# AI for personal brand content and thought leadership, in your voice · ScriptGrain

> Personal brand content only works if it sounds like the person. How founders, executives and consultants use AI to write thought leadership in their own measured voice, and score every draft before it goes out.

Canonical: https://scriptgrain.com/us/ai-for-personal-brand-content

# AI for personal brand content: thought leadership that still sounds like you

*By Jack Stovell [X](https://x.com/StovBuilds) · [Instagram](https://www.instagram.com/stovbuilds) · [LinkedIn](https://www.linkedin.com/in/jackstovell/) · published 2026-10-09*

AI can draft personal brand content, and it only works when the drafts sound like the person. That's why ScriptGrain measures the voice from writing you already have and scores every draft against it, so what goes out under your name still reads like you wrote it.

Here's the check. Paste something unmistakably yours (an email, a talk written up, an old newsletter) and a recent LinkedIn post into the free check below. The tool renders after it, and it'll score the post against your own writing so you can see the gap before anyone else does.

## Why personal brand content is a voice problem

People follow a person, not a company page. That's the whole premise of a personal brand, and it's also the whole risk. A generic draft under a real name reads flat, and it costs trust, because the reader thinks they know how you sound and a mismatch registers as something being off even if they can't name it.

This is where a general-purpose AI writer falls short. They generate fluent, structurally sound content that could have come from anyone. Fine for a product description. Not fine when your name is the byline and your reputation is the product. So the useful question is whether the draft sounds like you specifically, on a Tuesday, tired, with your usual habits intact.

A comms team's drafted posts can pull an executive's voice away from what they'd normally write, even when the team means well and the content is accurate. Effort is rarely the problem. Voice is measurable and specific, and a drafting process that never measures it has no way to notice the drift.

## What a thought leadership writer does, and what AI can take on

A good thought leadership writer doesn't invent your opinions. They shape the ones you already hold into something readable. The point of view stays yours; so does the judgment about what's worth saying and what isn't. That part doesn't get automated, and it shouldn't.

Drafting, format and consistency are a different matter. Turning a rough idea into a LinkedIn post with the right shape, keeping a newsletter's rhythm consistent week to week, matching the sentence patterns you'd actually use rather than generic ones: that's mechanical work dressed up as creative work, and it's exactly where AI earns its place.

ScriptGrain's Voice Brain keeps track of what you've already published, so a new piece builds on what you've said rather than repeating it. It runs a novelty check and can suggest what to write next, and it's free. This matters more than it sounds: thought leadership that just says the same three things in different outfits stops looking like leadership and starts looking like a loop.

## Measure the voice once, from writing you already have

You don't train this on nothing. ScriptGrain builds one voice profile from your own writing: emails, memos, posts, newsletters, talks written up, whatever exists already. It measures 45 attributes across eight groups, everything from sentence rhythm to how you use contractions to how you order a clause when you're building toward a point. Three or more pieces and about 3,000 words is what we recommend.

Once the profile exists, generation covers 16 content types, each with a sensible default length: a LinkedIn post defaults to 220 words, a newsletter to 500, a long-form article to 800, an X thread to about five numbered posts, plus email and more. You're not fighting the tool to hit a length that suits the format. It already knows.

And the profile isn't static. It connects to ChatGPT and Claude over MCP, so drafts written there get scored too, not just ones generated inside ScriptGrain. Wherever the draft comes from, it gets measured the same way.

## Score every draft before it carries your name

Every draft gets scored 0 to 1 against your profile. The scoring looks at two layers. Code counts the concrete stuff: sentence length and variance, contractions, commas, exclamation marks, semicolons, brackets, questions, ellipses, pronoun mix, article balance, word length, your own signature phrases. A small model judges the harder-to-count things: formality, humor, expressiveness, confidence versus hedging, opening style, argument structure, rhythm, sentence complexity, clause ordering, specificity. A weighted comparison runs through a fixed calibration to produce the final number.

Against your own profile, on-voice drafts typically land between 0.85 and 0.95. Partial matches sit at 0.5 to 0.8, close but with something named as off. A clearly different voice scores under 0.45. The score also names the difference, so you know what to fix rather than guessing.

Edits you save feed the profile's learning. Ghostwritten drafts work the same way: a founder's or executive's own writing becomes the reference, the ghostwritten draft gets scored against it, and the habits that slipped get named directly.

## Founders, executives and consultants

Founders write to be believed before they're big enough to be checked. If that's you, [the founders page](https://scriptgrain.com/us/for-founders) covers how the profile fits a founder's specific mix of writing.

Executives often work with a ghostwriter or a comms team, and a ghostwriter can [score drafts against the principal's writing](https://scriptgrain.com/us/founder-voice) rather than guessing at tone from memory. [The executives page](https://scriptgrain.com/us/for-executives) covers retainers and what changes when someone else is doing the drafting.

Consultants live and die by a distinctive point of view showing up consistently across formats. [The consultants page](https://scriptgrain.com/us/for-consultants) covers that case directly, and [LinkedIn posts in your own voice](https://scriptgrain.com/us/linkedin-posts-in-your-voice) covers how the platform's format can flatten your actual voice if you're not watching for it.

## Where AI should not write for you

Some things stay off limits regardless of how good the match score is. AI shouldn't write opinions you don't hold, even fluent ones. It shouldn't invent experiences you didn't have, even plausible ones. And it shouldn't produce anything you wouldn't say out loud, in person, to the people reading it.

For one person, two plans fit. Writer is $17 a month for one profile. Operator is $40 a month for three profiles, built for someone who writes for themselves and also for a company or a client. Neither plan writes your opinions for you. They just stop the drafting from sounding like nobody.

## Questions

### Does AI-written content ever really sound like me?

It can, if the draft is measured against your own writing rather than a generic tone setting. ScriptGrain builds a profile from your actual samples: emails, posts, newsletters, whatever you've already written. Drafts get scored against that profile, so "sounding like you" becomes something checked, not just hoped for.

### How much writing do I need to give it?

A profile can be built from a single sample of 50 words or more, but three or more pieces and about 3,000 words is what we recommend. Choose writing you did yourself, without an editor or an AI in between, so the profile measures you.

### Can a ghostwriter use this too?

Yes. The founder's or executive's own writing becomes the reference profile, and the ghostwritten draft gets scored against it. The scoring names which habits slipped, so edits are targeted rather than a full rewrite from scratch.

### What counts as a bad voice match?

Under 0.45 against your profile is a clearly different voice. Partial matches land between 0.5 and 0.8, close but with a named difference. On-voice drafts typically score 0.85 to 0.95.

### Will this write my opinions for me?

No. The point of view, the judgment about what's worth saying, and any experience you're drawing on all stay yours. AI can take on drafting, format and consistency. It shouldn't be asked to hold opinions or claim experiences on your behalf.