Posts that don't read as AI
Every other LinkedIn AI tool produces output you can identify as AI within two sentences. Postbrander has a 110-word banned list, 44+ banned phrases, and 30+ banned sentence patterns baked into every generation. Plus a live score on every post so you can see what slipped through.
The fatal pattern: negative parallelism
If your AI tool produces sentences like "This isn't a productivity hack. This is a mindset shift," it's flagging itself as AI in three seconds. The pattern is so over-represented in AI output that readers register the rhythm before they finish reading.
We block 14 variants of this pattern in every generation, including the sneaky disguised ones ("While X might seem right, Y is actually…", "Sure, X works. But Y is where the real…"). When the model tries to use one, we either rewrite it on the way out or flag it on the score badge.
"It's not about the prompt. It's about the context. You don't need better tools. You need better thinking. Stop chasing tactics. Start building systems."
"Most teams overthink prompts and underthink the context they give. Better tools without better thinking is just faster slop. Build the system once, the tactics fall out of it."
The banned word list
Every word below is blocked in every Postbrander generation. They're the statistical fingerprint of LLM output, picked up from peer-reviewed AI-detection research and from Wikipedia's own AI-style field guide.
delve, realm, harness, unlock, tapestry, paradigm, cutting-edge, revolutionize, revolutionise, intricate, intricacies, showcasing, showcase, crucial, pivotal, surpass, meticulously, meticulous, vibrant, unparalleled, underscore, underscores, leverage, leveraging, leverages, synergy, synergies, synergize, synergise, innovative, game-changer, game-changing, testament, commendable, emphasize, emphasise, emphasizes, emphasises, boast, boasts, groundbreaking, foster, fosters, enhance, enhances, enhanced, holistic, garner, garners, accentuate, accentuates, pioneering, trailblazing, unleash, unleashes, versatile, transformative, redefine, redefines, seamless, seamlessly, optimize, optimise, optimized, optimised, scalable, robust, breakthrough, empower, empowers, empowered, streamline, streamlines, streamlined, frictionless, elevate, elevates, adaptive, effortless, effortlessly, data-driven, insightful, proactive, mission-critical, visionary, disruptive, reimagine, reimagines, unprecedented, intuitive, leading-edge, democratize, democratise, accelerate, accelerates, state-of-the-art, immersive, predictive, proprietary, plug-and-play, turnkey, future-proof, paradigm-shifting, supercharge, supercharges, enduring, interplay, captivate, captivates, captivating
Phrases that mark you as AI
Dead transitions like "Furthermore" and "Moreover". Hype like "game-changer" and "10x your". Engagement bait like "Let that sink in" and "Read that again". Meta commentary like "In this post, I will…".
Every phrase below is blocked in every generation, and the post-processor strips the most common offenders even if the model still slips them in.
- "in today's"
- "in today’s"
- "it's important to note"
- "it’s important to note"
- "it's worth noting"
- "it’s worth noting"
- "in order to"
- "i'd be happy to help"
- "i’d be happy to help"
- "straightforward"
- "let's dive in"
- "let's explore"
- "let's unpack"
- "delve into"
- "at the end of the day"
- "moving forward"
- "to put this in perspective"
- "what makes this particularly interesting"
- "the implications here are"
- "in other words"
- "it goes without saying"
- "here's the part nobody's talking about"
- "what nobody tells you"
- "most people don't realize"
- "most people don't realise"
- "in this article, i will"
- "in this post, i will"
- "furthermore"
- "additionally"
- "moreover"
- "that said,"
- "that being said"
- "with that in mind"
- "it is also worth mentioning"
- "on top of that"
- "let that sink in"
- "read that again"
- "this changes everything"
- "are you paying attention?"
- "you're not ready for this"
- "you’re not ready for this"
- "10x your"
- "game-changer"
- "cutting-edge"
Every post gets a score
When you open any draft in Postbrander, you see an Anti-AI score (0-100) computed live against the same rules. Below 75 and you see exactly which rules fired and where in the text. Above 90 and you know it's clean enough to publish without thinking twice.
We don't auto-rewrite around banned words because that usually changes the meaning. You decide whether the word is load-bearing or a tic. Most of the time you'll find it's a tic.
We never send an AI reply verbatim
When someone comments on your post, Postbrander drafts a reply in your voice. It never posts that draft as-is. You have to change it in your own words before Send unlocks, and we never bulk-send replies.
This is deliberate. Richard van der Blom's 2026 research found 80% of the comments creators get in the first five minutes are AI-written, and that is exactly the behaviour LinkedIn's Authenticity Update set out to penalise. A draft saves you the blank page. The words that actually go out are yours.
Frequently asked questions
Why does AI writing sound like AI?+
Three reasons. First, AI models are trained on persuasive writing (TED talks, op-eds, marketing copy) that overuses certain patterns: negative parallelisms, rule-of-three lists, puffery. Second, AI defaults to hedge words and copulative avoidance ("serves as", "stands as", "represents") instead of plain verbs. Third, AI has no judgement about which words are statistically over-represented in its own output, so it reaches for the same 90 words again and again.
Can readers actually tell?+
Yes, and faster than you'd think. Peer-reviewed AI detection research and informal LinkedIn surveys both find that readers spot AI content within the first two sentences. The biggest single tell is the negative parallelism ("This isn't X. This is Y."). Once a reader notices one, the rest of the post is dead.
What's different about Postbrander's anti-AI approach?+
Three things. We bake a 90+ word banned list and 30+ banned-pattern list into the model's instructions for every generation, not just as a vague "sound human" prompt. We post-process the output to strip em dashes and dead transitions before you ever see them. We score every post out of 100 against the same rules and surface the score in the UI so you can see what slipped through.
Is the list public?+
Yes, the full banned-word and banned-phrase lists are on this page. The rules came from a personal style guide James (the founder) built over a year of editing his own LinkedIn output, then validated against academic AI-detection research and Wikipedia's own AI-style field guide.
Do I lose anything by using these rules?+
You lose the option to write a post that sounds like every other LinkedIn AI tool. That's the point. The rules cost you nothing real, banned vocabulary like "delve" and "leverage" was never load-bearing. Negative parallelisms are a tic, not an idea. Stripping them out actually makes your point clearer.