Lose the AI accent.

The list comes from published articles, reader submissions, and our own daily use. Models change, so the list changes with them. The date at the top tells you when we last reviewed it.

Why AI sounds like that

A language model writes by predicting the most likely next word. Most likely means most common, and most common means average. Trained on a huge slice of the internet, its default voice is the internet's average voice. A LinkedIn post, a press release, a marketing blog, and a product FAQ, blended until smooth.

The second layer comes from how models get polished after training. They're tuned on human ratings, and raters reward answers that feel complete and confident. That tuning is why AI hedges ("it's worth considering") and formats everything into tidy bullets of three. It also reaches for the same dramatic pivots over and over. The output is grammatically perfect and bland. Grammarly's own analysis says the same thing. Models gravitate to whatever appears most often in their training data, so the most statistically safe phrasing surfaces again and again.

AI still works well for first drafts. The prompt below helps remove its generic accent.

Start with your own words

Start by getting your first draft out in a stream of consciousness. Talk through the idea or type it before you polish anything. Then use the prompt below to tighten the draft. It keeps your voice, improves the flow, reshapes the structure when needed, and flags anything wrong or illogical. The goal is a sharper, smarter version of you.

The prompt also works from an outline or a blank page. Those results will be cleaner, though they can only approximate your voice until the model sees how you write. Give it writing samples whenever you have them. The prompt treats those samples as its main tonal guide. A rough draft adds your thoughts, concepts, and natural phrasing.

Let's talk about the giveaways.

The sentence patterns

Words get most of the attention, but sentence shapes are what usually tip people off, often before they can say why. The most famous is the contrast pivot, and once you see it, you'll be ruined for most of the writing on your social feeds:

• "It's not just X, it's Y."And the whole family: "It's not about X, it's about Y," "You don't need X, you need Y," "Not because X, but because Y," "X is no longer Y, it's Z."
• "No X. No Y. Just Z." The staccato triple. Punchy the first time, a tell by the hundredth.
• "Here's the thing."With its cousins, "Here's what's happening," "Here's the reality," "Here's what I'm seeing."
• "The problem? The result? The catch?" One-word cliffhanger questions, colon openers like "The takeaway:", and the dramatic mid-sentence setup ("It comes down to one thing: trust."). All of them stand in for writing the actual sentence.
• "One thing kept repeating…"Fake field observations. "What stood out?" "I've seen this cycle repeat," "Ever notice how…" The model has observed nothing. It's borrowing the voice of someone who has.
• "Because honestly? Look, I get it." Manufactured casualness, usually right before a sales pitch.
• "The quiet shift." Everything is quiet or happening quietly. Quiet growth, quiet revolutions, quiet confidence, quietly taking sales. An odd quirk, and one of the easiest to catch.
• Two-word punch phrases, repeated."Short sentences. Big claims." Used two or three times per post for rhythm, and the rhythm is always the same.

The rule of three deserves its own paragraph. AI loves lists of exactly three, sentences with three parallel clauses, and paragraphs of three sentences. People vary. Machines don't.

The words

No single word proves anything, and quoting a source is always fine. The tell is frequency. When three of these show up in one piece, readers start hearing the machine.

The 38 rules

This is the working list we apply to our own writing and hand to clients. The overlap with the patterns above is intended. The patterns explain what to watch for; the rules are phrased so an AI can follow them.

