THE NOISE

The AI Tool That Writes Your LinkedIn Posts Also Writes Everyone Else's

2026-09-20

Write in your own voice is one of the most common pitches on the Radar page right now, and it's an appealing one, since nobody wants their content to read like an obvious AI draft. The trouble is structural: most of these tools are trained on the same corpus of what a high-performing LinkedIn post looks like, the hook line, the short paragraphs, the reflective question at the end.

That means instead of sounding like the specific person using it, the output tends to sound like the genre the tool learned from. Enough people using the same tool ends up producing a wave of posts that all share the same structural tics, which is a genuinely recognizable pattern once you start noticing it in your own feed, not proof of anything, but hard to unsee.

Scroll any professional feed for ten minutes right now and the pattern is easy to spot once you know what you're looking for: a one-line hook designed to stop the scroll, a personal anecdote compressed into two sentences, a pivot to a lesson, and a closing question inviting engagement. That structure isn't wrong, exactly, it's just so common now that it's started to signal AI-assisted rather than authentic, even when a real person wrote every word by hand.

What actually makes writing sound like a specific person is the opposite of what these tools are built to produce: a real, dated detail from someone's actual experience, a name, a number, a specific moment that couldn't have come from anyone else. That's precisely the thing a generic drafting tool is least equipped to invent convincingly, because it has no access to it.

There's a useful test for this worth running on your own drafts, whether AI-assisted or not: could this exact post have been written about a different person's career, with only the name changed, and still make sense? If the answer is yes, the post is genre, not voice, regardless of who or what wrote it. A post that only makes sense as coming from you specifically has passed the test that actually matters.

It's also exactly why anything published on this site gets a specific fact-check pass before it goes up, rather than being left as smooth, generic-sounding prose. The sentence that makes someone stop scrolling is usually the one detail only the actual writer could have supplied.

The practical use of these tools, then, is for structure, not voice. Let AI handle the shape of a post: where the hook goes, how long each paragraph should run, where the call to action sits. The one true, specific detail that makes it sound like you and not like everyone else using the same tool should stay yours to write, and it's worth writing that one sentence first, before anything else, so the tool has something real to build structure around instead of inventing a generic anecdote to fill the gap.

There's a related risk worth naming for anyone building a personal brand around thought leadership specifically: audiences are getting faster at spotting the pattern, which means the cost of sounding generic is rising even as the tools that produce generic content keep getting better at the surface-level polish. The gap between looking professional and sounding like a specific person is where the actual differentiation lives now, and it's the one thing these tools structurally can't manufacture for you. The tools aren't going away, and they don't need to. What needs to change is treating them as a first draft of shape rather than a finished draft of substance, which is a small shift in how you use the tool but a large one in how the result actually reads to someone scrolling past it.