The difference between AI-generated text that sounds robotic and text that sounds human comes down to one thing: the specificity of your instructions.
"Write a blog post about productivity" gets you garbage. The prompts below get you content that readers can't tell was AI-assisted.
1. The Anti-AI-Sounding Writer
Write about [topic] for [audience]. Important rules:
- Write like you're explaining to a smart friend over coffee
- Use contractions (don't, won't, I'd)
- Vary sentence length: mix short punchy sentences with longer explanatory ones
- Include one specific personal anecdote or concrete example per section
- No phrases: "in today's fast-paced world", "it's worth noting", "at the end
of the day", "dive deep", "leverage", "unlock"
- Start paragraphs with different words (not "The", "This", "It" repeatedly)
- End with a specific actionable takeaway, not a generic conclusion
Why it works: The "banned phrases" list alone transforms output quality. Those phrases are how readers detect AI writing.
2. The Email That Gets Replies
Write a cold outreach email to [person/role] about [purpose].
Context: [why you're reaching out, any mutual connection]
Goal: [specific action you want them to take]
Tone: professional but human, not salesy
Rules:
- Subject line: under 6 words, creates curiosity
- First sentence: about THEM, not about you
- Body: under 100 words total
- One specific ask (not "would love to chat sometime")
- No buzzwords, no exclamation marks, no emojis
- Sign off with just your first name
Benchmark: Cold emails average 1-3% reply rate. Well-structured ones hit 10-25%.
3. The Blog Post Outliner
Create a detailed outline for a blog post about [topic].
Target reader: [who they are and what they already know]
Goal: reader should be able to [specific outcome] after reading
Length: [word count]
Structure:
1. H1: Title (include primary keyword, under 60 chars)
2. Hook paragraph: start with a surprising fact or bold statement
3. H2 sections: each one answers a specific question the reader has
4. For each section: main point + supporting evidence + practical example
5. Conclusion: specific next action (not "in conclusion...")
Include suggested internal links and external source references.
4. The Twitter/X Thread Creator
Turn this [article/idea/insight] into a Twitter thread.
Thread structure:
- Tweet 1: Hook (the most surprising or contrarian point)
- Tweets 2-6: One idea per tweet, each can stand alone
- Tweet 7: Summary + CTA
Rules per tweet:
- Under 280 characters
- No hashtags in the thread body (only tweet 7)
- Use line breaks for readability
- Numbers and specifics over adjectives
- Tweet 1 must make someone stop scrolling
5. The SEO Article Structure
Write a 2000+ word SEO article targeting the keyword "[keyword]".
Structure:
1. Title: include exact keyword, under 60 characters
2. Meta description: include keyword, 155 characters, compelling
3. First paragraph: answer the search intent in 2 sentences
4. H2 sections: each targets a related long-tail keyword
5. Include: numbered lists, comparison tables, FAQ section
6. Every claim backed by a specific example or data point
7. Internal linking suggestions: 3 related articles to link to
8. CTA: natural product mention in context, not forced
Target: position zero (featured snippet) formatting
The Framework Behind All of Them
Every writing prompt above uses the same structure:
- Who is reading this?
- What should they be able to do after reading?
- What makes good writing? (specific rules, not vague adjectives)
- What to avoid? (the banned phrases/patterns list)
Give AI these four things and the output quality jumps from "obvious AI" to "did a person write this?"
These prompts are from a collection of 170 covering writing, coding, creative, business, and automation.
Get all 170 prompts: The AI Toolkit 2026 — includes 50 free tools and 30 automation workflows.
What's your trick for making AI writing sound human? Share in the comments.
Top comments (0)