Remove AI Writing Tells From Drafts

You can spot it in one line: "In today's fast-paced world," "It's not just a tool, it's a partner," three adjectives in a row, and a closing line that restates the whole post. Readers stop trusting the text when they see these patterns. A generic "make it sound human" pass usually swaps one stock phrase for another or makes up details to sound concrete. This skill finds each pattern by name, cuts it, and leaves your meaning and facts alone.

What it does

• Finds and fixes over a dozen named patterns: stock vocabulary, "serves as" phrasing, "it's not X, it's Y" framing, trailing "-ing" clauses that tack on unearned conclusions, stacked hedges, reflexive lists of three, summary endings, and formatting habits
• Cuts before it rewrites, so the fix doesn't bring in a new cliché
• Never makes up numbers or examples to sound specific. It marks gaps as [FILL] placeholders for you to complete
• Leaves legitimate uses alone, such as domain terms and a single deliberate hedge, and tells you what it kept on purpose
• Audit-only mode explains why a draft reads as AI without rewriting it
• Writes new copy without these patterns and checks its own output before delivering

What you get

The cleaned text, plus a short change log grouped by pattern that shows each original fragment and what replaced it. It also lists any placeholders you need to fill in.

Best for

Marketers, founders, and writers who draft with AI and publish under their own name: LinkedIn posts, emails, landing pages, docs, and newsletters.