This might be controversial, but I'm convinced most founders are using AI completely wrong.
I see it all the time. They buy a subscription to the latest AI writing tool, feed it a one-sentence prompt, and expect magic. What they get back is a bland, soulless article that sounds like it was written by a corporate robot. It’s the definition of garbage in, garbage out. Then they complain that AI is “overhyped.”
The problem isn’t the AI. It’s the user. They’re using a supercomputer to do basic arithmetic. They’re using it as a replacement for writing, when they should be using it as a partner for thinking.
For the past 10 months, I went deep down the rabbit hole. I spent over $10,000 of my own money testing dozens of AI tools, trying to find a workflow that actually delivered results. Not just faster content, but better content. A workflow that could save me serious time without sacrificing quality. I run companies, I angel invest, I write books—my time is my most valuable asset. I needed a real solution, not a toy.
What I discovered is a framework that has fundamentally changed how I work. It’s not just about writing blog posts. It’s about drafting investment memos, preparing board updates, and even outlining chapters for my books. This system now saves me at least 7 hours a week, every single week. And the quality of my output has gone through the roof.
The Market is Full of Junk
Let's be honest. The AI productivity space is a circus. Every other day there's a new "game-changer" that promises to automate your entire life. Most of it is junk. They’re thin wrappers around the same core APIs, all competing on marketing fluff rather than real innovation.
My journey started with a simple goal: find the best AI writing assistant. I tested everything. The big names, the up-and-comers, the obscure tools from indie hackers. I tracked my time, the quality of the output, and how much editing was required. I was ruthless in my evaluation.
Here’s a hard truth: not a single one of them is a “fire and forget” solution. Not one. But some are dramatically better than others if you know how to use them. The secret isn’t finding the perfect tool; it’s building the perfect workflow.
Stop Using AI to Write, Start Using It to Think
The big mindset shift is this: stop treating AI as an intern you can delegate writing tasks to. Start treating it as a tireless, infinitely knowledgeable sparring partner. Its job isn’t to write for you. Its job is to help you think, to challenge your assumptions, and to do the heavy lifting of research and synthesis so you can focus on the high-value work of strategy and storytelling.
I call my system AI-Assisted Cognition (AAC). It’s a simple, three-step process.
Step 1: The Outline is Everything
I never, ever start by asking an AI to "write an article about X." Instead, I start with a conversation. I treat the AI like a new hire I’m briefing on a project.
My process looks like this:
Brain Dump: I throw all my raw, unstructured thoughts into the chat. I’ll say things like, "I want to write about the common mistakes founders make with AI. Key ideas are: focusing on the wrong metrics, using it for generic content, not having a real strategy. I want to include an anecdote about my experience at RemoteTeam and how we used automation to scale our content engine. The tone should be direct, a little provocative."
Socratic Sparring: I then ask the AI to act as a critic. "What’s weak about this argument? What are the counter-arguments? What am I missing?" This is where the magic happens. The AI will poke holes in my logic, suggest alternative angles, and force me to sharpen my thinking before I’ve written a single word.
Structured Outline: Once I have a clear argument, I ask the AI to structure it. "Create a detailed, multi-level outline for a 2000-word article based on our conversation. Include a hook, an introduction, three main body sections with supporting points, and a strong conclusion. For each point, suggest a potential anecdote or data point I could use."
This outlining phase is 50% of the work. It’s the most important step. The quality of your outline determines the quality of the final piece. A good AI tool for this is one that has a long context window and strong reasoning capabilities. You need it to remember the entire conversation and connect the dots.
Step 2: The "Skeleton Draft"
Only after I have a rock-solid outline do I ask the AI to start writing. But I don’t ask it to write the whole article. That’s a rookie mistake. I ask it to create a "skeleton draft."
I go section by section through my outline and have the AI flesh out the basic points. The goal here is not to get a polished draft. The goal is to overcome the "blank page" problem and get raw material on the screen. The output is often generic, and that’s okay. I’m going to replace most of it.
This is the trick that saves me hours. Instead of staring at a blinking cursor, I have a structured document filled with basic arguments and transitions. My job shifts from creator to editor and storyteller. It’s infinitely easier and faster.
Step 3: The Human Layer (Where the Real Value Is)
This is where I come in. I take the skeleton draft and I inject my voice, my stories, my opinions. This is the 20% of the effort that creates 80% of the value.
- I add personal anecdotes. Where the AI wrote, "It is important to use AI strategically," I write, "I remember one week at MovieLaLa, we were burning cash on content that wasn’t ranking. I realized our whole strategy was flawed..."
- I add hard numbers. Where the AI wrote, "AI can lead to significant time savings," I write, "We cut our content production time from 15 hours per article to 4, and our organic traffic went up 300% in 6 months."
- I add my voice. I rewrite the clunky, robotic sentences. I add short, punchy phrases. I make it sound like a human being, not a machine.
This final step is non-negotiable. It’s what separates great content from the sea of AI-generated mediocrity. The AI does the grunt work, but the human provides the soul.
The Tools I Actually Use
So, after spending over $10,000, what did I land on? My stack is surprisingly simple.
I won't give a comprehensive list of every tool I tried, but I'll categorize the ones that stood out and the ones that fell flat.
Top Tier - The Powerhouses:
- Anthropic's Claude 3 Opus: My go-to for the initial brainstorming and outlining (Step 1). Its massive context window and superior reasoning ability make it the best sparring partner I've found. It feels less like a tool and more like a brilliant, tireless collaborator.
- OpenAI's GPT-4: Still a beast for the skeleton draft (Step 2). It’s fast and incredibly good at taking a detailed outline and generating structured, coherent text. It’s the workhorse of my workflow.
Mid Tier - Good for Specific Tasks:
- Perplexity AI: Excellent for research. When I need to find specific data points or summaries of recent events, it’s often better than a standard search engine. I use it to feed my outline with facts and figures.
Bottom Tier - The Disappointments:
- Most "AI Writer" Tools: I’m not going to name names, but almost every tool marketed as an "AI writer" was a disappointment. They are too restrictive, their interfaces get in the way, and they encourage the exact kind of low-effort prompting that leads to terrible content. They are crutches, not power tools.
My advice? Don’t waste your money on a dozen specialized tools. Get a subscription to a state-of-the-art model like Claude or GPT-4 and master the workflow. That’s it. That’s the whole stack.
This is a New Muscle for Founders
Learning to use AI this way is like developing a new muscle. It feels awkward at first. Your first few attempts will probably still sound a bit robotic. But stick with it.
Start small. Try using the AAC framework for your next internal email. Then a short blog post. Then a bigger, more strategic document. Track your time. See the difference for yourself.
We are at a turning point. In the next few years, there will be two kinds of founders: those who master AI as a thinking partner, and those who are left behind, churning out the same generic content as everyone else.
Don’t be one of them. Stop looking for a magic button. The trick isn’t the tool, it’s the process. Now go build it.
Frequently Asked Questions
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.