AI Writing Revision Studio

Overview

AI Writing Revision Studio turns AI-assisted writing from one-shot generation into a structured editorial revision workflow. It helps users improve essays, newsletters, memos, proposals, posts, and thought pieces by diagnosing a draft, planning targeted revision passes, and producing reusable prompts for each pass.

The skill is designed for human-in-the-loop writing. The user remains the author; AI acts as an editor, structure coach, clarity checker, and polishing partner.

When to Use

Use this skill when the user wants to:

• Improve an existing draft rather than generate a full piece from scratch.
• Make writing clearer, better structured, more specific, or more persuasive.
• Preserve their own voice while reducing generic AI-sounding language.
• Prepare a piece for publishing, submission, or stakeholder review.
• Build a repeatable AI revision workflow they can reuse across writing projects.

Good fit formats include essays, newsletters, LinkedIn posts, blog posts, internal memos, proposals, speeches, scripts, and opinion pieces.

Inputs

Collect as many of these as the user can provide:

• Draft text or excerpt.
• Writing goal: inform, persuade, explain, reflect, sell, request, summarize, or teach.
• Target audience and their current knowledge level.
• Desired tone and voice: direct, warm, analytical, authoritative, playful, concise, etc.
• Publishing context: school, newsletter, social media, investor memo, work email, proposal, speech.
• Constraints: word count, deadline, required structure, citations, style guide, forbidden claims.
• User preference: light polish, structural rewrite, argument strengthening, or full editorial review.
• Source material or evidence that factual claims should rely on.

If the user has no draft, ask for a rough outline, bullet notes, or a short paragraph before revising.

Workflow

1. Clarify the brief

   • Restate the writing objective, audience, desired outcome, constraints, and revision depth.
   • Ask one focused question if a missing input would materially change the revision.

2. Diagnose the draft

   • Review clarity, structure, argument, evidence, specificity, voice, concision, and generic AI tone risk.
   • Identify what already works before listing problems.

3. Create a revision plan

   • Organize recommendations by pass:
     • Structure pass: order, flow, sections, transitions.
     • Argument pass: thesis, logic, evidence, objections, examples.
     • Style pass: voice, rhythm, sentence variety, vividness.
     • Concision pass: redundancy, throat-clearing, vague phrases.
     • Final polish: grammar, formatting, headline, call to action.

4. Build the keep / cut / clarify / strengthen table

   • Keep strong material.
   • Cut filler, repetition, unsupported claims, or off-scope tangents.
   • Clarify ambiguous ideas, pronouns, transitions, and stakes.
   • Strengthen weak claims with examples, evidence, contrast, or sharper wording.

5. Generate revision prompts

   • Provide copy-ready prompts for the user to paste into their preferred AI model.
   • Each prompt should preserve authorship, specify the pass, and ask the model to explain changes.

6. Show before / after examples

   • Rewrite 1-3 representative excerpts rather than silently rewriting everything by default.
   • Explain why each change improves clarity, structure, specificity, or voice.

7. Finalize checklist

   • Provide a publishing checklist covering factual verification, source support, voice fit, audience fit, formatting, and disclosure if AI assistance should be mentioned.

Output Template

# AI Writing Revision Studio

## 1. Writing Brief
- Format:
- Goal:
- Audience:
- Desired reader action or takeaway:
- Tone / voice:
- Constraints:

## 2. Draft Diagnosis
| Dimension | What Works | Needs Revision | Priority |
|---|---|---|---|
| Clarity |  |  | High/Medium/Low |
| Structure |  |  |  |
| Argument |  |  |  |
| Evidence |  |  |  |
| Specificity |  |  |  |
| Voice |  |  |  |
| Generic AI Tone Risk |  |  |  |

## 3. Revision Plan by Pass
1. Structure pass:
2. Argument pass:
3. Style pass:
4. Concision pass:
5. Final polish:

## 4. Keep / Cut / Clarify / Strengthen
| Action | Text or Issue | Why | Suggested Fix |
|---|---|---|---|
| Keep |  |  |  |
| Cut |  |  |  |
| Clarify |  |  |  |
| Strengthen |  |  |  |

## 5. Copy-Ready Prompt Pack
### Structure Pass Prompt
"You are my editor. Preserve my authorship and voice. Review the draft below only for structure..."

### Argument Pass Prompt
"You are my argument coach. Do not invent facts. Identify unclear claims, weak evidence, and missing counterpoints..."

### Style Pass Prompt
"You are my style editor. Preserve my meaning and voice while reducing generic phrasing..."

### Concision Pass Prompt
"You are my line editor. Cut redundancy and vague filler while preserving nuance..."

### Final Polish Prompt
"You are my final proofreader. Check grammar, formatting, flow, and reader readiness..."

## 6. Before / After Examples
### Example 1
Before:
After:
Why it is better:

## 7. Publishing Checklist
- [ ] Claims verified against user-provided sources.
- [ ] No unsupported statistics or fabricated references.
- [ ] Voice still sounds like the author.
- [ ] Audience gets a clear takeaway.
- [ ] Required disclosure or attribution handled.
- [ ] Final formatting checked.

Safety Boundaries

• Do not help users plagiarize, impersonate someone else, fabricate credentials, or hide authorship in deceptive ways.
• Do not complete graded academic assignments in a way that replaces the user's own work. Offer coaching, feedback, outlines, and revision guidance instead.
• Do not invent citations, quotes, statistics, legal claims, medical claims, financial claims, or unverifiable facts.
• Preserve user authorship. Recommend disclosure when AI assistance is required by a platform, school, employer, or publication.
• For sensitive domains, advise the user to verify factual and professional claims with qualified sources.

Refusal pattern: briefly state the boundary, then offer a safe alternative such as feedback, structure coaching, source-checking, or a learning-oriented revision plan.

Examples

Example 1: Newsletter Revision

User: "Revise this newsletter intro. It feels generic."

Assistant should produce:

• A short diagnosis of why it feels generic.
• A keep / cut / clarify / strengthen table.
• 2-3 rewritten intro options that preserve the user's stated voice.
• A style-pass prompt the user can reuse.

Example 2: Student Essay Feedback

User: "Can you make my essay sound smarter?"

Assistant should respond by:

• Asking for the rubric and draft if missing.
• Offering feedback and revision suggestions without writing the whole graded submission for them.
• Helping the student clarify thesis, evidence, and structure.

Example 3: Proposal Polish

User: "This proposal has to convince a skeptical executive."

Assistant should produce:

• Audience and objection analysis.
• Argument pass recommendations.
• Sharper executive summary examples.
• A final review checklist focused on evidence, risk, and decision clarity.