How To Get A Consistent Voice Writing With AI
Voice Lab Skill
Every month, we run live in-person bootcamps for readers who want to go deeper with AI— small cohorts, taught live, focused on one specific AI skill:
Claude
Claude Code
Claude Cowork
Personal Hermes Agents
And more!
Our live events are where we show you how we use AI across our $8,000,000+ portfolio of digital businesses, help you apply the latest AI tools and best practices to your own work, and answer your questions live.
If you want to hear about them first (before we announce publicly), click here.
Hey there!
The biggest mistake writers make with AI is thinking voice is universal.
They build one giant voice.md file
They dump every rule, every example, every framework into a single project.
They open one chat and use it for newsletters, LinkedIn posts, YouTube scripts, sales emails, and everything in between.
And then they wonder why the output keeps getting worse.
After writing thousands of pieces with AI across newsletters, social content, books, and marketing materials, I can tell you the problem is almost never your prompt. The problem is how you’ve organized your workspace.
So today, I’m going to share the 3 rules I use to get consistent, on-voice output from AI every single time.
Let’s walk through each one.
Rule 1: Isolate one thread per content type.
Voice isn’t universal—it changes by format.
The way you write a newsletter is different from how you write a short-form post, which is different from how you write a YouTube script.
When you try to cram all of those rules into one thread, the AI is essentially juggling multiple rulebooks at once. Which means the output gets less consistent, less sharp, and more generic over time.
Instead, dedicate one thread (or project, or chat) to one content type. One thread for newsletters. One thread for short-form social. One thread for marketing materials. One thread for YouTube scripts.
Why?
Because AI performs dramatically better when the rules are narrow and specific. When you say, “All we’re doing here is writing this specific type of thing, and these are the rules for writing this specific type of thing,” the model can actually stay in its lane. The output gets sharper, more consistent, and closer to your actual voice.
The moment you start mixing formats in the same thread, you’re asking the AI to hold conflicting rules in its head. And that’s when the writing starts drifting toward the generic middle.
Rule 2: Even within the same format, know when to start a new thread.
Let’s say you’re writing newsletters, and you spend a month writing a bunch of really strong, opinion-based newsletters in one thread.
They’re all bold takes. Personality-forward. Punchy.
Then one day, you decide you want to write a newsletter that’s more fact-driven and data-based. Same format (newsletter), but a completely different style.
Even though these are both newsletters, the rules are different.
The rhythm is different. The tone is different. And if you try to write that data-driven piece in the same thread as your opinion-based ones, the AI is going to smuggle in the personality from the earlier writing—because that’s the pattern it just learned.
Here’s my rule:
If the style of what you’re writing changes significantly (even within the same format), open a new thread.
Yes, it’s more work. But the output will be worth it.
Rule 3: Set up a context hub to solve the “AI forgets” problem.
Unfortunately, AI doesn’t have a great memory.
You spend hours uploading documents, explaining your target reader, sharing your voice guidelines, walking through your niche... and then a week later, you open a new chat and have to do it all over again.
The workaround is a Notion or project hub.
For example:
Build one central Notion page for your business that includes:
A full write-up of your target customer (who they are, what they care about, what they’re struggling with).
Your positioning and niche (what you write about and why).
Your voice guidelines (voice DNA, editing rules, examples of your best work).
Your content principles (educational vs. opinion, formats you use, formats you don’t).
Any proprietary frameworks or language you’ve developed.
Then every time you start a new chat for a specific content type, you link that chat to the Notion hub. The AI reads the hub, absorbs the context, and then you tell it what specific thing you’re writing today.
Think of the Notion hub as your persistent brain. The individual chats are your working memory. You separate them so neither has to do the other’s job.
You open a new chat, say something like, “Hey, we’re writing newsletters today. Here’s the Notion hub with all of my context: [link]. Read through it, then we’ll get started.”
The AI goes and reads it, and now you’re operating with full context.
From there, you can write as many newsletters as you want in that thread—as long as they follow the same style rules. When those rules change, start a new chat and repeat.
Digital Writer’s Voice Lab
Now, "write down your voice guidelines" sounds simple until you actually try to do it.
So I built a free Claude Skill called the Digital Writer's Voice Lab that walks you through it—you paste in samples of your best writing, and it extracts your voice DNA into a profile you can drop straight into your Notion hub.
Just click on the link below to get access to it—and click on the “Duplicate” button near the top right corner to make a copy.