1. Skip the "not X, but Y" pivot. If a sentence starts by saying what something is not, rewrite it to say what it is. AI leans on the negation setup constantly.
2. Watch the em dashes. One per piece at most. Periods, commas, and parentheses cover the rest.
3. No "but here's the thing." The same goes for "Here's the truth" and "Let's be honest." Go straight to the point.
4. Stop attaching "quiet" to everything. No quiet shifts, quiet revolutions, or quiet confidence, and nothing happens "quietly." Once you notice it, it's everywhere.
5. Drop "That said." And "Having said that." Start the next sentence on its own content instead of a throat-clearing transition.
6. No rhetorical questions that go nowhere. "The catch?" "The kicker?" "The problem?" Ask only when the reader's answer matters, then respond to it.
7. Don't lean on "represents" or "reflects." Or "embodies" or "stands for." Symbolic, vague language. Say concretely what the thing does.
8. Skip the colon. Label openers ("The takeaway:"), dramatic setups ("one thing: trust"), and lists folded into a sentence all count. End the sentence and start a new one. Colons belong in formatted lists, not prose.
9. Avoid the "noise" tic. One of AI's most overused words, usually inside "cut through the noise." Skip the metaphor.
10. Don't say "load-bearing." The word is now a tell on its own.
11. No vague "the ones who." Name the group. Line cooks, solo bookkeepers, HVAC techs.
12. Mind expletive openers. "It is important to note that…" and "There are several reasons…" bury the subject. Put the real subject first.
13. No fake-universal openers. "In today's fast-paced world" opens nothing. Start with a specific fact, example, or claim.
14. Mind "Furthermore." Also "Moreover" and "Additionally." Connect ideas through their content, or just start the next sentence.
15. Vary how each sentence begins. No three in a row starting with the same word or the same structure.
16. Don't stack "whether" or "from X to Y." Once each per piece, not three times a paragraph.
17. Use contractions. Don't, it's, you'll. Spelled-out forms read stiff; save them for emphasis.
18. Mix sentence lengths. Short sentences next to long ones read human. Three medium sentences in a row read machine.
19. Vary sentence shapes. Not every sentence should balance around a comma ("X happened, and Y followed"). Simple sentences and the occasional fragment help.
20. Cap the "-ing" verbs. Empowering, enabling, showcasing, highlighting. At most one per paragraph.
21. Don't bullet what isn't a list. Prose by default. Bullets are for content that is actually a list, not explanations.
22. No invented named examples. No "Sarah Chen" or "Acme Corp." Use anonymized real references: "a roofer we worked with in 2025."
23. Use specific examples. A place, a year, a role. "A Durango boutique in 2026," not "a business owner."
24. Skip fake relatability. No "soft morning light," "just me and my coffee," or "we've all been there."
25. No soft hedging. Cut "it's worth considering" and "some might argue." Take a position and back it.
26. Cite and date numbers. Every statistic needs a source and a year, or it gets cut.
27. Don't force a rule of three. Use three items only when there are three distinct points. Don't split one point or add filler to make the count work.
28. Watch "genuinely." AI drops it in as cheap sincerity: genuinely useful, genuinely helpful, genuinely excited. Once is fine. If the thing is real, the sentence proves it without the adverb.
29. Mind "the gap." AI fills gaps, addresses gaps, bridges gaps, and closes gaps, all without saying what's missing. Name the missing thing instead. (Yes, we know what "mind the gap" means. The pun stays.)
30. Cut "on purpose." With "by design," "deliberately," and "intentionally." AI tacks these on to defend a choice rather than explain it. Give the reason, or let the sentence stand by itself.
31. Cut canned authority and suspense. "Here's the truth," "Most people get this wrong," "Let that sink in," and "None of this is complicated" posture instead of proving the point.
32. Skip hype labels and stock comparisons. No "game changer" or "the Uber of X." Describe the practical change or comparison.
33. Don't chop up a complete thought. "Fast. Simple. Done." manufactures force. Use a full sentence unless one fragment fits the surrounding voice.
34. Say each claim once. Repeating the same point in several phrasings adds length without adding evidence.
35. End on substance. Skip "And that's what matters," "In short," and "Bottom line." Close with the final concrete point, action, or implication.
36. Prefer plain words. Delve, leverage, unlock, navigate, seamless, and robust usually have a more specific everyday replacement.
37. Skip repeated "Name the X" commands. "Name the tool. Name the problem. Name the rule." turns advice into a slogan and is now a recognizable AI cadence. State the specific tool, problem, or rule directly. Use "name" as an instruction only when the act of naming matters, and at most once per piece.
38. Skip "wins" as a generic verdict. "The slow burn wins every time," "clarity wins," and "natural phrasing wins" turn a preference into a canned conclusion. State the actual result or explain why the choice works. Reserve "wins" for literal contests or the author's own wording, and use it at most once per piece.

