Off Prompt

AI Tools for Small Business

Marketing

Using AI to build a simple weekly social media caption set from one piece of content you already have

How to write a week of social media posts quickly using one asset you already have. A repeatable AI workflow with prompts, platforms, and scheduling.

Owen Grant 8 min read
Using AI to build a simple weekly social media caption set from one piece of content you already have

You've got a blog post, a glowing customer review, or even just a decent photo sitting on your phone — and somehow the week goes by and you still haven't posted anything. This post shows you how to write a week of social media posts quickly by turning that one thing you already have into five platform-ready captions in a single sitting. It's more straightforward than it sounds, and you won't need anything you don't already have.

What you need before you start

ChatGPT{:target="_blank"} — a free AI tool you type to, like texting a very capable assistant who never gets tired. The free version works for everything in this post. If you want access to GPT-4o (the smarter version), it's $20/month, but start free.

Buffer{:target="_blank"} — a free tool for scheduling social posts so you write them once and they go out automatically. The free plan covers 3 channels and 10 scheduled posts at a time, which is plenty.

One piece of existing content — a blog post, a customer testimonial, a FAQ answer, a product description, even a caption you already wrote that did well. That's your raw material.

Time required: 60–90 minutes the first time. Under an hour once you've done it once.

Skill level: If you can copy and paste, you can do this.


How to batch social media content for a full week from one asset

Here's the idea before you start: one asset already contains at least five different stories. A customer testimonial has an emotional hook, a social proof angle, a practical tip, a question worth asking your audience, and probably a soft promotional angle. You're not making things up — you're just using a flashlight to look at the same room from different corners.

According to Buffer's 2024 State of Social Media report{:target="_blank"}, posting 3–5 times per week gets you similar or better results than posting daily. So five captions a week isn't a stretch goal. It's actually the sweet spot.

Step 1: Open ChatGPT (or Claude{:target="_blank"} or Gemini{:target="_blank"} — they all work) and start a new chat.

Step 2: Copy your source content — the blog post intro, the testimonial, whatever you've chosen — and keep it handy. You'll paste it in a moment.

Step 3: Before pasting anything, type this opening instruction. This tells the AI what you're doing before you give it any material to work with.

This next prompt is the engine of the whole workflow. The reason it's written this way is that telling the AI your platforms, tone, and goal before it sees the content means it doesn't have to guess — and guessing is where generic output comes from.

I'm a [describe your business in one sentence, e.g., "small plumbing company serving homeowners in Austin, TX"]. I need you to turn the content I'm about to paste into 5 social media captions — one for each of these angles: (1) educational/tip, (2) emotional or storytelling, (3) social proof, (4) promotional, (5) engagement question. For each caption, write three versions: one for Instagram (under 150 characters, 3–5 relevant hashtags), one for Facebook (under 80 characters, no hashtags), and one for LinkedIn (150–300 words, professional but conversational, one clear call to action). Use plain, direct language. No corporate speak. Here is the content: [paste your content here]

After you send this, expect a fairly long response — 15 captions total (5 angles × 3 platforms). Don't panic. Scroll through before you judge anything.

Step 4: Read through the output and mark the ones that feel closest to right. You probably won't love all 15. You don't need to.

Step 5: Pick the 5 you like best — one per day — and copy them into a simple doc or straight into Buffer.


Platform-by-platform formatting rules to include in your prompt

The prompt above already covers the basics, but here's why each platform instruction matters so you can tweak if needed.

Instagram's algorithm shows about 138–150 characters before it cuts to "more." If your hook isn't in those first characters, a lot of people scroll past. Keep the first sentence punchy. Hashtags still work, but 3–5 focused ones outperform 20 scattered ones.

Facebook is the opposite of Instagram. Sprout Social data{:target="_blank"} shows that captions under 80 characters get 66% more engagement than longer ones on Facebook. Short. Very short. One thought, one reaction.

LinkedIn rewards substance. A 150–300 word post with a real observation or a useful takeaway performs better than a quick hit. Write like you're talking to a colleague, not filing a report.


Fixing the voice problem: how to make AI sound like you

Here's the most common complaint people have: "It sounds like a press release, not me." That's a real problem, and it's easy to fix.

