The Solo Creators AI Studio./Every AI skill I build, a new drop each month, plus a direct line to me.
See inside the Studio
AI WRITING

Make AI Write Like You: The 12-Rule Writing System

You've been banning words one at a time. Here's the fix: give the machine an actual writing system, paste it in once, and stop editing the same voice out of every draft.

Make AI Write Like You: The 12-Rule Writing System
Ronnie Nijmeh
By Ronnie Nijmeh
Updated July 2026 · 9 min read
See inside the Studio
Every skill I build, yours to keep
A new skill drop every month
A direct line to me, an 18-year marketing vet
Built on $14M in sales, 550M emails

Key takeaways

  • A banned-word list is whack-a-mole. A writing system is a style you paste in once, and it's the only version that scales.
  • The fastest AI tell is corrective juxtaposition, the "not X, it's Y" shell. It gives a plain statement the shape of insight without adding a fact.
  • Orwell's six rules from 1946 do most of the work. Rule six, break any rule sooner than write something barbarous, is the one people skip and need most.
  • The 2026 patch handles the modern tics: no straw men, two examples not three, don't announce what you're about to say, and vary your sentence shapes.
  • The twelve rules stop it sounding like a machine. Only your own five lines stop it sounding like everyone else who pasted the same twelve rules.
  • Put the system in your global instructions, not in each chat. A system you have to remember is a system you skip on the day you're rushed.
  • I built a gate for this tic and it flagged 81 instances on my own site, then caught me planting a fresh one an hour later. Systems beat willpower.
The Solo Creators AI Studio

Get this skill, and every one I build.

The whole vaultA new drop monthlyBuild live with meYours to keep, forever
See inside the Studio

30-day money-back guarantee. Cancel anytime.

Questions people ask

What is the "not X, it's Y" pattern actually called?
The rhetorical family is antithesis, and the corrective form is correctio. Vox (@Voxyz_ai) calls the AI version corrective juxtaposition, which is the clearest name for it. It shows up as "the answer isn't X, it's Y" or "X isn't the problem, Y is." The structure is legitimate and ancient. JFK's "ask not what your country can do for you" is the same shape. The problem is frequency, not the device.
Why does AI use that pattern so much?
Vox's read is that it gives a plain statement the shape of insight without adding a fact, and that training supplies the pattern while post-training rewards the clean, decisive sound. That matches what you see in practice. The sentence feels like it concluded something, then you reread it and nothing new arrived.
Do I have to ban the pattern completely?
No, and you shouldn't. Rule seven says use it once per piece, maximum. Used once, deliberately, it's a strong move that writers have relied on for two thousand years. Used six times in one article, it's the thing that makes a reader feel like nobody was home.
Where exactly do I paste the writing system?
Put it in your global instructions so every new chat picks it up without being asked. In the Claude desktop app that's in Settings. If you work in projects or workspaces, each one can carry its own instructions on top. Pasting it into a single chat works, but only for that chat, and you'll forget on a busy day.
Will this make my writing pass an AI detector?
That's the wrong goal, and detectors are unreliable anyway. A US regulator found one popular detector performed no better than a coin toss. Write the system for your reader, who is a person deciding whether to trust you. Optimizing prose for a machine that guesses is how you end up with worse writing and a subscription.
How long does the voice-capture step take?
About ten minutes. Pull five to ten things you actually wrote and sent, hand them over with the prompt above, and ask for five short lines. The only real requirement is that the samples are genuinely yours. Feeding it writing you admire but didn't write produces someone else's voice, cleanly.
Does this work outside Claude?
Yes. The system is plain text with no tool-specific syntax, so it works in any assistant that lets you set persistent instructions. The mechanics differ (where the global instructions live varies), but the twelve rules and your five lines travel anywhere.
Isn't this just a prompt?
It's a prompt you write once instead of a correction you make every time, which is the whole difference. The reason it works is placement. Living in your global instructions, it applies to drafts you didn't think to supervise, and those are the ones where the robot voice usually slips through.
Found this useful? Share itPostShare
Ronnie Nijmeh
Written by Ronnie Nijmeh

Ronnie spent 18 years building a SaaS with a team of 20 that served over 650,000 customers, generated over $14M in sales, and sent over 550M emails. Now he's solo, solving real business bottlenecks and turning them into working AI skills, workflows, and automations. He teaches all of it, with direct access to him, inside the Solo Creators AI Studio Skool community. See what he's built →

“I completely rebuilt my site using the Website Launch Builder. If you've ever built a website, you'll really appreciate this skill. I'm blessed to be a member.”
Michael HattawayMichael HattawayFounder, Iron Strengthens Iron

Become a one-person team that outships a roomful of people.

You don't build any of it yourself. You get the finished AI skills, agents, and workflows, plus a direct line to me.

Built by Ronnie. 18+ years in marketing. $14M in sales. 550M emails. 650K customers.

See inside the Studio

30-day money-back guarantee. Cancel anytime.

See inside the Studio →