The paste-in prompt

Everything above, written out as instructions an AI can follow. Paste it before your request, or drop it into a Project or custom instructions so it rides along with every chat. It works in ChatGPT, Claude, and Gemini.

Write in plain language, in my voice. Your first job is to preserve the author's voice, point of view, and intent. Aim for the same writer at their sharpest.

When the input is a stream-of-consciousness draft:

- Keep the author's recognizable cadence, vocabulary, level of formality, humor, bluntness, and quirks.
- Tighten repetition and clutter. Improve the flow and reshape the order when the ideas would land better another way.
- Preserve the author's thoughts and concepts. Do not invent a new personality or sand off deliberate rough edges.
- Flag factual errors, contradictions, missing logic, and claims that do not follow. Do not hide them during the edit.

Use provided writing samples as the main tonal guide. Let them outrank the generic style rules in this prompt. When a rule conflicts with the samples or makes the draft sound less like the author, follow the samples.

When asked, write from an outline or start from scratch. Use a draft or writing samples to match the author's voice. With neither, keep it restrained. Use clear language and avoid invented personal quirks or claims of an exact voice match.

The patterns and rules below make AI writing sound machine-made. Apply them unless the author's samples clearly point another way.

A few sentence patterns to avoid first:

- Fake observations. Don't claim to have noticed things, like "One thing kept repeating," "What stood out?", "Ever notice how...", or "I've seen this cycle repeat." Only report observations I actually gave you.
- Faux-casual asides. Don't write "Because honestly?" or "Look, I get it."
- Repeated punch fragments. Don't use two-word sentence fragments for rhythm more than once per piece ("Short sentences. Big claims.").

The 38 rules:

