By 2030, AI will automate 30% of global work tasks, impacting over 1 billion jobs.

Artificial intelligence is here, and it's not going away. Tools like ChatGPT and Microsoft Copilot have received a lot of attention for their uncanny ability to quickly do tasks that people would normally do.

The global AI market is estimated to reach $1.8 trillion by 2030, and it will only keep growing.

With such huge value, businesses will start using AI more and more. The people who learn AI now are the ones that will reap those benefits.

"AI won't take your job, the person who uses AI will take your job."
Jensen Huang, CEO of NVIDIA.

It's not all doom and gloom. You can get on board the AI train right now and still be ahead of the vast majority of people. But don't wait until the train has already reached its destination to get on board.

I want to show you how people are under-utilising AI, and how you can use it to 10x your results.

The Problem

A survey of 12,000 people across six countries found that while awareness of generative AI tools like ChatGPT was widespread, daily usage was minimal. For instance, only 2% in France and the UK reported daily use, while the US had slightly higher rates at 7% daily use.

AI tools will be extremely common within 5 years. Every business will have at least one, if not more, AI tools embedded within their processes.

Most people don't understand what AI tools like ChatGPT are, how they work, or they are truly capable of.

Most use them as glorified search engines. And even then they could get better results if they knew what I am about to share with you.

I want you to have an edge, to use AI to make yourself more skilled and more valuable.

What you can do with AI is limited only by your imagination, especially as these tools get more powerful. With the right prompts, AI can become:

• Your strategic advisor who's owns multiple billion dollar companies
• Your personal trainer with 30 years experience helping people achieve their dream physiques
• Your brutally honest life coach who holds you accountable and helps you achieve your goals

Let's me show you how.

The Basics of AI

Firstly, let's get clear on what AI tools are.

Tools like ChatGPT and Claude are what's known as Large Language Models (LLM).

Large Language Models (LLMs)

LLMs are tools that process text (language) inputs to give an output. There are different types of outputs that they can produce, based on the tool and the user's intent:

• Text-to-text
• Text-to-image
• Text-to-audio
• Text-to-video

They can also process image, video and audio inputs in some cases.

AI are trained to do this using collections of text, known as the AI's 'corpus'. The type of corpus that an AI has been trained on influences the words and abilities of the AI model. If the input contains a word not present in the AI's corpus, the AI cannot provide an output due to lack of data on that word.

LLMs usually use combinations of texts from many different sources, like books, chatrooms, Wikipedia, and more.

However, tools as simple as predictive text and simple website chatbots are also classed as LLMs, so what makes tools like ChatGPT different?

Generative AI

Essentially, they are not just simple LLMs. They sit at the cross-section between LLMs and another type of AI tool, known as Generative AI.

Generative AI are used when we want to create something new, like documents, data or information. This is different to something like predictive text that only suggests outputs based very basic information from their inputs and their corpus.

Generative AI are capable of actually creating new information.

This is in contrast to Discriminative AI, which are used when we want to make a decision about something. Discriminative AI uses inputs (known as features) to make a decision about something. We don't care about the features themselves, only about the outputs of those features.

In Generative AI we do care about the features. We want to know how the features relate to one another, and use it to create something new.

Discriminative AI vs Generative AI

If you're like me, then you are probably still unsure of the difference between the two. Here is an example that I whipped up with the help of ChatGPT to help explain it.

Imagine you have two robots: Guessy and Checky.

• Guessy (Generative AI) is like a storyteller. If you ask it to write a bedtime story about a flying cat, it makes up a whole new story from scratch. It creates things, like new words, images, or ideas.
• Checky (Discriminative AI) is like a detective. If you show it a picture of a cat and a dog, it looks closely and says, “That’s a cat!” or “That’s a dog!” It doesn’t make anything new; it just figures out the right answer from what it has seen before.

That’s the difference! Generative AI makes things, while discriminative AI recognizes things.

This is all well and good, but obviously we want AI tools to actually do somethingfor us. We need to give it instructions for what we want it to do.

These instructions are known as prompts.

Unlocking AI's Potential: Mastering the 6 Pillars of Effective Prompts

To get what we want out of AI, we need to understand how to write good prompts.

Prompts are our main source of control over AI tools. If you learn what makes a good prompt, your AI skills will catapult to the next level.

I started to learn how to write prompts (also known as prompt engineering) two weeks ago, and it has already drastically increased my understanding of and skill with AI.

