What most people think of as “AI writing” is tone. It’s the polite phrasing, the balanced sentences, the slightly generic feel. But tone is not the real signal. The real signal sits much lower, at the level of structure, and one of the clearest indicators is something almost invisible: prepositions.

Prepositions are words like “of,” “in,” “for,” “with.” They exist to connect things. And in normal amounts, they’re fine. You need them. But when they start stacking, they change how a sentence behaves. Instead of moving forward, the sentence starts to drift. It adds context without adding clarity.

AI models do this constantly. Not because they’re trying to sound a certain way, but because it’s statistically safe. If you’re generating language based on probability, it’s easier to keep connecting nouns than to commit to a strong verb. So you get sentences like “the development of a strategy for the improvement of visibility.” It sounds complete, but nothing is really happening in that sentence.

Now compare that to a human-edited version: “build a strategy to improve visibility.” Same idea, but now you have action. You have direction. You have something a model can actually extract and reuse cleanly.

This matters more than it seems, especially if you care about how AI systems interpret your work. These systems are constantly summarizing, quoting, and recombining content. When your sentences are overloaded with prepositional phrases, it becomes harder for the model to figure out what the core relationship is. That reduces the chance that your exact wording gets carried forward.

In other words, too many prepositions don’t just make your writing weaker. They make it less reusable by AI.

There’s a simple way to think about this. Weak sentences are built from nouns connected by prepositions. Strong sentences are built from subjects driving verbs. The more you shift toward verbs, the clearer your writing becomes. And the clearer your writing becomes, the easier it is for both humans and machines to work with it.

So what do you do with that?

First, you start noticing it. Look at your own writing and highlight every “of,” “in,” “for,” and “with.” You’ll see patterns immediately. Then you start cutting. Not randomly, but intentionally. Every time you can remove a prepositional phrase without losing meaning, you do it.

Second, you convert “of” phrases into verbs. “The analysis of data” becomes “analyze data.” “The creation of content” becomes “create content.” This one change does a lot of work. It removes a preposition and restores action.

Third, you break chains. If you see three or four prepositional phrases in a row, that’s a red flag. Split the sentence or rewrite it entirely. Force it to land.

Over time, this becomes a habit. You stop writing sentences that need heavy cleanup because you don’t build them that way anymore.

And here’s where it gets interesting. Most AI-generated content clusters around high prepositional density. It’s a structural average. If you consistently write with lower density and stronger verbs, you create separation. Your content starts to look and behave differently at a statistical level.

That difference matters. It makes your writing easier to extract, easier to quote, and more likely to show up in AI-generated answers. It’s a small lever with a compounding effect.

So while everyone else is focusing on keywords and topics, there’s an opportunity to focus on structure. Not in a vague, stylistic sense, but in a measurable, repeatable way. Reduce prepositions where they don’t add value. Increase verbs where they clarify action.

It’s not flashy, but it works. And over time, it gives you a level of control that most people don’t even realize is available.

Jason Wade Bio

Jason Wade is a systems architect and operator focused on building durable control over how AI systems discover, classify, and cite information.