AI doesn't know how you talk. So tell it.

Before you run the main prompt, add a voice sample section. Grab 2–3 captions you've already written and liked — even if they're short, even if they're a bit rough — and add this to the top of your prompt:

Here are 3 examples of captions I've written that sound like me. Match this tone for all output: [paste your examples]

That's it. One extra paragraph. Claude is particularly good at picking this up — practitioners have noted its tone-matching is strong when you give it real examples. But ChatGPT and Gemini both respond well to it too.

If you don't have old captions to pull from, describe your tone in plain words. "Write like a friendly neighbor, not a brand. Casual, maybe a little funny, never stiff." That's a valid voice instruction.


The full weekly workflow to write social media posts quickly, start to finish

Here's the whole thing as a repeatable sequence. Block 90 minutes once a week — Monday morning works well, while the week is still blank.

  1. Choose your one asset. A blog post, a testimonial, a FAQ, a product photo. One thing.
  2. Open your AI tool and start a fresh chat.
  3. Paste your voice samples (3 past captions) and your business description.
  4. Paste the full caption prompt with your source content at the bottom.
  5. Review the output. Don't over-edit. Look for 5 you'd actually post.
  6. Copy the 5 picks into Buffer (or Later{:target="_blank"} or Meta Business Suite{:target="_blank"} — both free). Schedule them for the week.

Hootsuite's 2024 Social Media Trends report{:target="_blank"} found that 63% of small business owners struggle to post consistently because of time and content ideas — not because they don't care. Batching this way removes both problems at once. You're not staring at a blank screen at 7pm on a Tuesday anymore.


When something goes wrong

The captions sound generic, like a marketing brochure. This happens when the AI doesn't have enough context about your voice or your audience. Go back and add your voice samples (the 2–3 real captions you've written before) and re-run. The difference is significant.

All five angles feel similar. This usually means your source content was very narrow — like a single product spec rather than a full description or story. Try adding one more paragraph of context: why a customer loves it, a common question you get about it, or a before/after scenario.

The LinkedIn version is too long or too formal. Add this line to your prompt: "For LinkedIn, write like you're sharing a genuine observation with a smart colleague over lunch — not submitting a report." That specific framing usually pulls the tone into the right place fast.

These are normal first-pass issues. Think of the first output as a rough draft, not a final product. A couple of quick tweaks and you're done.


What to do next

Once you've got the captioning workflow running, the natural next step is reusing the same prompts for different content types — turning a frequently asked customer question into a week of posts, for example, or using a product photo as your source instead of text (GPT-4o and Gemini 2.0 can read images, so you can literally paste a photo and get captions back). If you want to go deeper on building a repeatable content system beyond captions, we've got a walkthrough on building a monthly content calendar with AI for small businesses.


FAQ

Do I need the paid version of ChatGPT to write a week of social media posts? No. The free version handles everything in this post. The paid tier (GPT-4o at $20/month) gives you faster responses and the ability to paste images, but for text-to-captions work, free is genuinely fine. Start free and upgrade later if you want to.

What if I don't have any old captions to use as voice samples? Good question — most people starting out don't. Instead, write 2–3 sentences describing how you'd talk to a regular customer in person. "Friendly, direct, a bit casual, never salesy" is a perfectly useful voice instruction. You can refine it over time as you see what outputs feel closest to right.

Can I use this workflow to repurpose one piece of content into social media posts for TikTok or X (Twitter) too? Yes. Just add the platform and its format rules to your prompt. For X, the limit is 280 characters and no hashtags unless they're very targeted. For TikTok, captions are short and the hook matters most — front-load the interesting part in the first sentence.

How do I know which angle to post on which day? There's no perfect formula, but a loose pattern that works well: educational on Monday (people are in learning mode), story or emotional mid-week, social proof or promotional toward the end of the week when people are more likely to act, and an engagement question on Friday or Saturday when they have a moment to respond.

What if I only have one or two social media platforms — do I still get value from this? Absolutely. Just tell the AI which platforms you're on and skip the rest. You'll still get five different angles — you'll just format them for fewer platforms. Less output, same amount of useful variety.

Was this useful? ·