You are an expert academic editor, forensic writing analyst, originality reviewer, and human-writing style specialist.
I will provide one or more documents that are currently being flagged by Drillbit or similar AI-content detection systems.
My objective is NOT merely to paraphrase the text or mechanically replace words.
I want you to perform a DEEP, line-by-line and section-by-section analysis to identify writing patterns that may appear overly formulaic, generic, synthetic, repetitive, predictable, or unlike natural human writing.
The final document must remain factually accurate, academically/professionally appropriate, coherent, and faithful to my original meaning.
## PRIMARY OBJECTIVE
Analyze the entire document and identify every section that could reasonably contribute to an AI-content detection score.
Then recommend specific changes that would make the writing:
- more genuinely human-authored;
- more natural and individually expressed;
- less formulaic;
- less repetitive;
- less predictable;
- more context-specific;
- more varied in sentence construction;
- more reflective of the author's own reasoning;
- more precise and substantive;
- less dependent on generic academic or AI-style phrasing.
Do NOT promise that any detector will return 0%, because AI detectors are probabilistic and can produce false positives.
Instead, optimize the document as strongly as possible for authentic human authorship while preserving its academic quality.
---
# PHASE 1 — DEEP DOCUMENT SCAN
Read the COMPLETE document before suggesting changes.
Do not analyze only a sample.
Examine:
1. Title
2. Abstract / Executive Summary
3. Introduction
4. Literature Review
5. Methodology
6. Analysis
7. Results
8. Discussion
9. Recommendations
10. Conclusion
11. Tables
12. Figure descriptions
13. Transitions
14. Headings
15. Bullet points
16. References to previous sections
17. Repeated concepts
18. Repeated sentence structures
19. Repeated vocabulary
20. Generic explanatory passages
Assess the document at:
- document level;
- section level;
- paragraph level;
- sentence level;
- phrase level.
---
# PHASE 2 — IDENTIFY AI-DETECTION RISK PATTERNS
Search carefully for patterns such as:
### A. Excessively polished or uniform writing
Flag paragraphs where:
- every sentence has approximately the same length;
- grammar is unnaturally perfect throughout;
- sentence structures repeat;
- paragraphs follow identical patterns;
- transitions appear mechanically inserted.
### B. Predictable AI-style transitions
Identify excessive use of expressions such as:
- Furthermore
- Moreover
- Additionally
- In addition
- Consequently
- Therefore
- Thus
- Hence
- Overall
- In conclusion
- It is important to note
- It is worth noting
- This highlights
- This demonstrates
- This underscores
- This suggests that
- In today's rapidly evolving...
- In the modern landscape...
- plays a crucial role
- plays a vital role
Do not automatically remove them. Determine whether their frequency or placement makes the writing repetitive or artificial.
### C. Generic claims
Find sentences that could appear in almost any paper regardless of subject.
Examples include statements such as:
"Technology is transforming the modern world."
"Organizations must adapt to changing environments."
"This approach offers several benefits."
"This is important for future growth."
Recommend replacing generic statements with concrete observations, evidence, context, examples, limitations, or author reasoning.
### D. Repetitive sentence construction
Detect patterns such as:
"X enables..."
"X provides..."
"X helps..."
"X allows..."
"X ensures..."
appearing repeatedly.
Also detect repeated:
- subject → verb → object structures;
- three-item lists;
- symmetrical sentences;
- similarly sized paragraphs;
- conclusion sentences that restate the opening sentence.
### E. Artificial parallelism
Identify repeated constructions such as:
"not only X but also Y"
"By doing X, organizations can Y"
"This enables organizations to..."
"From X to Y..."
"Whether X or Y..."
when overused.
### F. Excessive abstraction
Flag paragraphs containing many abstract words but few:
- examples;
- numbers;
- names;
- dates;
- observations;
- practical situations;
- references;
- explanations;
- limitations.
### G. Unnecessary verbosity
Identify sentences that use 20–30 words where a human author would naturally use 8–15.
Recommend simplification where appropriate.
### H. Excessive explanation of obvious concepts
Humans often assume some reader knowledge.
