I use AI 365 days a year to run a multi-million dollar business.
These 5 rules will 10x the quality of your AI output:
(thread)
Rule 1: Think autistic PhD student, not magic genie
AI is very smart but you need to be literal, detailed, and iterative.
Don't ask: "Help me with marketing"
Do ask: "Write 3 LinkedIn posts for entrepreneurs struggling with productivity. Each post needs a scroll-stopping first line, a brief story, and a soft CTA to download my time management guide."
Rule 2: Break everything into modules
AI works better when you give it one focused job at a time.
Instead of asking for a complete landing page, ask for:
- The headline first
- Then the subhead
- Then the bullet points
- Finally the CTA
Each piece builds on the last with full context.
Rule 3: Stack your prompts like building blocks
Never one-shot complex requests.
Start broad, then narrow:
"List problems faced by busy entrepreneurs"
"Focus on the top 3 productivity issues"
"Create a lead magnet for the biggest problem"
"Write the landing page copy for that magnet"
Rule 4: Push back and iterate
AI's first output will be mediocre. Always.
Use phrases like:
"This is too generic. Be more specific."
"This sounds like AI wrote it. Make it more human."
"This is boring. Add more emotion."
The magic happens in rounds 3-5, not round 1.
Rule 5: Document everything valuable
AI forgets everything between conversations.
Keep a master doc of:
- Prompts that work
- Outputs you approve
- Instructions for tone and style
- Examples of what you want
Feed this back to AI every time for consistent results.
Most people treat AI like a Google search.
One question, one answer, then wonder why it sucks.
These 5 rules turn AI from a toy into a business-building machine.
What used to take us 10 people and 3 weeks now takes 1 person half a day.
That's not hyperbole. That's what happens when you stop one-shotting and start treating AI like the professional tool it actually is.
Look, I get it. You want to build something but you're already working 60-hour weeks.
You don't have time to learn coding. You can't afford a team. You're stuck.
Until you learn how to accelerate your progress with AI.
In my next free masterclass, I'll walk you through the exact framework that lets you launch a business on nights and weekends with the help of AI.
This isn't theory.
This is the same system I used to build and sell companies for hundreds of millions.
But now you can do it without the startup capital, without the team, without the years of trial and error.
I used AI to reverse engineer how 12 billion-dollar companies got their first 1000 customers.
Here’s how to do it to find the opportunities in your market. (thread)
Step 1: Find their origin story using AI
Most successful companies have completely rewritten their history. The polished story isn't the real story.
Use this prompt: "Find the earliest archived versions of [company name]'s website, press releases, and founder interviews from 2010-2015. What was their original positioning and first target market?"
AI can dig up old TechCrunch articles, Wayback Machine snapshots, and forgotten founder tweets that reveal the messy truth.
Step 2: Identify what problem they ACTUALLY solved first
Airbnb wasn't about "belonging anywhere."
It was about cheap places to crash during conferences.
Netflix wasn't about "entertainment freedom."
It was about avoiding Blockbuster late fees.
Prompt: "Based on [company]'s earliest customer reviews and testimonials, what specific pain point did their first 1000 customers hire them to solve?"
Market research firms charge $30,000 for insights I get from ChatGPT in 30 minutes.
Here are the 7 prompts that turn AI into your $500/hour research analyst:
Prompt 1: Blind Spot Analysis
"Analyze what [top 3 competitors] are NOT addressing in their marketing, product features, and customer communications. What problems do they ignore or downplay?"
This reveals opportunities they're leaving on the table.
Prompt 2: Customer Frustration Mining
"Find complaints about [competitor products] on Reddit, G2, Amazon, and review sites. What do customers wish existed but no one provides?"
Your next product feature lives in their 1-star reviews.
If you have 60 minutes tonight, you can validate your business idea by tomorrow.
Here's the script:
Most founders waste months building products nobody wants.
I used to do this too.
Built entire companies around assumptions. Lost millions on "great ideas" that failed in the market.
Then I learned to kill bad ideas in 60 minutes instead of 6 months.
Here's my exact validation script with the benchmarks that matter:
Step 1: Problem Hunt (15 minutes)
Search Reddit, Facebook groups, and forums for people saying:
- "Why is there no..."
- "I wish someone would make..."
- "This sucks because..."
Look for these pain indicators:
- "I've tried 5 different solutions and they all..."
- "Paying $X but it doesn't even..."
- "Been looking for months and can't find..."
Copy 10 specific complaints. If you can't find 10 in 15 minutes, move to a different problem.
If I sold my company today, I'd start my next multi-million dollar business in 90 days using only AI. Here's how:
I'd create 5 AI specialists that work 24/7 for a fraction of what a single employee costs.
Here's precisely how I'd structure my AI dream team:
THE ANALYST
I'd upload a detailed vision document explaining my business model
I'd have it identify emerging market opportunities with hard data
I'd ask: "Find the top 3 underserved problems in profitable industries"
I'd demand specific research on customer behavior and spending patterns
I'd make it spot trends competitors are missing
THE GROWTH HACKER
I'd ask: "What's the fastest path to 10,000 customers with supporting evidence?"
I'd have it design rapid testing frameworks for every assumption
I'd make it identify the highest-impact growth levers specific to my business
I'd run every strategy through "What would break this?" analysis
I'd demand concrete timelines and success metrics for each test