How to Humanize a Ghostwritten Executive Post
AI exec posts invent opinions the leader never had. Lock one story they told you, the point they will defend, and what they refuse to say. Then match their spoken voice.
Ghostwritten executive posts fail when the draft has a thesis the leader did not choose. ChatGPT supplies the thesis for free: resilience, authenticity, “here is what this quarter taught me,” with a story that never happened.
Your job is to use their story and their limit on what is public. Human Writes only helps the sentences match how they talk.
Hubs: LinkedIn posts and newsletters, speaker bio, disclose AI on client deliverables.
What the executive will reject
| Draft habit | What they say in the review |
|---|---|
| Invented customer quote | “I never said we lost that deal” |
| Hot take they do not believe | “This is not my view” |
| Therapy tone | “I do not post like that” |
| Three lessons | “Pick one” |
| Their staff’s work claimed as theirs | A trust problem, not a style problem |
Story card (they fill or approve)
| Field | Example shape |
|---|---|
| One scene | A meeting, a launch, a mistake, with a time |
| What they decided | The actual call |
| What they want the reader to do | Hire, comment, change a habit, or nothing |
| Words they use | Pull from a transcript |
| Off limits | Unreleased numbers, other people’s performance, legal matters |
Prompt: “Write a 150–200 word first-person post from this card only. Use their vocabulary list. One scene, one point. No ‘in today’s fast-paced world.’ No lesson they did not write on the card.”
Before and after
AI ghost draft:
In today’s ever-changing landscape, I was reminded that leadership is about empowering people. This week I learned that vulnerability is strength.
After their card:
On Tuesday our onboarding call ran long because the new admin could not find last year’s export. I had described the folder in Slack and assumed that was enough. We now open the file together on the first call. Small, and it would have saved forty minutes.
The second post is something they can defend if a customer replies.
Voice match without impersonating a brand
Executives do not want “more thought leadership.” They want the draft to survive being read aloud in their kitchen. Collect a small style sheet from their last three posts or a transcript:
| Listen for | Then do this |
|---|---|
| Short sentences | Do not add a subordinate clause to sound smarter |
| Concrete nouns | Do not swap “Tuesday’s onboarding call” for “a recent customer interaction” |
| Rare jokes | One is enough. Models add a second and ruin both |
| What they never post | Politics, family, unnamed rivals. Leave those out even if the interview wandered |
Read the draft in their cadence, not yours. If you are a fast, witty writer and they are plain, the humanize pass should move toward plain. Say that in the prompt or in your own edit. “Make it punchier” is how ghostwritten posts drift back to the model.
The same scene should not also become a company blog with a different moral. Pick the channel. If the company page needs a customer story with an approved metric, that is a customer case study, not this post.
Workflow for the ghostwriter
- Interview for one scene. A voice note beats a questionnaire of “thought leadership topics.”
- Send the card back for a yes before you draft.
- Draft short. Executives cut more than they add.
- One Human Writes pass aimed at their spoken cadence, not at “more executive.”
- They post, or they explicitly tell you to post. Keep the approval.
- If AI helped the wording, follow the client’s disclosure rule. See client deliverable disclosure.
Do not reuse the same scene on the company page and the personal profile with different morals. Pick the channel.
Approval without a rewrite war
Send two versions at most: the card turned into prose, and one alternate opening if they hate the first line. More options invite them to mix sentences from different morals. Ask for edits in comments, not a fresh “make it more viral” prompt you then have to reconcile.
When they change a fact, update the card so the next post does not revive the old number. When they change only a word, keep it even if you liked yours. The byline is theirs. Your craft shows up as absence: no lesson they did not ask for, no adjective they would not say.
If the post is part of a weekly series, keep a list of scenes already used. Models love to regenerate “a hard conversation with the team” until every week is the same conversation. A list of used scenes stops that faster than a style guide.
Bottom line
A ghostwritten exec post is their memory and their point, in their length. Lock the card, get the yes, then humanize voice. Human Writes will not remember the Tuesday call for you.