ChatGPT Adoption Methodology

Here’s how I think people, organizations, and corporations should adopt AI through the ChatGPT Adoption Methodology.

About the author

Iliya Valchanov
CEO at

I teach data science and AI to 1.2M students on Udemy. I'm co-founder of 365 Data Science, 3veta, and Team-GPT. My work is mostly focused on helping people and organizations adopt AI.

ChatGPT is still pretty novel.

How do you learn it? How do you teach it? How to adopt GPT?

I’ve been thinking a lot about the best path to GPT adoption.

What is stopping teams from adopting GPT?

This essay is the optimal form I found to share my vision with you.

Here’s how I think people, organizations, and corporations should adopt AI through the ChatGPT Adoption Methodology.


I’ve been reading about GPT as much as I can.

I teach AI to 1.2M people, so I know a thing or two about it (my top course The Data Science Course: Complete Data Science Bootcamp, 630K students, 130K reviews).

But since GPT-4 came out I felt compelled to help people use it, understand it, and finally adopt it.

The problem is that many people who I respect are denying or soft-denying (not using) ChatGPT.

Maybe 1 in 5 of my friends really use it.

Seeing this pushback, I made a focused effort to understand GPT technology well enough to be able to teach it. This focused effort had one goal:

Create a ChatGPT Adoption Methodology.

The final result is a combination of a:

  • An interactive online course
  • A collaborative chat interface
  • Knowledge organization

Until the end of the article, I will try to explain the ChatGPT Adoption Methodology.

What is the need for this?

GPT is the best thing since the Internet and we need to master it.

2 billion people will have to adopt AI in the next 10 years.

Here’s how.

The ‘Pyramid of GPT Knowledge’

Many learning methodologies are based on hierarchical assumptions.

I’ve taken inspiration from some of them because I believe a hierarchical build-up is the right way to think about learning. Mainly because of the fallback mechanisms (you see something you don’t know and then you try to compare it with something you know. This really helps you accumulate new knowledge better and make logical fallbacks).

Either way, you have some fundament, then several layers of knowledge. The higher you go, the more masterful you are.

For example, school is the fundament of everything (for most people who were schooled). Afterwards, this knowledge is built upon in university, then work habits, etc. We develop knowledge: layer by layer. Professional knowledge assumes you’ve got the fundament sorted out.

If you have never been to school, everything else is very hard.

To adopt AI and to learn to utilize GPT models, we need to build a strong GPT fundament.

Someone needed to come up with a methodological approach we could use. But with ChatGPT being so new, we couldn’t find any good resources. So, we had to create the methodology.

Here is the ‘Pyramid of GPT Knowledge’ – a helper to explain the AI adoption methodology.

There are 3 layers:

  • Fundament: Build a strong fundament and it will hold the pyramid. This was my job with the ‘ChatGPT for Work: The Interactive Course’
  • Relevance and Specificity: The fundament is generic. How do you make the learning relevant and specific to the student. No matter the student…
  • Mastery: This is when GPT is seamlessly integrated in the mind

Ultimately, we want to reach Mastery, I presume. Otherwise, why read this essay?

AI Adoption must be = Mastery.

But there are so many problems.

Let’s start with the elephant in the room.

The Elephant in the Room: People don’t understand ChatGPT

‘Smart people’ are saying GPT is the future.

But not even they know why.

The truth is that a ‘sign-up’ in ChatGPT is not enough. You need to actually use it.

Many people create an account, have 1 conversation and go away. They never really understand it.

ONE conversation is NOT enough

My hypothesis is that there is a critical number of chats before you reach the ‘aha’ moment.

The moment when you simply know that GPT is something else entirely.

My guesstimate (based on the people I’ve helped learn) is that to ‘get ChatGPT’ you need 3-5-8 conversations on average, depending on the person and their own AI inclination. I’m convinced that at the 10th fruitful conversation, even the biggest sceptic shall be converted to a ChatGPT fan.

Either way, 3-5-8 chats are not many, but they are not few either.

How do I make people have these?

This became the premise of the ‘ChatGPT for Work: The Interactive Course’.

Make sure they have 10 successful conversations.

💡 These 10 conversations are from ‘ChatGPT for Work: The Interactive Course’

The course is interactive, if you take it inside the Team-GPT platform.

ChatGPT for Work: The Interactive Course contains:

• 100+ tried and tested prompts
• 100+ exercises
• 100+ extra notes and comments
• 1000s of use cases
• PDFs files for later reference

It is also completely FREE.

Start today and become a ChatGPT expert in less than 7 days!

The 10 conversations

This is the most important block of the Fundament.

Going through the ‘ChatGPT for Work’ course, practically means having 10 good conversations.

Of course, there are many other materials and useful knowledge BUT… 80% of the way is to make a person ‘feel’ the AI. Afterwards, they can and will explore it on their own.

How to ensure the conversations are good?

The holy grail of why the methodology works.

It is the prompt: ‘Ask me several clarifying questions to get more context before answering’.

This prompt makes ANY chat relevant to the person having it.

This ensures that these chats will be successful.

