AI Copied Me: Inside the CLEAR Copywriting Method

People tell me my copy sounds like AI. Here is the part that makes me laugh. Every large language model on the market was built by scraping the internet: textbooks, marketing books, and a decade of social media. My work is in all three. The machine did not teach me how to write. I was part of its curriculum.

That is only half the story, and it is the less interesting half. The models did not just eat the good stuff. They ate everything. That is exactly why the copy they hand you sounds the way it does.

Back in 2016, I Was a Nobody

In 2016 I was nobody on social media. My posts on Facebook and LinkedIn barely got a like, a comment, or a share, just a small circle of friends and family.

Some weeks the only comment was my mom saying “proud of you son.” I would delete the post out of embarrassment.

I did not suddenly get better at writing.

Starting in 2017 I deconstructed 10,000-plus social media posts, picking apart the psychology behind why some content spreads and most does not, subtle mechanics most people never notice, let alone name.

That's where CLEAR came from.

Before AI, I Was Already a Data Nerd

Long before anyone trained a model on scraped internet content, I had a real dataset of my own.

My teammate Clarence, who has been with me for over a decade, has spent that whole time collecting, organizing, and tracking every post I have published since 2016.

Tens of thousands of posts. Every one logged: content type, length, whether it carried media, theme, and the only numbers that actually matter, likes, comments, and shares.

Every month I pulled the top 10 most-liked and top 10 most-commented posts and studied them line by line. Every year-end, my team and I took the full dataset and used it to plan the next twelve months.

3,768 posts.
One year. Just Facebook.

That fact became its own post, because by then even the meta-commentary was working.

That is the difference between what I built and what a language model does.

A scraper pulls everything, the viral post and the ignored one, with no idea which is which. My dataset only ever had winners in it, because a human being sat down every single month and threw out everything that did not work.

Fast forward to today: 6.5 million people reached in the last 30 days alone. Millions of views a month is the floor, not the ceiling.

What C.L.E.A.R. Actually Stands For

Most people who quote the method have never seen it written down. Here it is, the same five steps I have taught from stages and inside university curricula.

• arrow_forwardC, Curiosity.The opening line exists to do one job: make them need the next line. It should leave three questions hanging. What happened? How did it happen? Why did it happen? My most-shared post of all time opens with two words. “It sucked!”
• arrow_forwardL, Lead Up.Answer the “what” and give them a reason to keep going. “I had to fire someone yesterday. I have not had to do that in over a year.”
• arrow_forwardE, Emotion.Say how it actually felt, not how you wish it felt. This is the step everyone skips because it requires being a person in public. “I tried to pretend that hiring someone new would make me feel better about it, but it didn't.”
• arrow_forwardA, Analyze. Explain why it happened and show your proof. This is where the reader starts nodding, or arguing. Both are good.
• arrow_forwardR, Reinforce. Land the lesson, tie it back to your opening line, and give people something worth reacting to. Reactions are the whole point. They are what the algorithm counts.

That is it. No swipe file, no template pack. Five steps that force a story to earn its own attention.

I have used it to sell directly off a post, not just collect likes. One CLEAR post did over $2,500 in 48 hours, and $400 of that came from running the exact same post as a $7 Facebook ad.

50,000 Posts. Written by Hand.

Over the last decade I have written more than 50,000 social media posts. Thousands of them went viral. Not one was written by AI, and most of them predate AI entirely.

Some were mine. Most were for clients.

I worked alongside Gary Brecka and Grant Cardone on the launch of 10X Health System. I worked with Agora Financial, one of the largest direct-response publishers in the world, teaching their publishers how to use AI to write long-form copy.

That is the part people get backwards. I was not taking a course on AI copywriting. I was teaching it to the people who write copy for a living.

Two different things, and the difference matters.

I own the Claude Community: over 1.5 million members, built from zero. And I am an AI Expert inside the world's largest AI community, four million members, a group I did not build and do not own.

Anyone can grow a room if they get there early enough. Earning expert standing in somebody else's four-million-member room is a different test, and it is the one that proves the method travels.

My posts inside that one group have pulled more than 100 million views. Not across my whole career, and not counting my own communities. One group, one method.

100,000,000+ views.
One group. One method.

In the Textbooks

Dr. Karen E. Sutherland teaches social media strategy at the university level, and my work is inside two of her Palgrave textbooks: Strategic Social Media Management: Theory and Practice (now in its second edition) and Artificial Intelligence for Strategic Communication.

Combined, they are used in university courses worldwide, over 300,000 copies in circulation.

In Dan Kennedy's Book

Dan S. Kennedy basically invented direct-response marketing as most people know it.

My work shows up in No B.S. Guide to Successful Marketing Automation, co-authored with Parthiv Shah, the no-holds-barred guide to technology, automation, and artificial intelligence. Sharing a cover with Dan Kennedy is not something I take lightly.

What the Machines Actually Ate

Every major language model was built the same way. Scrape as much human text as possible, then learn to predict the average of it.

