TL;DR Summary:
AI Tics Myth: Study of 1,000+ pages shows most criticized AI patterns like sentence starters and intro fillers have no impact on engagement.Not Only Hurts: Overused "not only... but also" constructions strongly correlate with lower reader retention and higher bounce rates.Conclusion Kills: Headers starting with "Conclusion" signal low quality, driving the strongest negative engagement at -0.118 correlation.Em Dashes Boost: Contrary to criticism, em dashes slightly improve engagement by enabling more nuanced, detailed writing.The AI Writing Tics That Actually Drive Readers Away
Content creators spend hours debating which phrases scream “AI-generated.” The internet overflows with lists of supposed AI tells. Em dashes get blamed. Certain sentence starters face criticism. But which AI writing tics actually hurt your content performance?
A new study analyzed over 1,000 content marketing pages to find out. The results challenge common assumptions about AI writing problems.
How the Study Worked
Researchers gathered content from 10 different websites across various industries. They looked for the most criticized AI writing patterns:
- “Not only… but also” constructions
- Sentences starting with “then,” “this,” or “that”
- Intro filler like “in this article” and “let’s explore”
- Headers beginning with “conclusion”
- Em dashes
The team standardized everything per 1,000 words. They excluded pages under 500 words. Then they measured engagement rates using Google Analytics data.
The engagement metric tracked sessions lasting 10 seconds or longer. This captures the moment readers decide to stay or leave.
The Surprising Results About AI Writing Tics
Most supposed AI tells showed no meaningful connection to engagement. But three patterns stood out.
“Not Only” Structures Hurt Performance
Phrases built around “not only… but also” showed negative correlations with engagement. When overused, these constructions drove readers away.
One blog post in the study used “not only” and “but also” twelve separate times. Readers noticed this repetitive pattern and bounced quickly.
Tools like Writecream help writers avoid these traps. The platform analyzes top-ranking content to show which patterns work and which ones hurt engagement.
“Conclusion” Headers Were the Worst Signal
Headers starting with “Conclusion” had the strongest negative correlation with engagement at -0.118. This was the clearest warning sign in the entire dataset.
These formulaic endings signal low-quality content. Readers often scroll to the bottom, see generic conclusion headers, and leave immediately.
Em Dashes Actually Helped Engagement
Here’s the biggest surprise. Em dashes correlated slightly positively with engagement rates. Despite widespread criticism, they didn’t hurt performance.
Writers who use em dashes often create more detailed, explanatory content. This nuanced writing style tends to engage readers better than flat, declarative sentences.
What Content Teams Should Do
Stop Obsessing Over AI Detection
Google doesn’t penalize content based on AI style markers. Most phrases people call “AI writing” don’t affect engagement at all.
Focus on reader value instead of avoiding specific words or punctuation marks.
Rethink Your Conclusions
Generic conclusion blocks turn readers away. Instead of obvious structure signals, blend your wrap-up into your analysis. Add new value rather than just summarizing.
Writecream‘s AI agent analyzes how top-ranking pages handle conclusions. This helps you see what actually works instead of following generic formulas.
Use Style Controls in Your AI Tools
Choose AI platforms that let you customize output style. Writecream offers detailed style settings to avoid repetitive patterns while maintaining your brand voice.
Set parameters to limit overused constructions. Generate multiple variations to find the best approach.
The Real AI Writing Problem
The issue isn’t specific phrases or punctuation marks. Bad AI writing happens when tools generate content without understanding what already ranks for your keywords.
Generic AI creates articles in a vacuum. It doesn’t know what your competitors do well or what readers expect from your topic.
Smart content teams use AI that analyzes search results first. They understand winning patterns before generating new content.
Moving Forward With AI Content
AI will keep playing a bigger role in content creation. But success comes from using it strategically, not avoiding it completely.
Write for your readers first. Choose tools that give you control over style and structure. Edit everything with engagement in mind.
The data shows most AI writing fears are overblown. Focus on creating valuable content that serves your audience.
Ready to see how your AI-generated content stacks up against what’s actually ranking in search results?
