorities, your goals, what you're thinking would be most helpful. Here's a few examples. Okay, I'm writing this company overview so I can reuse it in future marketing materials. I'm reviewing this transcript because I want to improve how we train new members. I'm sending this email to reconnect with a client I haven't spoken to in a while. What this changes is ChatGPT starts making smarter choices or decisions about tone or length or focus or format because it's starting to understand what matters most to you.

[D]

Okay.

[C]

It's a simple addition, but it does make a big difference. Okay, so part three is what does success look like? Getting the output you can actually use. So Once you've told ChatGPT what you're working on and why, the next step is is to define what a helpful result would look like. Not just a topic, but the actual form it should take and what you'll do with it. So are you looking for a bolded list or a first draft? A table, short summary. This is where you spell it out. Here's some examples. Success looks like a table that compares the last three years of data organized by region. I'd like a short list of clear takeaways I can copy into a presentation. I'm looking for A first draft response that sounds like it came from a real person, not a robot. Please give me the top three patterns you notice from this data along with one sentence insights for each. So what changes here is you stop getting technically correct answers and you start getting formats and outputs the way you want it.

[D]

Right.

[C]

This step can save you time in reformatting and rewriting. Okay, then, part four. Do you have any questions for me? I want to turn it into a partner. And this is the step that turns chatgpt from a basic assistant into a thinking partner. Once you've told it what you're doing, why you're doing it, and what success looks like, finish this by asking do you have any questions for me that would help you do your best work that can uncover missing context or unclear priorities or overlook details.

[D]

Right.

[C]

And you'll start getting follow up questions from ChatGPT to fill in the gaps.

[D]

Okay.

[C]

So you catch the misunderstandings before they happen and your first response becomes your final response. Right. Because you're working into this final one. So let's put this all together. I want to show you how the framework can be used. I want to use the following scenario. Let's say you're a team lead who needs to create training materials for your customer support team. They've been struggling with the most common questions about your software's support features, which is leading to longer call times and some frustrated customers. This is exactly the kind of task where most people get stuck with ChatGPT. It seems straightforward enough, but you need training materials. There are actually dozens of ways ChatGPT could interpret that request, but without the right setup, you're likely to get something generic and unusable. So let's see what happens when you approach this task the typical way versus using this four part. So a typical approach might be something like help me create training materials for customer support questions and reporting. Okay. For our software company, if that's it, you're going to start getting probably rambling answers. You might get something depending on how often you use ChatGPT, how much it knows you, knows your company, but it's going to start trying to give you output.

[D]

Okay.

[C]

And it's going to start giving you perhaps even some generic stuff.

[D]

Okay.

[C]

Might be technically correct, but it's not going to be where you want to go because the format is wrong.

[D]

Right.

[C]

And you end up spending too much time going back and forth. Here's how I would take that same request. Help me create training materials for our software company for customer support questions about reporting and turn it into the four part series. Okay, yes, it's longer. Yes, there's more details, but that's the point. Just like I would do with an intern, just like I would do with a teammate, I'm giving it context and some details. Okay, so here's how I would say this is the. After using the four part, I need to create a training document that helps our customer support team handle the most common questions we get about our software's reporting features. That's step one. Then I add step two. I'm creating this because our new team members are struggling with these questions and it's creating longer response times and some frustrated customers. I want something they can reference quickly during calls.

[D]

Okay.

[C]

Step three. What does success look like? Success looks like a one page reference sheet with top five reporting questions, clear answers for each one, and maybe a troubleshooting tip. It should be formatted so someone can scan it quickly while they're on a call with a customer. Do you have any questions for me to help you do your best work?

[D]

Okay.

[C]

Do you see the difference? It's longer, thorough. Those are the four parts right in place. What are we doing? Why are we doing it? What does success look like? And do you have any questions for me?

[D]

Okay.

[C]

That four part framework is going to help you get more out of ChatGPT regardless. And you're going to do this a few times and it's going to become common sense second nature. But it's something that people typically don't start with when they grab and open ChatGPT.

[D]

Right.

[C]

Because we're using those. We're used to those one sentence searches in Google and whatnot. Right. We're not used to talking to ChatGPT, so this is going to help you frame it. And I literally, by the way, have a post it note on my computer that has these four steps. What are we doing? Why are we doing it? What does success look like? And do you have any questions? So when I get to the point of putting things in, I want to remember those and intuitively, you know, you can just start to pull this out in your vernacular as you're talking to it. Do I say step one, step two, step three? I do not.

[D]

Right.

[C]

I'm just using it as a guide.

[D]

Okay.

[C]

I'm using as a guide. So I hope that's going to be helpful. And I mentioned at the beginning, switching topics here, that if you are a business owner that is interested in marketing your business or a marketing person in a business, I've got a free workshop online webinar right. Interactive session I'm going to do on September 17th. You'll find details on the homepage of chat GPT experiment.com I'm going to show you how to model and set up a project inside ChatGPT that can become your marketing strategy partner. Okay, so I'm going to show you how to build the instructions. I'm going to show you what files we're going to build, what documents we're going to add, and how we're going to use it all together so that you can have a thinking marketing project in ChatGPT that you can come back to over and over again. So find the registration link on chatgpt experiment.com and hopefully I'll see you there.

[D]

Okay, good.

[C]

Well, I hope that was helpful and perhaps I'll see you in a couple weeks. And remember, that conversation is going to be your best friend as using ChatGPT it's not about the perfect prompt. It's not not about the perfect instructions. It's really talking to it like you would talk to a person. So the more you do that, the more you become comfortable in giving it context and details and quite frankly asking it does it have any questions, the better off you're going to be. Okay, that's it for this week. Be curious as always and we will talk soon.

[D]

Bye bye.

[B]

This episode is brought to you by Progressive Insurance. You chose to hit play on this podcast today. Smart Choice. Make another smart choice with Auto Quote Explorer to compare rates from multiple car insurance companies all at once. Try [e-mail] Progressive Casualty Insurance Company and affiliates not available in all states or situations. Prices vary based on how you buy.

[E]

Hey, this is Dan Harris, host of the 10% Happier podcast. I'm here to tell you about a new series we're running this September on 10% happier. The goal is to help you do your life better. The series is called Reset. It's all about about hitting the reset button in many of the most crucial areas of your life. Each week we'll tackle a topic like how to reset your nervous system. How to reset your relationships. How to reset your career. We're going to bring on top notch scientists and world class meditation teachers to give you deep insights and actionable advice. It's all delivered with our trademark blend of skepticism, humor, credibility and practicality. 10% happier is self help for smart people. Come join the party.