Here’s the paradox every writer faces: The more you rely on AI, the more you risk sounding like everyone else using the same tools. After three years of writing with AI to publish a book, I learned that maintaining authenticity isn’t about avoiding AI — it’s about using it strategically while fiercely protecting what makes your voice unique.
The Day I Almost Compromised (and Lost) My Voice
In March 2024, I hit a wall. After weeks of productive AI collaboration, I read through my latest chapter and felt uneasy. The content was technically perfect — clear, logical, comprehensive. It was also completely soulless. As I used AI to parse through over 1,500 pages of interview transcripts, conduct research, and organize content, my latest chapter had become sanitized corporate speak.
My opening chapter about the kidney stone experience, which sits at the core of the manuscript, had been transformed from a visceral personal story into a clinical case study. Gone were the screams of pain, the dark humor about choosing an ER over a pipeline meeting, and the raw vulnerability that made the story resonate.
That’s when I realized: AI had become a crutch, not a supporting tool. If left unchecked, my writing core would become weak and anemic — much like skipping the gym for several weeks.
The Vulnerability Paradox
The more I used AI, the more I realized that vulnerability — the very thing AI struggles with — became my differentiation. AI excels at professional polish but stumbles at raw honesty.
Consider this passage about my COVID experience at Folloze: “Every day, I lost track of time and forgot to eat or drink water. If I remembered, my regular 5-day-a-week gym workout schedule was replaced by a short walk to the corner with my dog. These challenges would physically manifest as stress, spiking cortisol levels, and high blood pressure. I gained 20 pounds and found solace in Talenti Vanilla Bean ice cream.”
AI wanted to clean this up, make it more “professional.” But the specificity of Talenti Vanilla Bean ice cream, the image of walking my dog to the corner — these details create connection. They’re what readers remember.
The Voice Authentication Framework
As I wrote my book, I developed a voice authentication framework (VAF) for maintaining authenticity while leveraging AI’s power. Think of VAF as your accountability coach for maintaining your voice as you write.
Step 1: The Unfiltered Brain Dump
Before touching any AI tool, I now write — and regularly update — my raw thoughts in what I call “Randy uncensored” mode. This Google Doc file (and sometimes a pad of 3×5 cards by the nightstand) includes:
- Personal stories with all the messy details
- Opinions that might be controversial
- Analogies that come from my specific experience
- The occasional profanity (which gets edited out later)
- Stream-of-consciousness connections between ideas
Step 2: Write During Peak Authenticity
Capturing both the underlying passion of your manuscript and the essence of your voice is sacrosanct and must be done without any influence from AI. For me, this process meant protecting my morning writing time between 8am and 11am when my authentic, creative, and smart-ass voice flows most freely — before AI, before editing, before self-censorship.
After 11am, I turn into a pumpkin.
Step 3: The AI Enhancement (Not Replacement)
Only after capturing my authentic thoughts do I engage AI through the VAF, but with very specific parameters:
The Voice Preservation Prompt:
Review this draft while preserving my unique voice elements:
- Personal anecdotes and specific examples must remain intact
- Maintain conversational tone and occasional humor
- Keep contrarian opinions and strong positions
- Preserve emotional moments and vulnerability
- Only suggest structural improvements and clarity edits
Do NOT: sanitize, formalize, or make it sound academic
This approach helped maintain moments like when I describe vendor desperation: “My email inbox reflects the severity of this challenge: hundreds of urgent messages from sales representatives offering $100 Amazon gift cards or premium whiskey. The catch? I need to sit through a 30-minute product demo.”
Step 4: The Authenticity Audit
The secret to maintaining voice with AI lies in feeding it stories and perspectives only you possess. AI can enhance these unique elements but can’t create them from scratch.
For example, my story about Oracle CEO Larry Ellison’s network computer vision in 1995 works because I lived it — managing photo shoots of plastic prototypes while reporters pretended they were real computers. AI helped me structure this story for maximum impact, but the absurdist details (“those models were the most photographed chunks of colored plastic the world had ever seen”) came from my memory.
As Christine Heckart shares in my book: “I am a human catalyst that helps people envision and achieve more than they ever thought possible.” This purpose statement is uniquely hers — AI could never generate this level of personal clarity.
After AI enhancement, I run what I call the “Randy Test”:
- Would I say this in a conference room?
- Can my partner or friends recognize this as my writing?
- Does it include at least one story only I could tell?
- Are my quirks and speech patterns intact?
- Does it make me slightly uncomfortable with its honesty?
If it fails any of these VAF tests, I rewrite.
The Technical Tricks That Preserve Personality
1. Voice Sampling
To effectively train AI (or in my case, Claude) on your unique voice, create a “voice portfolio” by collecting 10–15 of your best writing samples — blog posts, emails, even casual Slack messages that capture how you naturally communicate. Start each conversation by providing 2–3 representative samples and explicitly asking Claude to analyze your style patterns: “Here are examples of my writing. Please identify my voice characteristics including typical sentence structure, favorite phrases, humor style, and how I transition between ideas.”
