01 — What You Get at the End
Four Assets You Walk Away With
Everything is platform-agnostic. If AI tools change tomorrow, your voice assets transfer in an afternoon.
Voice Vault
All of your best writing, organized by content type and ready for AI. Books, posts, newsletters, speeches, transcripts. Your permanent content library.
Voice DNA Document
A single document that captures exactly how you write. Opening patterns, sentence rhythms, vocabulary, emotional range, structural habits, and what you never do.
Working AI System
A configured Claude Project that generates content in your voice for any format. Books, speeches, social posts, newsletters. Your team uses it independently.
Calibration Process
A built-in feedback loop that makes the system more accurate over time. Every correction compounds. The more you use it, the better it gets.
02 — What the Voice DNA Captures
Your Writing Fingerprint
The Voice DNA Document is the most important piece of the entire system. Every section below is extracted from your actual writing.
Opening Patterns
How do you start a piece? A bold statement? A personal story? A direct greeting? Most people have 2–3 signature ways they open.
Sentence Patterns
How long are your sentences? Short bursts or flowing paragraphs? Where do you use line breaks? This is the rhythm of your voice.
Vocabulary & Word Choice
Phrases you use repeatedly. Words you gravitate toward. Language you avoid. Your fingerprint at the word level.
Emotional Range
How you express excitement, handle vulnerability, challenge your audience, and teach. The full spectrum of how you show up on the page.
Structural Patterns
How you structure a newsletter versus a social post versus a blog post versus a book chapter. Each format has its own architecture.
Beliefs & Guardrails
Ideas that show up in your writing again and again. Words you never use. Tones you avoid. The safety net that prevents AI from drifting.
03 — Step 1
Collect Your Voice
Gather your best 15–25 pieces of content into one organized folder. Variety matters more than volume. You want AI to see you when you're teaching, storytelling, motivating, being vulnerable, and being funny.
What to Collect
•Books and manuscripts (3–5 chapters that show different sides)
•Social media posts (your best 10–15 across platforms)
•Newsletter archives (your 5–7 best editions)
•Speeches and keynote scripts or transcripts
•Podcast transcripts (2–3 solo or deep conversation episodes)
•Blog posts and articles (your 3–5 best long-form pieces)
•Personal writing, journals, voice memos, emails you're proud of
Prioritize content you wrote yourself, especially anything from before you started using AI. That's your purest voice. Content that was AI-assisted may already have AI patterns blended in.
How to Organize It
Create a single folder your team can access. Google Drive, Dropbox, OneDrive, iCloud. Name it "[Your Name] Voice Vault." Inside, organize by content type: Books & Manuscripts, Social Posts, Newsletters, Speeches & Keynotes, Podcasts, Blog Posts & Articles, Personal Writing.
This folder is your permanent asset. AI tools will change. Platforms will evolve. Your Voice Vault works with any AI system, today or five years from now.
04 — Step 2
Build Your Voice Engine
This all happens in one place, and this project becomes your permanent home for creating content in your voice going forward.
1
Create a New Claude Project
Go to claude.ai, click Projects, and create a new one. Name it "[Your Name] Voice Engine." This is where you build your Voice DNA and create content going forward.
2
Upload Your Voice Vault
Add the files from your organized folder into the project's knowledge files. Books, posts, newsletters, transcripts. Everything you collected in Step 1.
3
Extract Your Voice DNA
Open a new conversation inside the project. Claude already has access to all the content you uploaded. Use this prompt:
Analyze the writing I've uploaded to this project. Create a Voice DNA Document that captures exactly how I write. Don't describe what I write about. Capture HOW I write: my opening patterns, my sentence rhythms, my vocabulary and phrases I use repeatedly, my emotional range, how I structure different formats, what I never do, and the beliefs and themes that show up across everything. Be specific. Use examples from my writing wherever possible. The output should be detailed enough that another AI reading this document could write new content that sounds like me.
4
Review and Refine
Read through the Voice DNA that Claude generates. Ask yourself: does this sound like me? Flag anything that feels off. Tell Claude what to fix. This usually takes 2–3 rounds.
Examples: "I never start sentences with 'And.'" or "I use way more commas than you're showing here." or "My newsletter always opens with 'What's good, it's [name]' and you missed that."
