How to make AI writing sound like you: the voice profile

A cover letter or a LinkedIn post has one job before any other: to sound like a particular person. How you write is part of your professional identity, and a generic AI tool hands everyone the same fluent paragraph. The voice profile is how FutureDay drafts in your words instead, so what you send under your name still sounds like you.

Why do AI cover letters and posts all sound the same?

Ask a generic AI tool for a cover letter and you get something fluent and almost identical to what everyone else asking it gets. With nothing real about you to work from, it reaches for the safest words it knows: "highly motivated", "results-driven", "a proven track record". Each fits every candidate, which is exactly why none of them describes you. Posts go the same way: thrilled to share, "Does this resonate?", and phrases you would never say out loud.

What does generic AI writing cost you in a job search?

A cover letter or a post exists to make a hiring manager want to meet you. Generic AI writing undoes that on contact. Anyone who reads a lot of applications learns the pattern, and the tell isn't bad grammar. It's the absence of anything a real person would say.

For an experienced professional, real and specific work gets flattened into everyone else's paragraph, and on LinkedIn it happens in public, under your name. The person behind the words disappears, which is the opposite of what the writing was for.

Why doesn't telling AI to "sound like me" work?

Because most people, asked to describe their own voice, freeze. You know your writing when you read it, but you can't hand a stranger, or a language model, a formula for it. And "sound more like me" means nothing to a tool that knows nothing real about you, so it falls back to the same safe tone it gives everyone. Doing it yourself works, but it means rewriting every draft by hand.

How does a voice profile make AI writing sound like you?

It starts in Settings, under "Your voice", with fifteen questions in three sections: Work & career, Personality & energy, and Values & how you operate. You can speak or type each answer, and none of them asks you to describe your writing. They ask about things you would talk about anyway: "What's a place or trip you'd happily talk about for twenty minutes if someone let you?" "What's a principle you won't compromise on, even when it would be easier to?"

You don't need all fifteen. The screen asks for at least one from each section and says a couple from each works better than fifteen in one, because FutureDay looks for the words you use across different topics. A word that turns up in your career answer and again in the trip answer is a real habit; one that appears once is noise. A bar shows how complete the profile is: one answer about your career and one about the trip take it from 0% to 30%, and a third answer in the same section doesn't move it.

You can also paste in something you've already written, an email, a post, a message, and tag it "All me", "AI helped polish it" or "Not sure". The tag travels with the sample, so writing an AI has already polished counts for less; polish smooths out the very habits the profile is looking for.

One answer in each section, or a single sample, is enough to press "Build my voice profile". The result appears on the same screen in plain words, so you can read what it picked up. Dictation filler like "you know" and "umm" is left out of it. "Rebuild my voice profile" runs it again when you add more.

Once it's built, switches under "Where it's used" decide where it goes. The "Content posts" and "Cover letters" switches start on: posts in the Post Writer and the Posting Plan, and the Tailored Cover Letter. The About section in the LinkedIn Optimiser stays off until you switch it on, and so does the top of a Tailored Resume: just the headline and summary, never the sections underneath.

Wherever it's switched on, the profile changes the wording and never the facts. Each writer is told the profile sets only which words you reach for and how plainly you say things, ranks below every other rule, and cannot add a claim, a figure, an employer or a result that isn't already in your material.

How can you tell if a LinkedIn post sounds like AI?

In the Post Writer, a signal check sits under the editor and reads the body of your post as you edit it. It gives a plain verdict, "Reads clean", "Some flags" or "Reads as AI", and flags each thing to fix: a stock phrase to remove, more than two em dashes, or a rhetorical question such as "Sound familiar?" A post with nothing to fix reads "Low AI signal. Strong writing." Open a post from your Posting Plan or Posting Calendar in the Post Writer and the same check reads it.

The same news, written two ways:

Put the first version in the Post Writer's editor and the check says "Some flags", with three things to remove: "thrilled", "passionate about" and the rhetorical question. The second says "Reads clean". The facts are identical; only the wording moved, which is all a voice profile is allowed to change.

The signal check lives in the Post Writer. A Tailored Cover Letter follows its own rules instead, which name phrases like "proven track record" and "fast-paced environment" and tell the writer to avoid them entirely. Either way the draft is yours to change, word by word, and FutureDay never posts or sends it. You do.

What changes once your writing sounds like you?

A cover letter stops reading like it could have come from anyone applying to anything, and starts reading like the specific person a hiring manager might want to meet. A post sounds like the colleague people already know, and in the Post Writer you get a plain verdict and exact fixes, not another layer of AI telling you it's fine. You decide where your voice is used, and nothing goes anywhere until you send or post it yourself.