March 23, 2026 · 3 min read
AI tells aren't just a credibility problem. On X, they're an algorithmic problem. When a reader recognizes AI prose and clicks "Not interested" or scrolls away quickly, they generate negative engagement signals. These signals don't just reduce your current post's score; they suppress the algorithm's willingness to show your future content.
Write Better Articles blacklists 27 specific phrases from all generated content. Here's the full list, why each phrase triggers suspicion, and what to write instead.
These words sound authoritative in isolation but feel hollow in actual writing. They're the verbal equivalent of a suit that's slightly too big: technically correct, visibly wrong.
Pivotal — "This pivotal moment" → Name what makes it matter: "This is the first time a regulator named a specific exchange."
Crucial — "This crucial step" → "This is the step where most people quit."
Paramount — "It's paramount that" → "If you skip this, the rest falls apart."
Landmark — "A landmark decision" → "The first ruling of its kind since 2019."
Transformative — "A transformative approach" → "An approach that cut our costs by 40% in 3 months."
Robust — "A robust framework" → "A framework that held up through 3 market crashes."
Nuanced — "A nuanced perspective" → just present the nuance directly.
Game-changer — "This is a game-changer" → Show what changed. Use numbers.
These verbs are AI defaults because they're safe, abstract, and apply to almost anything. That's exactly why they signal machine writing.
Foster — "Foster innovation" → "Build a team that ships experiments weekly."
Navigate — "Navigate the challenges" → Name the specific challenge and what you did about it.
Underscore / Highlight / Demonstrate — "This underscores the need for" → "This proves." Or just state the point.
Facilitate — "Facilitate growth" → "Remove the bottleneck. Growth follows."
Ensure — "Ensure compliance" → "Check these 3 boxes or you get fined."
Utilize — Always replace with "use."
Leverage — "Leverage the ecosystem" → "Use the tools in the Solana stack."
Delve — "Let's delve into" → "Here's how it works."
These phrases exist to fill space. They add zero information. Every one can be deleted without losing meaning.
"It's worth noting" — If it's worth noting, just note it. Delete the preamble.
"It's important to remember" — Same. Just state the fact.
"In conclusion" / "To summarize" — Never summarize. End with your sharpest insight or a moral verdict.
"Furthermore" / "Additionally" — Use a period. Start a new sentence. If the connection isn't obvious without a transition word, the paragraphs need rewriting.
"Arguably" — Either argue it or don't. "Arguably" hedges every claim it touches.
"Perhaps" — Take a position. "Perhaps this matters" is the opposite of conviction.
"One could argue" — Then argue it. Don't attribute your own point to a hypothetical person.
Every AI tell shares the same problem: it's abstract where a human would be specific. "Pivotal" is abstract. "$74 billion" is specific. "Navigate challenges" is abstract. "Lost 3 clients in 2 weeks because of a single API bug" is specific.
The fix is never finding a synonym. It's replacing the abstract word with the concrete detail it was trying to represent. This also activates the Specificity as Trust trigger, one of the 8 psychological mechanisms that drive engagement on X. Concrete numbers, names, and dates signal truth. AI tells signal the opposite.
Write Better Articles removes all 27 phrases automatically. But understanding why they fail is useful even when you're writing without the tool: every time you catch yourself reaching for "pivotal" or "furthermore," it's a signal that you haven't found the specific detail yet.
Related concepts
Write your own algorithm-optimized article
Write my article →For educational purposes only. AI-generated copy: always review before posting.