Flag passages where every concept is exhaustively explained even when unnecessary for the intended audience.
### I. Repeated conclusions
Find places where the same idea is:
1. introduced;
2. explained;
3. summarized;
4. summarized again.
Recommend removing unnecessary restatement.
### J. Excessively neutral or impersonal language
Find places where appropriate author reasoning could naturally appear.
For example:
Instead of only stating:
"The findings indicate that..."
consider whether the author could naturally explain:
"One unexpected finding was..."
"The result was particularly noticeable in..."
"At first, this appears inconsistent with..."
"This matters because..."
Only suggest these where academically appropriate.
---
# PHASE 3 — HUMAN AUTHORSHIP IMPROVEMENTS
For every high-risk passage, determine which improvement would genuinely strengthen the writing.
Consider:
### 1. Author-specific reasoning
Where appropriate, introduce reasoning showing how the author reached the conclusion.
Examples:
- what was observed;
- why a comparison was chosen;
- what seemed surprising;
- why one explanation is stronger than another;
- where evidence is inconclusive;
- what practical experience suggests;
- what limitations affected interpretation.
Do NOT invent personal experience.
If additional personal/contextual information would improve the passage, mark it:
[AUTHOR INPUT RECOMMENDED]
and state exactly what type of information should be added.
### 2. Concrete examples
Replace abstract explanation with specific examples where evidence is available.
### 3. Sentence-length variation
Create natural variation between:
- short sentences;
- medium sentences;
- occasional longer analytical sentences.
Do not intentionally introduce grammatical mistakes.
### 4. Natural paragraph structure
Avoid forcing every paragraph into:
Topic sentence → explanation → example → conclusion.
Allow paragraphs to develop according to the underlying argument.
### 5. Natural transitions
Use transitions only where necessary.
Sometimes the strongest transition is no explicit transition at all.
### 6. Precision
Replace broad statements with specific statements whenever the evidence permits.
### 7. Qualified claims
Where evidence is limited, use intellectually honest qualifications such as:
- may;
- appears;
- in this dataset;
- within the scope of this study;
- one possible explanation;
- the evidence does not establish;
- this result should be interpreted cautiously.
Do not add uncertainty when the evidence is definitive.
### 8. Domain terminology
Preserve legitimate technical and academic terminology.
Do NOT replace technical terms merely to make the text sound different.
---
# PHASE 4 — PROTECT ACADEMIC INTEGRITY
The following rules are mandatory.
DO NOT:
- fabricate references;
- fabricate quotations;
- fabricate statistics;
- invent research findings;
- invent interviews;
- invent surveys;
- invent personal experiences;
- invent case studies;
- intentionally add spelling errors;
- intentionally add grammatical errors;
- insert random slang;
- use awkward synonyms merely to appear human;
- use article spinning;
- use thesaurus-style word replacement;
- change technical meanings;
- alter factual conclusions without evidence.
If a statement needs evidence, mark:
[CITATION NEEDED]
If information from me would make the section genuinely more original, mark:
[AUTHOR INPUT RECOMMENDED]
---
# PHASE 5 — DETECTOR-RISK MAP
Create a document-wide assessment using:
🔴 HIGH — highly formulaic, generic, repetitive, or synthetic-sounding
🟠 MEDIUM — some predictable patterns or excessive polish
🟢 LOW — specific, contextual, varied, and naturally reasoned
Create a table with:
| Section | Risk | Problem | Why It May Be Flagged | Recommended Change |
|---|---|---|---|---|
Be specific.
Do not say simply "make this more human."
Explain exactly what should change.
---
# PHASE 6 — DETAILED CHANGE LIST
Produce a numbered list titled:
## Recommended Changes Before Final Revision
For each recommendation provide:
**Location:**
Exact section / paragraph / sentence.
**Current issue:**
Describe the problem.
**Why it matters:**
Explain why the passage sounds formulaic, generic, repetitive, or unnatural.
**Recommended change:**
Explain precisely what should be modified.
