Preview
What it is.
Ours, and we sell the service it runs inside. Two passes run at the end of our content chain: a Content Brand Enhancer at v11 that applies brand voice across a draft, and a Humanizer that strips the sentence patterns which read as machine-written.
You get a draft that survives a read-aloud. The v11 is the signal worth reading, because each version came from a failure someone caught in review, so a pattern list like this gets built over time and not written once.
What you get.
- A brand voice pass applied consistently across a draft
- Removal of the sentence patterns that read as machine-written
- A versioned pattern list, so a corrected pattern stays corrected
- A late-stage pass that runs once the figures are already verified
How to set it up.
Keep voice and pattern removal as two passes. Enforcing a brand voice and deleting AI tells are different edits and they interfere when combined.
Build the pattern list from your own rejected drafts. Ours is at v11 because each version answers a failure someone caught in review.
Run it after fact-checking. Rewriting a sentence before the figure in it is verified means checking it twice.
Keep the brand voice definition in one place the skill reads, so the pass is the same for every writer.
Version the skill and note what each version corrected, which is how a voice pass stops regressing.
For a public equivalent, copy-editing runs 7 sequential sweeps with a reason attached to every change.
Pricing Plans
Not sold as a skill.
It runs inside client engagements. The published account dates from April 2026.
Use cases
The pass before publishing
A fact-checked draft still reads like a machine wrote it. Run the pattern pass and the recognizable sentence shapes get stripped, so the piece survives a read-aloud before it goes out under your name.
One voice, many writers
Several writers feed one brand account. The voice pass reads a single stored definition and applies it to every draft, so the output matches no matter who wrote the first version.
Growing your pattern list
You're building an equivalent pass in-house. Version the pattern list and add an entry each time review catches a new tell, since a list built from your own rejected drafts catches more than a generic one.
Best for
AI-assisted drafts going out under your name
The recognizable patterns are what a reader notices first, and a versioned list catches them more reliably than a fresh instruction each time.
Several writers, one voice
A brand voice pass the skill reads from one place beats a style guide nobody opens.
Building your own voice pass
The v11 detail is the lesson: build the pattern list from your own rejected drafts instead of a generic list of AI tells.