This really is the key to having the ‘aha’ moment.

What are the 10 conversations?

I had to pick 10 for the ‘ChatGPT for Work’ course.

I don’t love the choice, but I think it is a good balance.

Other people using this methodology should pick theirs carefully.

I’m leaving descriptions below them as to why I think they are useful:

1. Write an email

This is the most basic use case I could think of. At the same time, every email is different. I believe many people have tried to write an email with ChatGPT unsuccessfully because they didn’t know how to talk to ChatGPT. In the course, we ensure that the email will be good.

2. Conflict resolution at work

This is a soft skill. Everyone has had some kind of conflict at work. It is a very broad issue, but at the same time very personal to the student. It opens the door to all types of other soft skills, relationship advice, etiquette, and more. I think both employees and managers can quickly see several applications.

3. How to learn to code

This is obviously the learning use case. You can dig as deep as you want in these. ChatGPT can endlessly answer you. I don’t particularly like this because of the whole ‘data until 2021’. When you are trying to learn something, GPT models will always be a bit outdated. Still, I have learned many things from ChatGPT and I’m sure this prompt will be useful to many people.

4. Have better sleep

I wanted to touch on the ‘health-related’ capabilities of using ChatGPT. However, without getting into ‘doctor territory’. Either way, ChatGPT has been known to be very helpful in these. I think a lot of the people relate to this use case and can have very fruitful conversations.

5. Fix frozen phone

Hard(er) to Google. I was looking for example of an issue that people have had that are hard to Google. I’m sure you can Google it, but ChatGPT is just instant. Instead of asking someone else, you can now ask ChatGPT for everything. It is like your techy smart friend who knows everything. But instant and better.

6. How to sell a product better

This is the brainstorming use case and/or strategy. Whatever conversation you have, you will receive a reply with many different ‘pieces of advice’ by ChatGPT. It provides so many suggestions that at least one of them will be useful. If the chat is long enough, I think any marketer would stick with GPT forever.

7. What is string theory

This is the research use case. Obviously GPT models are amazing at things that are written in books, textbooks and other scientific texts. That’s what the example aims to show.

8. Write a poem

This is not useful for work. However, it helps the person discover the whole ‘constraints’ and ‘adjustments’ technique on their own. Apart from this it is kind of ‘the fun exercise’.

9. Write an essay

The essay here is shown right after poem. Placing them one after the other shows the user how versatile ChatGPT is and how useful it is on both topics. Also… people need to know what ChatGPT can do, because all students can use GPT for their essays – it is this simple. This will be a very big problem for the whole educational system going forward, so I wanted to raise awareness.

10. Write a report

This is the ‘structured text creation’ use case. ChatGPT is amazing at repetitive tasks. And I couldn’t remember anything more repetitive than reporting (reporting is repetitive by definition).

These 10 conversations are by no means perfect. I’m sure they are not all relevant for everyone.

However, these are so universal that they are sufficient to gain the critical 3,5,8 conversations.

All of them show a different side of ChatGPT.

This should be about enough to ‘get it’.

And remember, we are using the prompt: ‘Ask me several clarifying questions to get more context before answering’. This greatly helps the favorable outcomes.

Also an ‘Unlimited Exercise Generator’ is provided with the course. It helps people find relevant questions for them.

This is also the main reason why the course is called ‘The Interactive Course’. The exercises make it interactive. Also, once in the GPT platform, you will also start talking about other things with the AI, facilitating the student to interact even more.

Once the 10 conversations are had, the rest is simple.

The Promise

The ‘ChatGPT for Work: The Interactive Course’ makes the promise that you’ll easily become a good prompter if you follow these rules:

  • Have 10 high-value conversations with the AI
  • Learn a couple of simple tricks (let’s go through this step by step, ask me clarifying questions, give 5 variations, use less adjectives, etc.)
  • And continue asking the AI any question you have, any problem you face, try to test the boundaries of the tool

The course also features extra materials to supplement the student.

The video part is ~1 hour long, so the course is short and therefore ‘easy to start and finish’.

All the videos are extremely short and practical. They focus on the ‘Prompting Methodology’ and the different ‘Techniques’ but not on ‘work use cases’.

After completing it, the students get the one thing they need: the high school of GPT.

The ‘ChatGPT for Work’ course is the Fundament of the GPT knowledge.

It equips the student with everything they need to move up the pyramid.

Relevance and Specificity

We are still on the path of AI adoption.

Imagine you have managed to get someone to go through the ‘ChatGPT for Work: The Interactive Course’.

What is the next step? How do you make it relevant to them? No matter who they are…


Team-GPT is a collaborative environment where teams adopt AI. The key there is that the whole team enters the platform and can organize its knowledge.

Team-GPT’s main question:
How do you make ChatGPT chats RELEVANT and SPECIFIC?

Customizable workspace in the software.

An adoption software.

Used by in-house GPT evangelists.

In-house GPT Evangelists

Right now, less than 10% of the people in companies are ‘really excited’ about ChatGPT.

I call these GPT evangelists.

