- COOLDEEP AI
- Posts
- The AI sentence everyone keeps writing
The AI sentence everyone keeps writing
McKinsey, Cisco write it!!! You probably wrote it today.
In partnership with
Are you looking to get in front of 150,000+ AI readers?
Sponsor CoolDeep AI Newsletter. Get $50 Dollar off on first Ad!
Was this email forwarded to you? Subscribe here!
Today’s Sponsor
100+ Claude Code hacks to ship code 10X faster
Top engineers at Anthropic and OpenAI say AI now writes 100% of their code.
If you're not using AI, you're spending 40 hours doing what they do in 4.
These 100+ Claude Code hacks fix that and help you ship 10x faster.
Sign up for The Code and get:
100+ Claude Code hacks used by top engineers — free
The Code newsletter — learn the latest AI tools, tips, and skills to code faster with AI in 5 minutes a day
Hey
Last month, I sent a newsletter and got a shocking reply.
Someone replied back with three words: "reads like AI."
I sat with that for a minute. Because I had used Claude to help draft it. I thought I had cleaned it up. I thought I had made it mine. But something in there gave it away.
And when I went back and read it, I saw it immediately.
"It is not about the tool. It is about how you use it."
That right there. That one sentence. That is the giveaway.
I had no idea this pattern had a name. But apparently every AI model defaults to it, hundreds of times, without being told to stop.
The pattern nobody talks about
There is a construction that AI writes constantly.
It goes like this: dismiss one idea, replace it with another, sound wise.
"It is not X. It is Y."
Sounds sharp. Sounds like insight. But it is one of the most detectable AI writing patterns that exists right now.
Barron's and AlphaSense tracked this pattern inside Fortune 500 company filings. In 2023, it showed up 50 times. By 2025, over 200 times. That is a 4x jump in two years.
And it is not just small companies. Cisco, McKinsey, Accenture, Microsoft. All caught using it in official published content this year.
Here is the Cisco line that was flagged: "In 2025, AI will not just be a tool; it will be a collaborator."
Here is the McKinsey version: "These systems are not just executing tasks; they are starting to learn, adapt, and collaborate."
Read them again. You have seen this sentence shape everywhere. In newsletters. In LinkedIn posts. In emails from people who are clearly using AI and do not realize they are giving themselves away.
I was one of those people.
BTW, have you joined CoolDeep AI daily alert? Keep getting AI updates the moment they drop on your Phone. It hardly takes 5 seconds to opt in. Get CoolDeep AI Alerts.
I am launching a CoolDeep AI community on Skool. What do you want most inside it? |
Why this happens!!!
When you open Claude and say "write me something," the output sounds helpful. It sounds structured. But it also sounds like it came from a machine trying very hard to sound wise.
The problem is not that people use AI. Everyone uses AI now.
The problem is that most people use AI, accept the first draft, and send it.
They do not train Claude on their own voice. They do not tell it what to stop doing. They do not give it any context about who they are and how they actually think.
So Claude gives them the default. And the default includes negative parallelism, banned words like "leverage," "delve," "seamless," "transformative," dead openings like "In today's world," and a hundred other patterns that readers have learned to recognize in 2026.
Your reader knows. They might not say it. But they know.
How to fix this inside Claude?
This is the actual solution. Not a prompt trick. A proper system.
The fix is a single file called anti-ai-writing-style.md
You create it once. You upload it to a Claude Project or Claude Cowork folder. From that point forward, Claude reads it before it writes anything for you. It knows what to ban, what to avoid, and what to rewrite when it slips up.
Here is what goes inside the file:
Step 1: Ban the negative parallelism pattern
Tell Claude to never use the "It is not X, it is Y" construction. Not in any form. Not in headings. Not across two sentences. Not even disguised with softer words like "Sure, X works, but Y is where the real value is."
That pattern alone will clean up a massive part of what makes your writing sound like a robot wrote it.
Step 2: Build a banned word list
Delve. Realm. Harness. Unlock. Leverage. Synergy. Transformative. Seamless. Empower. Streamline. Elevate. Holistic. Game-changer.
Add these to your file. Claude will stop using them.
Step 3: Add your rhythm rules
Tell Claude how you write. Short sentences. Vary the lengths. No em dashes. No "In today's world" openings. No fake transitions like "Furthermore" or "Moreover."
Step 4: Add an audit command
After Claude gives you a draft, type: "Audit this against my anti-ai-writing-style file."
Claude will go back through what it just wrote, catch the patterns it missed, and rewrite the problem areas. You are using Claude to correct Claude. And that works.
Step 5: Keep the file alive
Words that sounded fine in 2024 are dead giveaways in 2026. "Unlock" used to pass. It does not anymore. Every few months, go back through your recent drafts, spot what is starting to feel robotic, and update the file.
The file is your taste, written down. The more specific it gets, the more your writing starts to sound like you again.
Why Readers Are Obsessed With These 12 Free AI Prompts
Want an A+ when it comes to AI? Subscribe to The Current written by Kim Komando and get a free AI Prompt Hack Pack filled with powerful prompts you can enter into ChatGPT to save you hours (and money) every week.
*Partnered
Finally…
People are going to keep using AI. That part is done. The question now is whether your output sounds like 10,000 other people using the same tool with zero customization, or whether it sounds like you.
The difference is a file. One afternoon of setup. And a habit of actually checking what Claude gives you before you send it.
I used to accept the first draft and hope for the best. Now I have a system that makes Claude write in my voice from the start, and then audit itself when it slips.
That is the whole thing. No special prompt. No secret model. Just a file and a process.
If your writing sounds like AI right now, do not be embarrassed. I was there. But do something about it today.
Check out the full Substack post here from Ruben Hassid where he explained this in deep.
CoolDeep AI
Helping you stay ahead with AI & productivity
AI is fueling the Fourth Industrial Revolution and these 10 stocks are front and center.
One of them makes $40K accelerator chips with a full-stack platform that all but guarantees wide adoption.
Another leads warehouse automation, with a $23B backlog – including all 47 distribution centers of a top U.S. retailer – plus a JV to lease robots to mid-market operators.
From core infrastructure to automation leaders, these companies and other leaders are all in The 10 Best AI Stocks to Own in 2026.
*Partner
ICYMI Last Newsletters:
How did you like today's newsletter? |
Was this email forwarded to you? Subscribe here.
P.S. If CoolDeep AI helps you with your AI skills, I also run a daily email newsletter about Digital Product Business, email marketing, automations, Instagram growth and much more. Click here to join for free.
P.P.S. Reply to this email and let me know what else you want me to cover in the next edition. Keep getting AI updates the moment they drop on your Phone. It hardly takes 5 seconds to opt in. Get CoolDeep AI Alerts.
Not subscribed to CoolDeep AI yet?
*Disclosure: Some of the content in our free newsletter and on our website includes paid placements, marked as “Sponsored”, “Partnered” or “Ad”, and may contain affiliate links. If you choose to click or make a purchase, we may earn a small commission. We are not directly connected to the brands we feature. This helps us keep the newsletter free for you.
If you find our work valuable, your support through these links truly helps us continue creating it.
Reply