How I Use AI to Write Better Social Media Captions

My name is Jeffrey Mdala, an AI Engineer & Founder based in Lusaka, Zambia. I run the Zambian Online Education Company (ZOEC), where I built eskulu, an AI-powered learning platform helping students across Zambia access ECZ-aligned notes, quizzes, past papers, and AI support. Over the years, I have spent a lot of time not only building AI systems, but also learning how to use AI practically in everyday digital work.

One of the simplest and most useful ways I use AI is for social media captions. If you create TikToks, Instagram Reels, or image posts, you already know that sometimes the hardest part is not the content itself — it is finding the right words to package it. You may have a great post ready, but then you sit there wondering what caption to write.

That is where tools like ChatGPT can help.

In this article, I want to share a simple method I recommend for generating better captions with AI — especially if you want something short, natural, and tailored to your content without sounding forced or cheesy.

Why captions are harder than people think

In Zambia and across Africa, more young people are building brands online, marketing businesses through social media, and using content to create opportunities. I have seen this firsthand through my own work in education technology, software, and digital products. Whether you are promoting a business, sharing your work, teaching something, or just trying to stay consistent online, captions matter.

A good caption can:

• Help people understand your content quickly
• Increase engagement
• Give your post personality
• Make your message more memorable
• Save your audience from scrolling past

But writing captions consistently is mentally draining. Sometimes you simply run out of ideas. That does not mean your creativity is gone. It just means you need a better system.

My simple AI method for TikTok and Instagram Reels

If I have a TikTok or Instagram Reel, the easiest approach is to take the script or the main message behind the content and paste it into ChatGPT. Then I give a direct instruction like:

Generate a caption for this Reel.

That is the basic version. But if you stop there, the result may feel too generic. AI often defaults to captions that sound overused, cliché, or too dramatic. That is why I recommend being specific about the style you want.

For example, I would refine the prompt with instructions such as:

• Do not be cheesy
• Do not be cliché
• Keep it short and simple
• Use emojis only if necessary

Those extra instructions make a big difference. Instead of getting one weak caption, you begin to shape the output toward your brand voice.

This is one of the biggest lessons I have learned as someone working in AI: the quality of the result depends heavily on the clarity of the instruction. Whether I am building AI-powered education tools like eskulu or using generative AI for content workflows, the principle is the same — better prompts lead to better outcomes.

Ask for volume, not just one idea

One trick I really like is asking for many options at once. Instead of saying “generate a caption,” I prefer something like:

Generate 40 caption ideas in 10 different styles.

This is powerful because it gives you range. You are no longer trapped with one AI-generated sentence. You now have multiple directions to choose from.

You might get captions that are:

• Minimal and clean
• Playful
• Bold and attention-grabbing
• Professional
• Curious and conversational
• Youthful and energetic

From there, you can pick one as it is, combine two ideas, or edit one to sound more like you. AI should not replace your judgment. It should give you a strong starting point.

This matters especially for creators, founders, and small businesses in African markets where time and resources are limited. If you are running your own brand, doing your own marketing, and building your own product, efficiency matters. I understand that personally because I have spent years building platforms like Zedpastpapers and eskulu while also balancing technical work, strategy, and communication.

How to use AI for image captions too

The same concept works for images.

If you are posting a photo on Instagram, Facebook, LinkedIn, or another platform, you can upload the image into an AI tool and simply say:

Caption this image.

Then, just like with video content, add style instructions. For example:

• Keep it natural
• Make it short
• Avoid sounding forced
• Give me multiple options

This is useful if you are posting:

• Personal brand photos
• Product images
• Event pictures
• Educational graphics
• Behind-the-scenes content

Instead of spending 20 minutes thinking of a sentence, you can get a list of usable options in seconds.

What makes AI captions actually good?

Not every AI-generated caption is worth posting. The best ones usually have a few qualities:

• They sound human
• They match the content
• They are concise
• They fit the platform
• They reflect your personality or brand

If the caption sounds like something everyone else is writing, it will not stand out. That is why I always recommend telling the AI what to avoid, not just what to do.

For example, saying “do not be cliché” can be just as important as saying “make it engaging.”

This is also where your own voice matters. AI can help you move faster, but authenticity still comes from you. In my own journey, whether I was building educational tools during COVID-19, growing platforms used by hundreds of thousands of learners, or representing innovation from Zambia in rooms that included national leaders and investors, I have learned that technology works best when it amplifies a real human mission.

Why this matters for African creators and businesses

Across Zambia and the continent, AI is no longer something distant. It is becoming a practical tool for students, entrepreneurs, educators, marketers, and developers. I have seen that shift through my own work as an AI engineer, and it is one reason I remain deeply committed to building useful AI solutions locally.

When I built eskulu, the goal was not to use AI for hype. It was to solve real learning problems for Zambian students. That same mindset applies here. Using AI for captions may seem small compared to building a full platform, but it is still part of a bigger movement: helping Africans use technology to work smarter, communicate better, and create more opportunities.

Practical AI use cases matter because they lower barriers. A student entrepreneur in Lusaka, a small business owner in Kitwe, or a creator in Nairobi should be able to use these tools to improve their output without needing a full media team.

That is the future I believe in — one where AI is not just imported technology, but something we actively apply to local challenges and local ambition.

My final advice

If you are struggling with social media captions, do not overcomplicate it.

Start with these steps:

• Paste your Reel or TikTok script into ChatGPT
• Ask it to generate a caption
• Tell it what style to avoid, especially cheesy or cliché wording
• Ask for short, simple options
• Request many variations, such as 40 ideas in different styles
• For images, upload the image and ask for caption ideas the same way

The goal is not to let AI think for you. The goal is to use AI to remove friction so your creativity can move faster.

I have spent years self-teaching, building, and experimenting with technology — from coding as a student to winning innovation recognition like the X Pitchathon Business With a Purpose award, and continuing to deepen my AI skills through certifications such as AWS Lambda Foundations. One thing I know for sure is that simple tools, used well, can create a real advantage.

If you are a creator, founder, school, or business looking to integrate AI into your workflow, I would love to connect. You can also explore what I am building through eskulu and the wider work we are doing at ZOEC to make AI more useful in Zambia.

For AI consulting, software development, or collaboration, email me at jeffmdala@gmail.com.