Team-GPT’s hypothesis:
You can find at least 1 AI evangelist in a team trying to adopt AI. Find these people, get them in Team-GPT. Provide them with ‘ChatGPT for Work’ course. This is the fundament. Then through the customizable software make it easy to organize knowledge in folders, subfolders, saved prompts which are for your own industry. Once the workspace is set up, everyone else from the team joins and gets an ‘aha’ moment.

This GPT Evangelist knows their own team and the use cases. They must be a domain expert.

A good prompt engineer is a good DOMAIN EXPERT.

Let’s give an example, because this is important.

Anyone who finishes the ‘ChatGPT for Work’ course will be better than me at their own domain. For instance, any lawyer will be a much better LAW prompt engineer than me. Why? Because they can define better GOALS and you can provide better CONTEXTabout law.

The GPT Evangelist must be a domain expert.

Not a young recruit who doesn’t understand the business. It has to be someone who knows how the company works. If you decide to get outside help, make sure the prompt engineers or consultant is a domain expert in your field.

Either way, this Evangelist could be internal or external (with the detail from above).

The Evangelist would use the Team-GPT software to prepare the workspace for their team. They will also be the main drivers (other names I’ve heard: ‘AI adoption project manager’)

💡 Team-GPT: Where people, organizations, and corporations adopt GPT models

Team-GPT is a collaborative environment for teams to use ChatGPT.
Team-GPT Enterprise allows organizations to deploy Team-GPT on their own server, have their own private database and use the Microsoft Azure OpenAI Service.
This is most private and secure way to utilize GPT-4 in your company.

What needs to be done?

The job of the GPT Evangelist will be to bring the domain expertise to the software.

Example chats and prompts need to be prepared. This could also be a ‘view only’ template (this is what ‘ChatGPT for Work’ is inside Team-GPT – a template workspace). Any company can create their own templates that are relevant for the company.

This is a crucial step in the adoption as every person needs to see the direct benefit to their work.

The only way to reach Relevancy and Specificity is to bring the specific domain expertise INSIDE of a collaborative knowledge sharing platform (i.e. Team-GPT).

This relevant and specific knowledge is organized by the GPT Evangelist through the software capabilities.


Finally, at the top of the pyramid is Mastery.

Mastery is when it becomes hard to distinguish if something was done by human or AI.

Mastery is when you stop thinking about ‘how can the AI help me’ and just know it.

Mastery is when you dream bigger because the AI is already an extension of yourself.

The ‘extension’ part is key.

Let’s see how GPT models can be an extension of ourselves.

ChatGPT is a revolutionary TOOL

Tool is the key here.

When the first humans started using TOOLS: hammerstones, stone cores, sharp stone flakes, this allowed them to hunt, build houses and so on. Without going through the whole history of humanity, tools have been very useful and were practically the big reason why humanity was evolving.

When talking about a TOOL, there is a very interesting phenomenon.

I’ll quote MIT Press on this:

When a tool in your hand “becomes part of you,” it’s not just a metaphor… It’s real. Your brain makes it real.

This is very important:

Even if the brain just pretends that the tool is part of its body, then the tool is a part of its body.


When you are inside a car, especially driving, you become ONE WITH THE CAR. You stop thinking about the car, you ARE the car. In every maneuver in your head, you are no longer just human. You are also a car.

This is Mastery – when your brain makes it real. When you become ‘the car’.

In the same way, pilots are one with the plane, captains are one with the ship.

If you don’t drive a car, you are experiencing a similar feeling when holding: a tennis racket, baseball bat, knife, phone, computer mouse, or any other object that can be used as a tool.

You are ‘extended with these tools’.


So, we know 2 things:

  • ChatGPT is the most useful tool ever
  • Human brains can make it feel like a real extension

The user stops thinking about ChatGPT as a tool.

The user starts thinking about ‘Human+AI’ as one whole.

During ‘ChatGPT for Work: The Interactive Course’ we mention ‘assisted intelligence’ and ‘assisted brain’. This is what is meant.

Once Mastery is achieved (Human+AI level), the it’s not just a metaphor… It’s real. Your brain makes it real.

Time is no longer measured as: I need X hours to do this on my own. ->
Time starts being measured as: Since I am using ChatGPT, I need Y hours to do this.

Mental limitations are unlocked by the brain, because it is extended (assisted).

Bigger goals can be conceived and executed.

Can you afford to skip on AI adoption?

The times they are a-changin’

People who adopt GPT early will have an unfair advantage.

Unfair, because Human+AI can do much more than Human.

With each year that passes, the early adopters of GPT models will have bigger and bigger rewards for using the technology. They will not only be more productive but also disproportionately more capable.

But should there be a gap at all?

With the proper adoption approach, we don’t have to leave anyone behind.

Upskilling everyone will elevate humanity as a whole.

This GPT adoption methodology works.




If you found this resource useful, please share with 1 friend.

It helps a lot.

About the author

Iliya Valchanov
CEO at

I teach data science and AI to 1.2M students on Udemy. I'm co-founder of 365 Data Science, 3veta, and Team-GPT. My work is mostly focused on helping people and organizations adopt AI.