5
Save and Install
Save the Voice DNA as a markdown file (your portable backup) and paste the full document into the project's custom instructions. Every future conversation inside this project will write in your voice without you having to ask.
You are the final authority. AI surfaces the patterns. You confirm what's accurate and correct what isn't. Nobody knows your voice better than you.
05 — Step 3
Test & Calibrate
Your Voice DNA from Step 2 is a strong starting point. Calibration is where you catch the things it gets slightly wrong and fix them permanently. Each correction makes the system more accurate.
1
Generate Something Real
Open a new conversation in your Voice Engine project. Give it something real. A story that happened to you, an idea you've been thinking about, a lesson you learned. The more specific you are, the better the output.
2
Read It Out Loud
Don't just scan it. Say it like you're talking to someone. Ask yourself: does this sound like me? Not whether you agree with the content, but whether the rhythm, tone, and feel match how you actually communicate.
3
Flag What Doesn't Sound Right
Tell Claude exactly what's off. Be specific. "I would never open a paragraph that way." "That word isn't me." "This section is doing too many things at once."
4
Update the Voice DNA
Tell Claude: "Update my Voice DNA Document with these corrections. Add each one to the section where it belongs. Give me the full updated Voice DNA so I can copy it." Then save and re-install the updated version into your project instructions.
What to Expect
First Test
You'll find 3–5 things that need correcting. Completely normal.
Third or Fourth
Down to minor tweaks. The system is dialing in.
By the Tenth
Publishing with light edits. The system sounds like you.
The key rule: Every correction goes into the Voice DNA, not stored somewhere else. One living document that gets more precise every time you use it.
06 — Step 4
Create
The system is live. The Voice DNA handles how you sound. You handle what you say. The more specific and real you are with your input, the better the output. Your Voice Engine is one piece of the broader AI context foundation your business runs on.
What You Can Create
•Social media content (LinkedIn, Instagram, Facebook, X)
•Newsletters (share the insight, AI drafts it in your voice)
•Speeches and keynotes (topic, audience, key stories in, draft out)
•Website copy, bios, talking points, interview prep
•Books and long-form writing (chapters drafted in your voice)
•Thank you notes, pitch decks, book forewords
How Your Team Uses It
Shared Access
If your team is on a Claude Team plan, they access the same Voice Engine project directly. Everyone works inside the same project. The Voice DNA keeps everything on voice.
Individual Setup
If each team member has their own account, share the Voice DNA file with them. They paste it into their own project instructions and upload the Voice Vault content. Same output, different account.
The biggest tip: Don't type "write me a LinkedIn post about leadership." Instead, speak what's actually on your mind: the story from this morning, the lesson from a conversation, the idea you can't stop thinking about. When you bring the depth and the Voice Engine brings your voice, that's when the output feels like you actually wrote it.
07 — Common Questions
Tips & FAQs
How accurate is it? Will it actually sound like me?+
First drafts land at 75–80% voice accuracy. That means AI does the heavy lifting and you're editing, not writing from scratch. The calibration loop pushes accuracy higher with every cycle. Within the first few weeks of active use, most people see 90%+. The more you use it, the better it gets.
What if I don't like what it produces?+
That's the calibration process working exactly as designed. When something doesn't sound like you, flag it. Each correction gets absorbed into the Voice DNA permanently. The first few rounds are the most active. After that, the corrections get smaller and less frequent.
What happens if AI changes? Am I locked into one tool?+
No. By design. The Voice Vault lives in your cloud storage. The Voice DNA is a text file. If the entire AI landscape shifts, your voice assets move to the next platform in an afternoon. You own the asset forever.
How is this different from telling AI to "write in my style"?+
Telling AI to write in your style gives you generic output with a thin layer of personality. The Voice Engine gives AI a detailed map of how you actually communicate: your sentence rhythms, your vocabulary, your emotional range, your story structure, the phrases you use, the words you never use. Then the calibration loop catches every place AI drifts off voice and corrects it permanently. Anyone can give AI a style instruction. The Voice Engine gives AI your actual fingerprint.
Can it handle different types of content?+
Yes. The Voice DNA captures how you write across different formats. Your Instagram voice, your book voice, your speech voice, your newsletter voice. They're all variations of you, and the system understands the differences. When you ask for a social post, it writes like your social posts. When you ask for a book chapter, it writes with the depth your longer work has.