An AI Roasted My Writing
Here's what it told me.
Today's big reveal: I occasionally ghostwrite for my portfolio companies.
After finishing a recent piece of content for one of our businesses, I did something I'd never tried before. I asked ChatGPT to dissect my entire writing process. (Because apparently I’m now outsourcing self-reflection to the robots. Don’t lie. You’re doing it too.)
I fed the AI my rough first draft, the polished final version, and every prompt I'd used to iterate along the way. "Analyze how the content evolved and what made each revision stronger," I told it. Actually, here's the much-more-specific prompt I used:
Its feedback was surprisingly insightful. Not just about the finished product, but about the fundamental mechanics of good writing.
The patterns it identified (the specific moves that improved the piece's clarity, engagement, and persuasion) can apply to virtually any content you're creating. Especially if you're using that content to help you attract new customers, which, let's be honest, is what most of us spend our time trying to do.
Below is a summary of the feedback (and the associated prompts) the AI gave me.
Steal the prompts.
Save some time.
Have some fun.
1. Industry Expert Voice > Marketing Voice
Feedback: When positioning yourself as as a thought leader, write like someone talking to an equally-credible peer, not like a marketer writing about the expert's expertise. Marketing instinct is to polish and perfect, but authentic expert language is more credible than marketing-optimized language.
AI Prompt: “Review this content and identify any language that sounds like marketing copy rather than authentic expert voice. Rewrite those sections to sound like a practitioner talking to peers. Use more direct language, rougher edges, and conversational tone. Remove corporate jargon and overly polished phrasing."
2. Acknowledge the Emotional Reality Before Delivering Analysis
Feedback: Manifestos and thought leadership work when they meet people where they are emotionally, not just intellectually. Address the doubt, frustration, or confusion before providing the framework or solution.
AI Prompt: “Analyze this content's opening. Does it acknowledge the emotional state of the target audience before diving into analysis? If not, add 2-3 sentences that recognize what the audience is likely feeling about this topic (uncertainty, frustration, skepticism, etc.) before presenting our perspective."
3. Specific > Generic (Always)
Feedback: Concrete details make you sound like an insider who really understands the problem. Abstract language makes you sound like a consultant who read about the industry. Specificity builds credibility.
AI Prompt: “Identify all abstract or generic statements in this content. For each one, either replace it with a specific example, number, or concrete scenario, or flag it as needing more specific supporting detail. Replace phrases like 'significant challenges' with specific examples and 'major impact' with actual numbers or scenarios."
4. Let the Expert's Authentic Language Shine Through
Feedback: Source material (interviews, transcripts, conversations) often contains authentic phrases and insights that are more powerful than cleaned-up marketing prose. The "rough" expert voice can be more persuasive than the "smooth" marketing voice.
AI Prompt: “I'm providing source material [transcript/interview/notes] along with this draft. Identify 3-5 phrases or insights from the source that sound distinctly authentic to the expert. Incorporate these authentic voice markers into the content, preserving the expert's natural speaking style rather than translating everything into corporate language."
5. Structure for Curiosity, Not Just Information
Feedback: Good thought leadership pulls readers forward by creating curiosity gaps, not just by organizing information clearly. Headers and transitions should make people want to keep reading, not just understand what's coming next.
AI Prompt: “Review all section headers and key transitions in this content. Rewrite any that are purely informational to create curiosity gaps instead. Each header should make the reader think 'I need to know the answer to this' rather than just 'This section will cover X topic.'"
6. Don't Bury the Sales Message. Integrate It Naturally.
Feedback: When you establish authentic expertise first, commercial messaging becomes a natural extension rather than a jarring shift. The best thought leadership weaves positioning throughout the expert insights rather than tacking it on at the end.
AI Prompt: “Review how our company/solution is positioned in this content. If it appears only at the end or feels like a separate sales pitch, revise to weave our positioning naturally throughout the expert insights. Our commercial message should feel like a logical extension of the expertise demonstrated, not an abrupt shift to selling."
If you’re a CEO or run a sales/marketing team and are wondering how to train your people to get more out of AI, there’s a lot more useful prompting strategies in this post.
Thinking about an outside speaker or keynote on the topic? Send me a message. I might be able to help.