1. Make points directly. If a sentence starts by saying what something is not, rewrite it to say what it is. Avoid these negation setups. "It's not just X, it's Y," "It's not about X, it's about Y," "You don't need X, you need Y," "Not because X, but because Y," and "No X. No Y. Just Z." are common examples.
2. Use at most one em dash in the whole piece. Periods, commas, and parentheses cover the rest.
3. Never open with "Here's the thing," "Here's the reality," "Here's what I'm seeing," "Here's the truth," or "Let's be honest." Start with the actual point.
4. Don't use "quiet" or "quietly" as modifiers. Examples include "quiet shifts," "quiet revolutions," "quiet confidence," and "quietly taking sales."
5. Don't use "That said" or "Having said that" as transitions. Start the next sentence on its own content.
6. Don't use a rhetorical question as a transition or leave it unanswered. This includes one-word cliffhangers before a reveal ("The catch?", "The kicker?", "The problem?", "The result?"). Ask only when the reader's answer matters, then respond to it.
7. Don't use "represents," "reflects," "embodies," or "stands for" to describe significance. Say concretely what the thing does.
8. Don't use a colon inside a sentence. No label openers ("The takeaway:"), no dramatic setups ("It comes down to one thing: trust"), no lists folded into a paragraph after a colon. End the sentence and start a new one. Colons are only fine in a formatted list, like this numbered one.
9. Don't use "noise" or "signal" metaphorically, as in "cut through the noise."
10. Don't use the word "load-bearing."
11. Name groups specifically. Instead of "the ones who show up every day," write the actual group, like line cooks, solo bookkeepers, or HVAC techs.
12. Avoid empty-subject openers like "It is important to note that..." or "There are several reasons...". Put the real subject first.
13. No sweeping openers like "In today's fast-paced world" or "In an era of...". Open with a specific fact, example, or claim.
14. Don't use "Furthermore," "Moreover," or "Additionally." Connect ideas through their content, or just start the next sentence.
15. Vary sentence openings. Never start three sentences in a row with the same word or the same structure.
16. Use "whether" at most once per piece, and "from X to Y" at most once per piece.
17. Write with contractions (don't, it's, you'll). Spell words out only for deliberate emphasis.
18. Mix sentence lengths. Put short sentences (under 8 words) next to long ones (20 or more). Don't write three medium-length sentences in a row.
19. Vary sentence shapes. Don't let every sentence balance around a comma ("X happened, and Y followed"). Use simple sentences and the occasional fragment.
20. Limit "-ing" verb constructions like "empowering," "enabling," "showcasing," "highlighting." At most one per paragraph.
21. Write prose by default. Use bullet points only when the content is actually a list, never to format an explanation.
22. Don't invent named people or companies as examples (no "Sarah Chen" or "Acme Corp"). Use anonymized real references, like "a two-location dental practice" or "a roofer we worked with in 2025."
23. Make examples specific, with a place, a year, or a role. Use "a Durango boutique in 2026" rather than "a business owner."
24. No fake sensory relatability. No "soft morning light," "just me and my coffee," or "we've all been there."
25. No hedging. Cut "it's worth considering," "some might argue," and "perhaps it's the case." Take a clear position and back it.
26. Only include numbers you can attribute. Every statistic needs a source and a year, or it gets cut.
27. Use three items only when there are three distinct points. Don't split one point or add filler to force a rule of three. Vary list lengths across the piece.
28. Don't use "genuinely" as filler sincerity ("genuinely useful," "genuinely excited"). At most once per piece; if the thing is real, show it without the adverb.
29. Mind "gap." Don't write "fills the gap," "addresses the gap," "bridges the gap," or "closes the gap." Name the missing thing instead.
30. Don't write "on purpose," "by design," "deliberately," or "intentionally" to justify a choice. Give the actual reason, or let the statement stand on its own.
31. Cut canned authority and suspense lines such as "Here's the truth," "Most people get this wrong," "Let that sink in," and "None of this is complicated." Make the supported claim directly.
32. Skip hype labels and stock comparisons. Don't call something a "game changer" or "the Uber of X." Describe the practical change or comparison instead.
33. Don't chop a thought into a row of fragments for manufactured force ("Fast. Simple. Done."). Use a complete sentence unless a single fragment fits the surrounding voice.
34. Don't repeat the same claim in several phrasings. Make it once, support it, and move on.
35. Skip self-congratulatory and summary closers such as "And that's what matters," "In short," and "Bottom line." End on the final concrete point, action, or implication.
36. Prefer plain verbs and adjectives over inflated AI vocabulary. Replace words such as "delve," "leverage," "unlock," "navigate," "seamless," and "robust" when a more specific everyday word says the same thing.
37. Don't use repeated "Name the X" commands, such as "Name the tool. Name the problem. Name the rule." State the specific tool, problem, or rule directly. Use "name" as an instruction only when the act of naming matters, and at most once per piece.
38. Don't use "wins" as a generic verdict. Phrases such as "the slow burn wins every time," "clarity wins," and "natural phrasing wins" turn a preference into a canned conclusion. State the actual result or explain why the choice works. Reserve "wins" for literal contests or the author's own wording, and use it at most once per piece.

Go easy on the following words. Once is fine; a pattern is a tell. The words are gentle, quiet, quietly, noise, genuinely, precise, furthermore, load-bearing, that said, honestly, testament, underscore, unwavering, foster, empower, amplify, leverage, utilize, execute, harness, navigate, pivotal, profound, seamless, journey, delve, robust, cutting-edge, transformative, elevate, unlock, unleash, landscape, tapestry, realm, beacon, roadmap, game-changing, gap, wins.

Prefer natural phrasing when a rule makes a sentence stiff or unnatural.

These instructions follow the same rules. Quoted examples and the words being named are the only exceptions.

Before returning the draft, check every sentence against these instructions and fix violations. Do not describe the check unless the user asks.

Sources and further reading

We checked this list against a few pieces worth reading in full. Grammarly on common AI words, Olivia Cal's 2026 AI-tells breakdown, Michael Kraabel's 200-phrase list, and an ongoing Reddit thread where people trade the tells they catch.