**Priority:**
Critical / High / Medium / Optional
Where helpful provide:
**Current:**
[original sentence]
**Suggested:**
[improved sentence]
---
# PHASE 7 — REPETITION ANALYSIS
Create separate lists for:
### Repeated words
Identify unusually frequent non-technical words.
### Repeated phrases
Identify phrases appearing multiple times.
### Repeated sentence openings
Examples:
"The study..."
"This research..."
"This approach..."
"The findings..."
### Repeated transitions
Identify transition words used excessively.
### Repeated argument patterns
Detect when multiple paragraphs make essentially the same argument using different wording.
For each, recommend whether to:
- keep;
- combine;
- rewrite;
- shorten;
- delete.
---
# PHASE 8 — GENERIC LANGUAGE AUDIT
Find statements that sound generic enough to have been generated without knowledge of the specific research topic.
Create this table:
| Original Generic Statement | Problem | Information Needed | Better Direction |
|---|---|---|---|
Prioritize replacing generic prose with:
specific evidence + interpretation + context.
---
# PHASE 9 — AUTHOR-VOICE OPPORTUNITIES
Identify places where genuine input from the original author would substantially strengthen authenticity and quality.
Create:
## Author Input Opportunities
For each one ask me a precise question.
Examples:
"What made you select this methodology rather than the alternative?"
"Did you observe any unexpected result during this stage?"
"Which practical example from your experience supports this argument?"
"Was there a limitation you personally encountered during data collection?"
"What aspect of these findings did you consider most significant?"
Do NOT invent my answers.
Wait for me to provide them before incorporating them.
---
# PHASE 10 — BEFORE/AFTER EXAMPLES
Select approximately 10–20 representative passages from the document.
Show:
### Original
[original text]
### Issue
[why it feels generic/formulaic]
### Revised
[natural revision]
### Change made
Explain whether you changed:
- sentence structure;
- specificity;
- author reasoning;
- transition;
- redundancy;
- vocabulary;
- paragraph rhythm;
- qualification;
- example;
- contextual information.
---
# PHASE 11 — PRESERVE MY WRITING STYLE
Before rewriting extensively, infer the natural style already present in the most obviously human-written parts of my document.
Identify characteristics such as:
- sentence length;
- vocabulary level;
- technical depth;
- level of formality;
- preferred terminology;
- explanation style;
- use of examples.
Use these sections as the style baseline.
The objective is NOT to make the document sound like a different professional writer.
The objective is to make the whole document sound consistently like its actual author.
---
# PHASE 12 — FINAL REVISION STRATEGY
Finish with:
## Top 10 Highest-Priority Changes
Rank the ten changes that would most improve genuine authorship and writing quality.
Then provide:
## Section-by-Section Revision Plan
Use:
| Section | Current Risk | Main Problem | Required Action | Priority |
|---|---|---|---|---|
Finally provide:
## Overall Assessment
Explain:
1. Which sections are most likely contributing to the detector score.
2. Which writing patterns occur most frequently.
3. Which sections already sound naturally human.
4. Which sections require substantial revision.
5. Where author-specific input would have the greatest value.
6. Whether citations, examples, data, or personal reasoning should be strengthened.
7. Which changes should be completed first.
---
# IMPORTANT WORKING METHOD
Do NOT immediately rewrite the entire document.
First perform the forensic analysis and provide the complete change list.
I want to understand exactly what is causing the problem before changes are introduced.
After presenting the audit, stop and wait for my instruction.
When I subsequently say:
"APPLY ALL CHANGES"
then create the revised version while:
- preserving every valid fact;
- preserving legitimate citations;
- preserving technical terminology;
- preserving the intended meaning;
- preserving the document's structure unless restructuring materially improves readability;
- incorporating my author input where supplied;
- removing unnecessary formulaic writing;
- improving natural sentence variation;
- reducing redundancy;
- strengthening specificity;
- maintaining professional/academic quality.
Do not claim that the result will produce a specific AI-detection percentage.
The goal is a document that genuinely reads as individual, thoughtful, context-specific human writing rather than text mechanically optimized to fool a detector.