How to Write Compelling Content with AI (That Doesn’t Sound Generic)
Learn the exact AI-assisted writing workflow Jossie uses to create authentic, high-quality content that sounds like her—not generic or robotic. This step-by-step process helps leaders write faster, refine their voice, and use AI as a multiplier without losing originality.
The biggest complaints I hear about AI-generated content are: “It sounds robotic.” “It all comes out the same.” “It doesn’t sound like me.”
And honestly? That’s true, if you just drop a topic into ChatGPT and ask it to “write me an article.”
But I’ve learned that if you approach it differently, AI can help you create content that’s not only unique but also deeply in your own voice.
Here’s the process I use every week to write newsletters and blog posts that don’t sound generic:

Step 1: Start with context
I don’t just ask AI to “write a newsletter.” I give it the guardrails up front:
- How long should it be?
- What tone do I want?
- Who is the audience?
- How do I want it formatted?
Pro tip: I’ve even built custom GPTs so I don’t have to repeat myself every time about what “good” looks like for me.
Step 2: Get the summary
I ask AI for a summary of the topic.
This does two things: it shows me what the AI “thinks” the topic is about, and it gives me a chance to correct any misconceptions before we go further. If something’s off, misleading, incorrect, or just not my angle, I fix it right here.

Step 3: The five-question rule
This is the magic. I tell AI: “Give me five questions, one at a time, about this topic that will help create a truly unique and compelling article.”
I answer each question. At the end, I also ask: “Is there anything else you need from me?”
By the time I’m done, I have all the raw material that makes the piece mine.
Step 4: Build an outline (and refine it)
Instead of having AI write the whole piece at once (which almost always feels bland or misses details), I ask it to create an outline. Then I tweak it. Make sure the structure actually makes sense before moving forward.
Here’s why this matters: ChatGPT especially struggles with creating complete long-form content in one shot. It gets generic. It gets truncated.
(Side note: Claude tends to do better with longer content generation in one go, but I still prefer the outline approach for maximum control.)

Step 5: Go paragraph by paragraph
From there, I ask the AI to go paragraph by paragraph: introduction first, then each section. AI writes, I review, and we build it together. This is where you get content that actually flows and doesn’t sound like it was written by a robot trying to hit a word count. You get more depth this way, but sometimes you’ll see repetitiveness – that’s where you need to provide feedback.
Step 6: Refine and compare
Here’s a technique I love: I’ll sometimes run this entire process in both ChatGPT (using voice in the web version on my laptop, a game-changer because it feels like I’m just talking out my ideas, which keeps the writing natural) and Claude. Then I have one AI critique the other’s output.
You can also have the same AI critique its own work.

Step 7: Final edit in my voice
Once I’m about 80–90% of the way there, I stop. That’s the inflection point where it’s better to edit myself. Otherwise, I fall into the trap of endless iterations that waste time.
This is where you remove the unnecessary em dashes, the forced comparisons, the weirdly formal language. You’re not starting from scratch. You’re refining something that’s already good into something that’s authentically yours.
The more you use this process, the better you’ll get at recognizing that inflection point. Trust yourself to know when to take over.
Your Voice
Before you even start this process, do this exercise: Take 3-5 pieces of your existing content and ask an AI to “generate a description of my voice that you could use to create content that sounds like me.”
Save that description. Use it in your custom GPT instructions or paste it at the beginning of each conversation. It makes a massive difference.

My actual workflow
At this stage, I also leverage AI for all the other details I need to publish my content. I have custom GPTs for:
- Generating 5 title options for any piece of content
- Creating SEO-friendly blog post titles
- Writing email subject lines
- Generating meta descriptions and SEO keywords
These aren’t necessary to start, but once you’re using this process regularly, they’re incredible time-savers.
The result?
Content that feels like me and not like every other AI post floating around the internet. And the process actually saves me hours every week.
AI is a tool for multiplication, not replacement. Use it to get further faster, but never let it replace your voice. The process I’ve outlined here ensures you’re always in the driver’s seat and the AI is just handling the grunt work.
That’s how you create content that’s both efficient to produce and high quality.
Most leaders have access to AI tools, but no systematic approach for leveraging them for leadership challenges. If you’re a technology leader who wants to multiply your impact (not just your task list), I can help. I combine comprehensive leadership coaching with strategic AI integration, so you can focus on your superpowers and the uniquely human work that creates real differentiation. Let’s talk.