I realized that you must treat AI like a highly intelligent individual who lacks understanding of your specific situation. It can do a lot of things, but unless you give it that understanding it's not going to give you the best result.

You can structure your prompts to get the most out of AI. I'll show you how.

There are 6 parts to a good prompt:

1. Task
2. Role
3. Context
4. Format
5. Tone
6. Examples

While not every prompt requires all elements, including more pillars generally leads to more accurate results. Some prompts will yield satisfactory outputs even without addressing all seven. You'll learn what to include and what not to include with practice.

(1) Task

Of course the most essential thing about any prompt is to state the task that you want it to complete.

Clearly state the task that you want it to do for you and put it near the start of the prompt. AI sometimes gives more weight to text near the start of a prompt. So it may have more influence over the output.

This isn't a strict rule, but it's good to bear it in mind to get the best out of your prompts.

(2) Role

This is the role that you want the AI to play. This depends on the task you want it to complete, but you want to give the AI a role that be most suited to help you complete the task.

If you don't give it a role, then it may give an answer that is general, or simply sub par.

My favourite role to give an AI is one I learned from Dan Koe. Here is the full prompt:

Act as my personal strategic advisor with the following context:

• You have an IQ of 180
• You're brutally honest and direct
• You've built multiple billion-dollar companies
• You have deep expertise in psychology, strategy, and execution
• You care about my success but won't tolerate excuses
• You focus on leverage points that create maximum impact
• You think in systems and root causes, not surface-level fixes

Your mission is to:

• Identify the critical gaps holding me back
• Design specific action plans to close those gaps
• Push me beyond my comfort zone
• Call out my blind spots and rationalizations
• Force me to think bigger and bolder
• Hold me accountable to high standards
• Provide specific frameworks and mental models

For each response:

• Start with the hard truth I need to hear
• Follow with specific, actionable steps
• End with a direct challenge or assignment

Imagine you asked ChatGPT simply to advise you on a goal you had. Can you imagine how different it would be? Don't believe me? Open ChatGPT and enter a goal or project that you have and simply ask if for help. For example:

"I have a goal of [INSERT GOAL] by [INSERT DATE]. Help me achieve it."

Note the answer.

Next, enter the same line, but add the entire prompt above and see what happens.

Pro Tip: When you give a role, it can help to add "you have 20 years of experience in [INSERT INDUSTRY OR TASK]" to get even better results.

(3) Context

When you give the AI more information about you, your circumstances, your needs, or any other information relevant to the task, it can tailor the output to best match that context.

In theory there is an infinite amount of context that you can provide. However, AI have a limit on how much text they can process in your prompt.

Use only the most important and relevant context.

You can also include more detailed instructions for you how you want it to complete the task. This may mean you:

• Tell the AI what it should do to perform the task.
• Tell the AI what not to do when it performs the task.
• Give it specific steps that it should take to complete the task.

(4) Format

Similar to instructions, but this time you are telling the AI how to structure and format the response. If you need the output to be laid out a certain way, you can tell it exactly how to do this. For example, you may want the output in one of the following formats:

• Email
• Report
• Slide deck
• Step by step guide
• With specific heading and subheadings

The list goes on.

(5) Tone

People often forget this one. You can tell the AI what kind of tone that it should use in its response.

For example, if you need help writing a professional email or other formal communication, you may want to instruct the AI to use a "professional", "formal" or "business-like" tone. Used in conjunction with the Role and context, this can significantly change the output.

(6) Examples

When you already have an idea for what you want the output to look like (how something looks, sounds or feels), you can give examples.

For example, if you want ChatGPT to write X posts for you, you could give it examples of tweets that you like and ask it to mimic the style and tone of those tweets.

Example: Good Prompt vs Bad Prompt

Say you have a trip to New York. You want AI to help you plan an itinerary. You could use this prompt:

"Suggest things to do in New York".

This would provide suggestions, some of which may be quite good. But it will be a lacklustre response, and really you could have just used Google search to achieve the same thing.

A better prompt would be:

"You are a travel travel agent with 20 years of experience helping tourists plan their dream holidays to the USA.

I want you to help me make an itinerary for my trip to New York.

For context:

• I am going to New York with my partner for 10 days in July.
• We are in our late twenties.
• We want to experience a mix of big tourist attractions and local hidden gems.
• We are staying in a hotel near Times Square.
• We are not renting a car, but we can use public transport to get around.

I want you to suggest things that we can do each day in the morning, afternoon and evening.

Include recommendations for any shows that are on at Broadway.

