←Back/Content/ChatGPT
BeginnerContent

How to Create a 'Mega Prompt' in ChatGPT to Refine and Humanize Your Writing

Turn recurring writing preferences into an editing prompt that flags generic language and model habits, but keep the writer in charge of the draft, factual content, exceptions, and final voice.

How to Create a 'Mega Prompt' in ChatGPT to Refine and Humanize Your Writing

Source episode · 30:00

How I AI: Lee Robinson's Workflows for Resilient Code with Cursor and Sharper Writing with ChatGPT

with Claire Vo

Lee shows a writing mega-prompt built from personal preferences, banned generic words, and recurring model patterns, then explains that he prefers to draft himself and use AI as a reviewer rather than a ghostwriter.

Read the full episodeWatch on YouTubeListen on SpotifyListen on Apple

Before you start

What you need

  • A draft written or outlined by the author
  • Several examples of the author's preferred voice
  • A short list of personal verbal habits and generic claims to question
  • Known model patterns that repeatedly appear in edits
  • Facts, quotes, links, names, and calls to action that must remain accurate

What you’ll make

An edited draft that preserves the author's argument and voice while identifying vague claims, repeated habits, and formulaic model language for the author to accept or reject.

Tools used

Step by step

The workflow

Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.

3 steps

Step01

Build Your 'Banned Words & Phrases' List

Build a short, evidence-based watchlist from your own repeated habits, vague marketing claims, and feedback from trusted editors. For every flagged word, record the problem and at least one specific alternative rather than banning it blindly.

Example prompt
Review this draft for my watchlist: [words and phrases]. For each match, quote the sentence, explain whether it is vague, repetitive, or justified, and propose a more specific alternative. Do not rewrite the full draft yet.

Start with patterns you can point to in your own drafts. A short list with examples is more useful than a universal blacklist.

Step02

Identify and Ban Common LLM Patterns

Add recurring model structures such as canned enthusiasm, false contrasts, excessive lists, choppy affirmations, generic conclusions, and unsupported significance. Include examples of the author's actual voice so the prompt has something positive to preserve.

Example prompt
Also flag these model patterns: [patterns]. Preserve the rhythm and phrasing in these voice examples: [examples]. Do not remove a construction merely because it appears on the list; explain why it weakens this sentence.
Step03

Integrate the Prompt into a Human-First Writing Process

Draft the substance yourself, or use AI only to escape a blank page. Then run the mega-prompt as an editorial pass, inspect a change list, accept or reject each material edit, and read the final piece aloud.

Example prompt
Act as an editor, not a ghostwriter. Preserve my argument, facts, examples, quotes, links, and voice. Return: factual issues, vague or formulaic passages, proposed line edits, and a clean revision. Mark any edit that changes meaning. Do not add claims or a generic conclusion.

What good looks like

  • Every factual claim, quote, link, number, and commitment remains unchanged unless the author explicitly corrects it.
  • The editor explains flagged language and proposes specific alternatives rather than applying a universal ban.
  • The revision keeps deliberate quirks and exceptions that sound like the writer.
  • The final draft is read and accepted by the author rather than copied directly from model output.

Build your next product with ChatPRD

Turn an idea into a PRD, user stories, and a plan.

Try ChatPRD free

After the steps

Runbook notes

How to recover when the loop fails and where human judgment helps.

Recover

If it goes sideways

A long banned-word list removes the writer's personality and makes every sentence plain
Treat rules as flags, use voice examples, and let the author keep intentional phrasing.
The edit changes a product claim, quotation, date, link, or promised action
Lock factual spans before editing, compare them after the rewrite, and correct from the source.
The prompt targets current model tics but produces a new set of formulaic replacements
Review the full cadence and argument, not only word matches, and update examples when a new pattern recurs.
Sensitive workplace text is pasted into an unapproved consumer account
Use an approved environment, redact unnecessary details, or edit locally when the content is confidential.

Keep building

Related workflows

Create an Automated Workout Tracker with a Simple Claude Voice Note
beginnerPersonalClaude

Create an Automated Workout Tracker with a Simple Claude Voice Note

Log your workouts effortlessly without ever opening a spreadsheet. This workflow uses the Claude mobile app to parse a simple voice note and automatically update your workout tracking spreadsheet in the background.

Open
How to Rebuild Your Gmail Inbox Inside Claude to Manage Email
advancedOperationsClaude

How to Rebuild Your Gmail Inbox Inside Claude to Manage Email

Create a custom email management interface within Claude to read, triage, and draft replies without opening the Gmail tab. This workflow helps you reclaim focus and compound your learnings by training your AI on every interaction.

Open
Automate Client Proposals and Onboarding with a Claude 'Pipeline Operator'
intermediateSalesClaude

Automate Client Proposals and Onboarding with a Claude 'Pipeline Operator'

Build a fully automated system in Claude to generate personalized, interactive HTML proposals and client onboarding materials, saving hours on administrative work and creating a high-touch experience.

Open

Start shipping
better products.

Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.

Start building freeRequest a demo
Free to start
No credit card
SOC 2 certified
Enterprise ready