Then, when requesting content, reference this analysis: “Using the voice patterns you identified, help me write…” The key is consistency — remind Claude of your voice preferences in each session, as it doesn’t retain information between conversations. I’ve found that providing specific examples of phrases I use (“Here’s the thing…” or “Let me explain with a short story”) and my structural preferences (like using threes: “The good, the bad, and the ugly”) yields remarkably accurate voice matching. Most importantly, always review and inject your own flourishes — Claude can approximate your patterns but can’t replicate the experiences and emotions that truly make your voice unique.
The Voice Training Prompt:
Analyze these writing samples for:
- Recurring phrases and expressions
- Sentence rhythm and length variation
- Types of analogies and metaphors used
- Humor style and frequency
- Emotional temperature and vulnerability level
Create a "voice profile" to maintain consistency
2. The Spice Rack Method
Like cooking, I learned to add “flavor” after AI provided the base. My VAF spices include:
- Industry war stories
- Contrarian takes on conventional wisdom
- Specific vendor callouts (when appropriate)
- Self-deprecating humor about my control freak tendencies
- References to kidney stones and avoiding pipeline calls (apparently my signature move)
3. The Interview Integration Technique
When weaving in quotes from my 50+ interviews, I preserved each person’s unique voice by capturing their speech patterns, not just their words. For instance, Kevin Gaither’s directness: “In a product-led growth environment, sales leaders need to check their ego at the door.”
Five VAF Prompts for Preserving Your Voice
1. The Voice Capture Prompt:
Analyze my natural writing style in this sample. Identify:
- Unique phrases I use repeatedly
- My typical sentence structure
- How I use humor or emotion
- My favorite analogies/metaphors
- Words I tend to avoid or overuse
2. The Authenticity Check Prompt:
Compare this AI-enhanced text to my original. Flag anywhere that:
- Sounds too formal or corporate
- Loses personal anecdotes
- Removes controversial opinions
- Sanitizes emotional content
- Changes my natural rhythm
3. The Story Enhancement Prompt:
Help me expand this personal story while maintaining my voice:
- Keep all specific details intact
- Suggest where to add sensory details
- Identify natural places for humor
- Maintain my conversational style
- Preserve the emotional core
4. The De-Robotification Prompt:
This text sounds too AI-generated. Help me:
- Add personal asides and parentheticals
- Insert conversational transitions
- Include minor tangents (as I naturally do)
- Add rhetorical questions
- Restore my speech patterns
5. The Regional Flavor Prompt:
Ensure this text maintains my [regional/cultural] communication style:
- Natural expressions and idioms
- Cultural references that resonate
- Appropriate level of directness
- Humor that translates to my audience
- Professional yet personable tone
When to Ignore AI Completely
Some sections demanded pure human touch:
- Personal failure stories
- Controversial industry opinions
- Emotional revelations about burnout
- Specific critiques of vendor practices
- Predictions about the future
As Neil Cohen notes in my book: “What’s fundamentally missing from performance marketing conversations is the critical role of brand and empathy. We’ve elevated digital marketing and algorithms to be the miracle cure for pipeline problems, but this is misleading.”
This kind of provocative stance requires human conviction that AI can’t replicate.
The Ultimate Test: Beta Reader Feedback
The real proof came from beta readers. Comments like:
- “This sounds exactly like you’re in the room explaining it”
- “I can hear your voice when I read this”
- “This is so Randy”
- “The kidney stone story made me laugh and cringe — perfectly you”
One reader noted: “Chapter 10 felt different — more polished but less personal.” That chapter had the highest AI involvement. Message received.
The Bottom Line on Voice
After three years of AI collaboration, here’s what I know: Your voice isn’t just your writing style — it’s your experiences, biases, failures, and triumphs all shining through the words. AI can polish your thoughts, organize your ideas, and accelerate your production. But it can’t replace the messy, beautiful, distinctly human elements that make readers trust you.
In your day-to-day writing, experiment with the Voice Authentication Framework, best practices, and prompts in this blog. Do what works for you to ensure you don’t lose your voice for the sake of efficiency.
As I conclude in my book: “The difference now is that we’re not just changing engines on the plane but changing engines on the rocket. The destination may be uncertain, and the journey will be challenging, but with some assembly required and the right mindset, you’re more than equipped for the adventure ahead.”
That optimistic-yet-realistic tone? That’s pure Randy. AI helped me structure and refine it, but couldn’t generate the feeling behind it.
Next week in Part 4, I’ll share the unexpected discoveries from this journey — including why AI made me a better human writer and what this means for the future of B2B marketing content.
Ready to find your authentic voice in the age of AI? Get your copy of “Some Assembly Required” at https://www.amazon.com/dp/B0FJZDXZCK
Coming Next:
- Part 4: Unexpected Discoveries & Future Implications
