AI can dramatically increase your social media output — but used badly, it produces generic content that tanks your engagement. This guide shows you how to use AI as a multiplier for your ideas, not a replacement for them.

The Right Mental Model

AI for social media should:

  • Turn your ideas into more polished posts
  • Repurpose your long-form content into multiple formats
  • Generate variations for testing
  • Draft posts faster so you publish more consistently

AI should NOT:

  • Replace your perspective and opinions
  • Generate generic “5 tips for…” posts with no point of view
  • Write posts that could have been written by anyone about anything

The difference between AI-powered accounts that grow and ones that plateau is whether there’s a real human perspective behind the content. AI is the word processor; you’re the writer.

Step 1: Build Your Voice Document

Before using AI, create a 1-page voice document:

## My Voice and Perspective

**Industry/niche:** [What you write about]

**My point of view:** [What I believe that others might disagree with]

**Topics I cover consistently:** [4-6 topics]

**Writing style:**
- [Characteristic 1: e.g., "Use simple words, not jargon"]
- [Characteristic 2: e.g., "Short sentences under 20 words"]
- [Characteristic 3: e.g., "Personal anecdotes, not case studies"]

**Things I don't do:**
- [e.g., "Don't use corporate speak or buzzwords"]
- [e.g., "Don't write list posts without genuine insight"]

**Example posts I love (paste 3-5 examples)**

Feed this document to Claude as a system prompt or paste it at the start of every session.

Step 2: Post Writing Workflow

Starting From an Idea

Your idea: “Most startups waste money on paid ads too early”

Prompt:

I want to write a LinkedIn/Twitter post about this idea: [your idea]

My voice: [paste voice document]

Constraints:
- LinkedIn: max 300 words, no bullet lists, conversational
- Twitter: max 280 characters, punchy opener

Write 3 versions with different angles:
1. Story-led (start with a specific anecdote)
2. Contrarian take (challenge the conventional wisdom)
3. Data/observation (start with a surprising insight)

The Editing Step

AI gives you a draft, not a final post. Edit it to:

  • Replace generic phrases with specific details from your experience
  • Add a story or example that only you could tell
  • Adjust the voice to sound more like you
  • Add your actual opinion at the end if it’s missing

Step 3: Repurposing Long-Form Content

Your best content already exists — in blog posts, podcast episodes, newsletter issues. Repurposing it is often more effective than creating from scratch.

From Blog Post to Social Posts

Prompt:

I wrote this blog post: [paste full article or key excerpts]

Extract 8 ideas worth posting on LinkedIn and Twitter/X.
For each idea:
1. The core insight in one sentence
2. A LinkedIn version (150-250 words, narrative format)
3. A Twitter thread version (5-7 tweets)

My voice guidelines: [paste voice doc]

From Podcast/Video Transcript

Here's a transcript of a podcast episode: [paste transcript]

Find the 5 most quotable moments — things I said that 
are either surprising, counterintuitive, or highly specific.

For each:
1. The exact quote
2. Context needed to understand it
3. A standalone social post version (50-100 words)

Step 4: Platform-Specific Formats

LinkedIn

What works: insights from experience, professional lessons learned, career perspectives.

LinkedIn post structure that works:

Hook line (makes people stop scrolling)
[blank line]
Short sentence that deepens the hook.
[blank line]
The actual substance in 3-6 short paragraphs.
[blank line]
A specific takeaway or observation.
[blank line]
Question to prompt comments.

Prompt:

Rewrite this post for LinkedIn:
[draft]

LinkedIn format:
- Short hook line that stops scrolling
- Short paragraphs (3-4 lines max)
- End with a genuine question that invites discussion
- Keep under 250 words
- Don't use emojis at the start

Twitter/X

What works: hot takes, threads with insight, commentary on industry news, specific data points.

Thread structure:

Tweet 1: The main claim or surprising insight
Tweets 2-5: Evidence, examples, or expansion of each point
Tweet 6: Contrarian view or nuance
Tweet 7: The practical takeaway
Tweet 8: Call to action or discussion question

Instagram (Captions)

For creators who use Instagram, captions need a different approach — shorter, more visual storytelling.

Write an Instagram caption for a photo of [describe photo].
My audience: [description]
Caption style: first line hooks, personal story of 3-4 sentences, 
relatable insight, question at end, relevant hashtags (5-8, not more)

Step 5: Content Calendar

Monthly Batch Creation

Sit down once a month and create your content calendar for the next 4 weeks:

Prompt:

I want to create a 4-week LinkedIn content calendar.
I post 3x per week.

My content pillars: [3-4 topics you cover]
My voice: [paste doc]
Goal: [grow followers / build brand / generate leads]

For each of the 12 posts, give me:
- Post theme/topic
- Core insight or angle
- Format (personal story / insight / data / question)

Keep variety across the month.

Then use this calendar to prompt specific drafts throughout the month.

Maintaining Authenticity at Scale

The #1 mistake with AI social content: publishing AI’s first draft.

Review every post for:

  1. Does this sound like me? (Not “sounds reasonable” — sounds like me)
  2. Is there a specific detail only I could know?
  3. Would this be useful to one specific person I know?
  4. Does it have an actual point of view, or is it “on one hand… on the other hand”?

Good AI-assisted content uses AI for speed, then adds your perspective for value. It’s a 70/30 split: 70% of the work faster, 30% human editing that makes it worth reading.