Textbooks got scraped. Two of them have my method in them.

Marketing books got scraped. Dan Kennedy's has my work in it.

Social media got scraped. That is 50,000-plus of my posts, tracked and refined over ten years.

Nobody asked me. Nobody asked you either.

Go ask ChatGPT, Claude, or Gemini who Jeff J Hunter is. They will have something to say. Not because I bought placement, there is no placement to buy. Because I published in public for a decade and the crawlers took all of it.

Here Is Why AI Copy Sounds Like AI Copy

But my work is a rounding error in that dataset, and so is every other good writer's.

The models also ate every abandoned blog, every desperate cold email, and every “I'm humbled and excited to announce” post nobody read. There are vastly more bad marketers on the internet than good ones. It is not close.

So when you type “write me a Facebook post,” you get the mathematical center of a mostly mediocre pile. Bland, hedged, three adjectives deep, terrified of saying anything.

Average in.
Average out.

That is not the model failing. That is the model doing exactly what it was built to do. It is hitting the target. The target is just the middle.

Which means the entire game is aim. A method like CLEAR is a set of instructions for which fraction of everything it learned to actually reach for.

Then I Taught One to Write Like Me

This is the part where my ego dies.

My top three performing posts of the past year were not written by me. They were written by an AI copywriter I trained on my own method, my own tracked dataset, and my own voice.

One of them:

2,727,710 reached.
765,095 clicks.

3,256,911 impressions. 83,599 reactions, comments, and shares. From a post I did not write.

I was training models to write viral social copy before most marketers had heard the phrase prompt engineering. Not because I am a better technologist than they are. Because I already had the one thing a model cannot scrape: ten years of my own data telling me exactly which posts worked and why.

A generic model averages everyone. Mine was pointed at the winners, and it knew what a winner looked like because a human being had labeled them one month at a time.

I Was Early on Claude, and I Took Heat for It

I was the first person in my network to go all in on Claude. Not ChatGPT, which is what everyone else in marketing was demoing at the time. Claude.

I caught a lot of flack for it. Colleagues told me I was helping automate copywriters out of a job. Some said AI writing was a fad that would embarrass anyone who leaned on it. A few stopped taking my calls.

I was not guessing. I had spent a decade tracking, line by line, what makes a piece of writing spread. That is exactly the background that lets you look at a language model and see precisely what it can and cannot do, and where a human still has to hold the wheel.

I did not think it was the biggest opportunity of the decade. I thought it was bigger.

So I proved it on myself. I built an entire product, AI Copy Tools, using AI for the product, the marketing, the emails, and the sales page. It did $70,000 in a single weekend. Entrepreneur Magazine ran the story on page 52. I broke down exactly how that launch worked here.

Today I run the Claude Community, over 1.5 million members and the most active room of Claude users on Facebook. The people who told me I was early stopped saying it a while ago.

It Is Not Just Me

A method that only works for the person who invented it is not a method. It is a personality.

I ran a challenge: write one CLEAR post, today or tomorrow. Thirteen posts came back. They pulled 4,560 reactions, comments, and shares between them.

Facebook's average engagement rate sits under 3 percent. To earn 4,560 engagements at that rate, those thirteen posts had to be seen roughly 150,000 times.

On the Cover Line of Entrepreneur

The March/April 2025 issue of Entrepreneur Magazine put “10X Your Sales” on the cover.

My $70K weekend launch made the cover line: “The Marketing Automation That Built a $70K Launch,” page 52. I told the full story of that launch here, if you want the breakdown.

Quoted in Success Magazine

Success ran a feature called “Building $10M+ Businesses” in their May/June 2024 issue, Tony Robbins on the cover, and pulled me in directly:

Zero Paid Ads Since 2016

Here is the number that actually matters.

$10,000,000+.
Zero paid ads.

Every dollar came from organic content built on the same method sitting in those textbooks. Not a funnel hack, not a growth-hacking trick, just copy that works, published consistently, for over a decade.

I write long-form. I write story. That is the part of copywriting I actually love, the part most people skip because it is slower and harder than a swipe file. It is also the part that compounds.

Last 28 days, on my personal profile alone: 6,184,741 views, up 255 percent, and 578,422 engagements, up 309 percent. That is one man and a method, not an ad budget.

Ten years of writing has taught me that people are not accusing me of writing like a robot because I sound generic. They are accusing me because I sound consistent, and consistency at this volume looks automated even when every word is mine.

So yes, my copy sounds like AI to some people. The irony runs the other direction.

AI copied me.

It copied a lot of people worse than me, too, and that is the real problem. The model will hand you the average of everything it swallowed unless you tell it precisely where to aim.

CLEAR is the aim. Ten years of tracking what worked, compressed into five steps, so you are not stuck asking a machine to guess what good looks like.

If you want the full method and the AI copywriter I built on top of it, that is what the AI Persona Method is.