emily in ai
How to Use AI for Social Media Content
Guides

How to Use AI for Social Media Content

I used to spend entire Sundays batch-creating content and still feel behind, then I figured out how to actually use AI for social media, and now I batch a month's worth of posts in like three hours.

Okay so here's the thing about AI and social media content: most people are using it wrong. They're typing "write me an Instagram caption" and getting something so stiff and generic that they have to rewrite the whole thing anyway. That's not AI helping you. That's AI wasting your time.

I went through that phase. Spent weeks being disappointed before I figured out what actually works. And now I genuinely love using AI for content, like, I look forward to my batch days. Wild, right?

Stop asking AI to write your posts. Start using it to think.

The biggest shift for me was treating AI less like a ghostwriter and more like a really fast brainstorming partner. When I sit down for a content batch session, I don't open Claude and say "write 30 Instagram captions." I start by dumping my brain into it.

I'll say something like: "Here's what I'm thinking about this week. I've been trying out this new AI tool for my morning routine, I had a frustrating experience with a chatbot that gave me wrong info, and I've been thinking about how my content strategy has changed since I started using AI. What are 10 different angles I could take across different platforms?" And then I actually read what it gives me and pick the ones that feel real.

That's the move. Use AI to surface angles you already had somewhere in your head but couldn't articulate yet. It's faster than staring at a blank document for 45 minutes, which is what I used to do.

Platform-specific is everything

Here's what I mean about most people doing this wrong: they generate one post and then copy-paste it everywhere. That's actually worse than not using AI at all, because it looks lazy. Instagram, LinkedIn, TikTok, and Threads all have totally different vibes and audiences. A LinkedIn post that performs well sounds nothing like a TikTok caption that performs well.

I have a little system now. When I'm generating content for different platforms, I give the AI a specific persona brief for each one. For LinkedIn: "professional but not stiff, genuine personal insight, first person, 150-200 words." For Instagram: "casual, one main idea, short sentences, conversational tone." For TikTok captions: "under 100 characters, curious or punchy."

I paste those briefs in every time, and the output quality jumped dramatically. The AI isn't psychic, it needs to know what you're going for, and if you don't tell it, it defaults to this weird formal-but-friendly corporate voice that doesn't sound like anyone.

The hook problem (and how AI actually helps here)

Writing hooks used to be the thing I spent the most time on. Like, 70% of my content creation time was just trying to figure out how to start. AI is genuinely good at this, but only if you give it something real to work with.

I'll write my main point first, the actual thing I want to say, and then ask the AI to give me 10 different opening lines for it. Not 10 variations of the same line, but 10 different approaches: a bold claim, a personal story opener, a stat, a question, a surprising contrast. Then I pick the one that feels most like me and actually edit it.

That last part matters. Edit it. I can always tell when someone just copy-pasted an AI hook without touching it. There's this slightly-too-polished quality to it, and your regular audience will notice even if they can't articulate why. Take the AI's idea, make it yours.

Repurposing existing content is where AI genuinely shines

This is my favorite use case and I don't talk about it enough. If I've written a long-form piece, done a YouTube video, or published a newsletter, I can feed that into AI and ask it to pull out the five most standalone insights and reframe each one for a different platform.

I did this with a newsletter I wrote about how I organize my AI tool subscriptions. That one piece became eight separate posts across three platforms, each totally different, each feeling native to the platform. It would've taken me probably four or five hours to do that manually. With AI it took maybe forty minutes, and a lot of that was me editing and actually thinking about which pieces to use.

Real talk: this is what "content repurposing" was always supposed to be, but it was too tedious to actually do consistently before AI made it fast.

What AI is bad at (so you know where to actually spend your time)

I'm not going to pretend AI does everything well. It doesn't.

AI is bad at your specific voice when it doesn't have examples. If you haven't given it samples of your actual writing, it will default to something that sounds like a slightly more interesting version of a press release. You have to train it on you, paste in your best-performing posts, explain what worked, tell it what you want to sound like.

AI is bad at current references. It doesn't know what happened last week, it doesn't know about the trending audio everyone's using right now, it doesn't know that a particular format is totally overdone on your platform at this moment. That cultural currency is yours to bring. AI just helps you build the structure around it.

And honestly? AI is bad at replacing the part of content creation that requires you to actually live your life and have experiences worth sharing. That's still on you. AI can help you say it better, say it faster, say it across more platforms. But you have to have something to say first.

A simple system that actually works

Here's what my actual batch process looks like now. I give myself a brain dump session first, just talking to the AI about what's been on my mind, what I've been testing, what frustrated me, what surprised me. Then I ask it to identify the most interesting angles. Then I develop those angles into platform-specific content one by one, with my platform briefs pasted in. Then I do a pass editing everything to sound more like me.

Start to finish, a full month of content across three platforms takes me about three hours. It used to take me a full weekend and I still felt behind. I could be wrong about what works for everyone, but this process has been consistent for me for six months and I'm not changing it.

Emily in AI

Emily in AI is a plain-English guide to AI tools, tips, and beginner guides. Every tool gets tested and written up without the hype or the jargon, so you can figure out what actually helps. New posts every week.

About Emily in AI →