From the Conch team
Why AI phrases keep getting flagged — and what to do about them
ChatGPT, Claude, and Gemini all learned to write from the same internet. That is why they default to the same handful of phrases: "delve into", "it is important to note", "in the realm of." These words are not wrong — they are just predictable. Turnitin's AI detector, GPTZero, and Winston AI all measure phrase-level predictability as one of their primary signals. The more your essay relies on these verbal habits, the higher your AI detection score climbs. This tool flags every phrase from our manually curated list of 80+ patterns so you can swap them before submitting. The replacements are not cosmetic — they make your writing sound like it came from a person with a point of view.
Conch Tip: Cluster replacements. If you have 5+ flagged phrases in one paragraph, rewrite the entire paragraph. Spotfixing one phrase in a sea of AI patterns rarely changes detection scores significantly.