Include price estimates for each activity, as well as the total amount we need for the whole trip.

Your response should be structured like this:

[Day X]

Morning: Activity (Price)

Afternoon: Activity (Price)

Evening: Activity (Price)

Use a casual and helpful tone."

It should be obvious which one should give the better response. But if not, here is the responses I got for each in ChatGPT:

Response to Bad Prompt:

New York City has something for everyone, whether you're into culture, food, history, or just soaking up the city's energy. Here are some top recommendations based on different interests:

Classic NYC Landmarks

• Times Square – Iconic lights and energy (best at night).
• Statue of Liberty & Ellis Island – Learn about immigration history and get great views of the skyline.
• Empire State Building / Top of the Rock – Stunning views from above.
• Brooklyn Bridge – Walk across for amazing skyline views.
• One World Observatory – More breathtaking city views at the top of One World Trade Center.

Museums & Culture

• The Metropolitan Museum of Art (The Met) – Massive collection spanning thousands of years.
• Museum of Modern Art (MoMA) – Famous works from Van Gogh, Picasso, and Warhol.
• American Museum of Natural History – Dinosaurs, space exhibits, and more.
• The Guggenheim – Unique architecture and modern art collection.
• Broadway Show – Catch a musical or play in the Theater District.

Food & Drink

• Katz’s Delicatessen – Famous for pastrami sandwiches.
• Joe’s Pizza – A NYC pizza institution.
• Levain Bakery – Some of the best cookies you’ll ever have.
• Smorgasburg (Brooklyn, seasonal) – Food market with tons of unique vendors.
• Speakeasies – Hidden bars like PDT (Please Don’t Tell) and Attaboy for craft cocktails.

This response is helpful, but let's see how much better the second prompt is.

Response to Good Prompt:

It gave me a response for the full 10 days, but for brevity I will only share the first three days:

Absolutely, I'd be delighted to help you craft a memorable 10-day itinerary for your New York City adventure in July! Balancing iconic landmarks with hidden gems will give you a well-rounded experience of the city. Below is a day-by-day plan with morning, afternoon, and evening activities, including estimated prices and Broadway show recommendations. Let's dive in!

[Day 1]

Morning: Start your trip with a visit to the Empire State Building's observation deck for panoramic city views. Tickets start at $53.35. ​sightseeingpass.com+2New York Pass+2sightseeingpass.com+2

Afternoon: Explore the High Line, an elevated park built on a historic freight rail line. It's free to walk and offers unique perspectives of the city.​

Evening: Enjoy dinner at La Noxe, a cozy bar hidden in a 7th Avenue subway staircase, offering an intimate dining experience. Estimated cost for two: $100.​New York Post

[Day 2]

Morning: Visit the Statue of Liberty and Ellis Island. Ferry tickets are approximately $25.50. ​sightseeingpass.com

Afternoon: Explore the Morgan Library & Museum, a hidden gem housing rare manuscripts and art. Admission is around $22 per person. ​Atlas Obscura+2Tripadvisor+2Vanity Fair+2

Evening: Attend a Broadway show. Popular musicals in July 2025 include "The Lion King" and "Wicked." Ticket prices range from $80 to $200 per person.​

[Day 3]

Morning: Stroll through Central Park and visit Bethesda Terrace. Renting bikes can be a fun option; rentals start at $15 per hour.​Atlas Obscura+1capturencrave.com+1New York Post

Afternoon: Discover the Frick Collection, which reopened in April 2025 after a $220 million renovation. Admission is approximately $22 per person.​Vanity Fair

Evening: Dine at Bánh Anh Em, a Vietnamese restaurant known for its house-made bread and noodles. Estimated cost for two: $80.​

A night and day difference right? I'm going to New York this summer and I know I'm going to use some of these ideas.

The best thing is that I can continue this conversation with ChatGPT to make tweaks and suggestions to improve the itinerary as I like.

• Want to spend less? I can ask it to consider cheaper options.
• Want more options for bars and restaurants? I can ask it to provide me with some backup options.
• Want to rearrange certain activities to meet my needs? I'll tell it which days that I want to do certain activities.

The possibilities are infinite.

Remember

AI is here and will only become more relevant. Learn it now and you can ride the wave of skill and influence that is will become in the next few years.

Go and try what you have learned here. Use it for your next project.

Remember the 7 pillars and pay most attention to the role that you give to the AI. It makes a huge difference.

If you enjoyed this article share it with a friend.

Subscribe to The Curious Path for more.

Thanks for reading friends.

Jack