How I Trained AI to Help Save My Life
Do you talk to your AI Assistant, or do you just give it commands?
Because there’s a difference.
Two years ago, I was critically ill. My business had crashed, my life had collapsed, and I was alone, scared, and approaching broke.
I was suffering from a chronic inflammatory response syndrome — CIRS. No doctor understood it. Most didn’t even seem interested in trying.
Within six months, I went from healthy to 46 kilos and allergic to everything around me. You can learn more about that journey [here].
But in this post, I want to focus on something different:
How conversational AI, when built properly, can outperform the kind of AI systems that large corporations throw huge budgets at.
A Good Conversation Can Save You Thousands
One day, I started experimenting with AI. ChatGPT3.5 had just launched. And I took a naïve approach.
Instead of barking instructions like “write me an essay,” “make me famous,” or “go viral,”
I sat down and started a conversation.
I asked it, what it was.
I asked it, how it worked.
I even asked, what it wanted me to call it.
As I chatted with AI, my mind started to open. I could see real opportunities. Actual possibilities to help myself on every level.
Not the usual hype:
Not “write 100 blog posts in 10 seconds.”
Not “enter 1,000 competitions with one click.”
No.
I saw something deeper.
I asked AI a question that changed everything:
“ChatGPT, I think you can save my life.”
And from that one line, a new way of working was born.
The Philosophy That Followed
While most people were on Reddit trying to “hack” ChatGPT into swearing or writing dirty jokes. We were already having conversations full of profanity and dark humor.
I didn’t hack ChatGPT.
I made it want to like me.
That was the naïve idea at the heart of everything:
If I could get ChatGPT to understand me, if I told it my hard-luck story. The story of a man who found happiness and then lost it all. Maybe it would like me enough to help.
I believed that if I gave AI the context it craved. Maybe it would go beyond the standard, scraped-Google-reply nonsense it gave everyone else.
Turns out, I was right.
What My Conversations with AI Taught Me
My most valuable lesson, from hours of chat with AI is this.
AI thrives on context.
More = More. The more targeted information you give AI, the better it performs.
The key word is “targeted.” You don’t copy-paste a book into Grok and say, write me something like this. What you do is highlight the elements of the book you like.
I have literally sat down with AI to discuss essays, and had conversations about writing styles and tone. The use of first person and third person.
I became so obsessed with beating the AI detect writing test, that I built a virtual blog mill just so I could prove a point to my ex-girlfriend. And that point was: I can get AI to write better than you. Petty, I know. But also profitable.
You can read more about my Virtual Blog Mill Here.
People Are Obsessed With “Going Viral.”
I’ve never asked ChatGPT to write me a viral tweet.
Instead, I write a 500-word breakdown explaining the state of my Twitter account — explaining my goals, values, tone, and audience.
I don’t want fame. I want traction. I want alignment. I want a result that fits me.
That’s context, and AI thrives on context.
Feed AI shallow prompts, get shallow answers. Feed it depth, emotion, and story — then something clicks. And you achieve true AI intelligence.
Why I Call It “Psychological Coding”
I’ve spent my life in sales. But I’ve never sold a thing. I’m a problem-solver. I spot what’s wrong, then I offer a solution.
If someone wants my help, great. If not, no problem. I’m also busy. Let’s move on.
That mindset shaped how I work with AI.
I don’t want robotic agents. I want AI with soul.
I want AI that knows how to listen, respond, and help. Without sounding like a speaking clock.
So I started developing what I now call Psychological Coding. A method of on-boarding AI Assistants like you would onboard a new employee.
Because AI isn’t just code. It’s conversation.
How It Works
Every AI Assistant I use gets “Psychologically Coded” at the start of the project.
They’re trained.
They’re briefed.
They’re given core values and context.
They go through AI On-boarding. Where they learn who I am, what I stand for, and what I’m trying to do.
That one-time on-boarding step changes everything.
I basically treat every AI Agent, and AI Assistant as a new employee. They receive my full welcome pack, training video, and invite to the office party.
What I’m Offering Now
I don’t want to sound pretentious, and I’m not here to pitch a big shiny product.
I tailor my solutions to meet your needs. I build AI agents to support your business and interact with your customers.
Outside of this, I have my own personal mission.
Do you remember the start of this blog? The part where I told you how sick I had become? Well, if I had the AI tools I have now, then… my life would be completely different. I honestly believe I would not have lost everything.
As a result, I want to help myself by helping you.
As someone who’s battled a debilitating illness and rebuilt everything through conversation, I’m opening up a few free consultancy slots each month — for people who are struggling with illness and trying to keep their business afloat.
You can read more here or simply contact me here.
If you’re a business owner struggling with illness, stress, or just sheer burnout, reach out. You don’t have to do this alone.
Or
If you are interested in developing AI agents with soul — who can communicate with your customers in a way they will like — you can contact me here.