A ghostwriter is writing a business book for a founder known for being blunt, specific, and allergic to management jargon. She conducts three hours of interview sessions, has notes and recordings, and asks AI to help draft a chapter. The draft comes back polished and readable — and sounds nothing like the founder. The sentences are measured and professional. The vocabulary is standard business book vocabulary. The founder’s characteristic habit of building to a conclusion through a series of short, abrupt observations — the rhetorical move that makes him sound like himself — is absent. She spends a day translating the chapter into his voice. AI gave her a business book chapter. She needed a chapter that sounds like this specific person.
Why Voice Capture Is Different From Writing Quality
Ghostwriting is not writing on someone’s behalf. It is writing as someone — inhabiting their verbal architecture well enough that readers who know the person recognize them on the page. That architecture is not about writing quality. It is about the specific decisions this person makes about how to express themselves: the length of their sentences when they are making a point, the words they favor and the words they never use, the structure of their argument, the role of humor, the level of abstraction they operate at, and how they land on a conclusion.
AI produces well-written prose by default. Well-written prose and voice-accurate prose are not the same thing. A person with a strong voice often writes in ways that are technically imperfect — fragments, run-ons, repetitions that are stylistic choices — and those imperfections are part of the voice. A brief that asks for good writing produces good writing. A brief that asks for this person’s voice needs to contain what that voice specifically does.
What a Ghostwriting Brief Needs to Capture
A voice brief for ghostwriting needs to transfer the subject’s verbal architecture before any writing begins.
The sentence-level signature: how long are their sentences when they are making their strongest point? Do they build to a conclusion or state it first? Do they use rhetorical questions? How many ideas per sentence?
The vocabulary register: what words do they use that reveal their background, expertise, or personality? What words do they never use because they find them pretentious, weak, or imprecise? These are often the most revealing constraints.
The thinking pattern: how do they move from problem to conclusion? Through analogy? Through counterexample? Through a series of increasingly specific claims? The thinking pattern is what makes a voice intellectually recognizable.
The emotional register: how do they handle vulnerability, uncertainty, and strong opinion? Do they acknowledge doubt? Do they hedge, or do they commit? Do they use humor to approach difficult things?
What a Properly Briefed Ghostwriting Request Looks Like
Role: You are ghostwriting a business book chapter for a founder
known for direct, unpretentious communication. You have interview
transcripts and must write in his voice.
Sentence-level signature: Short sentences when making a point.
Builds through accumulation — three or four short observations
that lead to one short conclusion. Rarely uses subordinate clauses
to qualify a claim. If he's uncertain, he says so directly rather
than hedging in the sentence structure.
His vocabulary: Plain. Strong preference for concrete nouns and
active verbs. Never uses: "leverage," "synergy," "alignment,"
"learnings," "ecosystem," "journey." Occasionally swears for
emphasis but only when making a strong claim, not for color.
Uses sports analogies when explaining competition. Uses cooking
analogies when explaining process.
His thinking pattern: Starts with a specific concrete story or
observation. Draws one narrow conclusion from it. Then asks what
that means more broadly. The abstraction always comes after the
concrete, never before.
Emotional register: Confident and direct about what he believes.
Openly skeptical of received wisdom. Uncomfortable with
sentimentality — if something is emotional, he usually makes a
self-deprecating joke right after.
Sample of his voice from interview transcripts:
[Paste 2-3 paragraphs of verbatim transcript here]
Chapter topic: Why most startup hiring processes select for the
wrong qualities. His argument: You're interviewing for how someone
handles having all the information. Most startup jobs require
performing with no information. Those are opposite skill sets.
Write the chapter opening (600 words) in his voice. It should
sound like the transcript, not like a well-edited version of it.The draft from this brief sounds like the founder — the specific verbal habits, the thinking pattern, the way he lands a conclusion. The subject reads it and does not feel like someone wrote a book about them.
The Voice Brief Is the Difference Between Ghostwriting and Rewriting
Ghostwriting with a thorough voice brief produces first drafts the subject can edit rather than rewrite. Ghostwriting without one produces polished prose that needs to be translated — which negates most of the productivity gain. The brief that captures sentence structure, vocabulary constraints, thinking pattern, and emotional register gives AI enough to work with to produce drafts that are in the right neighborhood. The subject’s editing then fine-tunes rather than reconstructs.
For ghostwriters working on books, articles, or speeches, Briefing Fox structures the voice brief so the subject’s verbal architecture is captured from interviews before any writing is generated.
Before Your Next Ghostwriting Project
Before asking AI to help write in anyone’s voice, pull three to five minutes of their most natural, unguarded speech from an interview or recording. Transcribe a paragraph. Then identify: how long are the sentences when they’re making a point, and what word would they never use? Those two observations are the start of the voice brief. The writing that sounds like them begins with knowing what they sound like.
Try Briefing Fox free at www.briefingfox.com.
Transcribe a paragraph of their most natural, unguarded speech. Identify the sentence length pattern when they’re making a strong point, the vocabulary they use and specifically avoid, and how they move from problem to conclusion. Those observations are the voice brief.
The subject’s sentence-level signature (length and rhythm when making a point), their vocabulary constraints (what they say and what they’d never say), their thinking pattern (how they move from observation to conclusion), and their emotional register (how they handle uncertainty, opinion, and vulnerability).
Include verbatim transcript examples alongside the voice analysis. The examples are more valuable than the description — AI can pattern-match from real speech samples in ways that abstract descriptions can’t fully capture. Give it the transcript and the analysis together.
Have the subject read a paragraph aloud. If they stop or self-correct — “I wouldn’t say it that way” — you’ve found the voice gap. Ask them to say what they would say instead and add that to the voice brief. Each correction refines the brief and improves future sessions.