Reverse Engineering Content with AI

How to Turn Content Envy into Your Next Great Article, Email, or Social Post

I was up late the other night, doom scrolling on my phone. You know how it is. My wife was asleep, the house was quiet except for the gentle hum of the server rack for my home automation projects, and my Australian Shepherd was twitching in his sleep at the foot of the bed, probably dreaming of chasing squirrels.

And then I saw it.

It was a LinkedIn post from a fellow consultant. It was perfect. It was sharp, insightful, had a touch of vulnerability, and a crystal clear call to action that didn’t feel slimy. The comments were blowing up. And my first, gut level reaction was a familiar, sinking feeling in my stomach.

“Fuck, I wish I wrote that.”

We’ve all been there. It’s that unique cocktail of admiration, jealousy, and frustration that I call “Content Envy.” You see a competitor’s blog post that perfectly articulates an idea you’ve been struggling with. You read an email newsletter with a subject line so good you have to click it. You watch a 30 second video that explains a complex topic more elegantly than you’ve managed in a 2,000 word essay.

In the creator economy of 2025, where everyone from a solopreneur to a Fortune 500 company is a publisher, the pressure to produce is immense. That feeling of content envy can be paralyzing. It can make you feel like you’re perpetually a step behind, that all the good ideas are already taken.

But what if I told you that feeling of envy is actually one of the most valuable assets you have? What if you could turn it into a repeatable system for generating your own brilliant content?

For the last few years, especially since the big AI boom kicked off around 2022, most people have been using AI for content in the most basic, and frankly, least effective way. They go to ChatGPT and type (DON’T DO THIS):

“Write me a blog post about the benefits of time management for small business owners.”

And what they get back is… fine. It’s a C+ essay. It’s grammatically correct, factually adequate, and has as much soul as the bottom of your shoes. It’s blocks of generic advice that reads like it was written by a committee of robots, because it was. It has no point of view, no personality, and no chance of standing out.

When I first started my consultancy, I fell into this trap myself. I was trying to scale my output and thought AI was a magic content factory. I learned this one the hard way: content that doesn’t sound like a real person wrote it gets ignored.

The real power of these Large Language Models isn’t as a content factory, its to ask it to teach you, and then help you apply the lesson.

Today, I’m going to show you a method I’ve refined over dozens of client projects and hundreds of hours of my own work. We’re going to reverse engineer great content using AI, so you can turn that feeling of “I wish I wrote that” into “I’m so glad I wrote this.”

The Lightbulb Moment with a Client

Last year, I was working with a fantastic business based just outside of Houston. They sell high-end, durable portable generators. They’re passionate people who make incredible products, but their marketing felt… flat. It was very “Here’s our new generator. It has these features. Please buy it.”

The founder, Darren, called me for a consultation, and he was audibly defeated.

“James,” he said, “did you see the new email campaign from [Big Competitor Name]?”

I had. It was brilliant. The subject was something like “Stuck in the dark? Our power moves with you.” The email told a short, compelling story about a product tester losing power to their business during an emergency. It was vivid, relatable (especially for the folks in TX), and it masterfully wove in the benefits of their new generator without ever feeling like a hard sell.

“We could never write something like that,” Darren said. “It’s just… so good. We don’t have that in us.”

And that’s when it clicked for me. I told her, “Darren, you’re looking at it as a finished painting but overlooking the process. We’re going to use AI to figure how they wrote that, and then we’re going to create an email just as good that speaks to your people.”

He was skeptical, but willing to try.

This is what we did. Instead of asking an AI to “write a cool email about our new dual fuel generator,” we started by giving it the competitor’s email and a very different kind of instruction.

Step 1: The Deconstruction Prompt

The first step is to play the role of a detective. You give the AI the piece of content you admire and ask it to break it down into its fundamental components. You’re not asking for a summary; you’re asking for an architectural blueprint.

Here’s a version of the prompt we used. I call it the Content Deconstruction Prompt.

Prompt: