Latest / The Side Hustle Career Podcast with Fexingo: Second Income, Freelancing While Employed / How a Part-Time Freelancer Built a 5000 Monthly Side Hustle in AI Prompt Engineering
Transcript
- Lucas: You know that moment when someone says 'just learn to code' and you think, 'okay, but that's a two-year commitment and I have a day job'? Luna: Every single time. Lucas: Well, there's a new kind of technical skill floating around — one that doesn't require a computer science degree, and it's actually paying pretty well as a side hustle. I'm talking about AI prompt engineering. Luna: Prompt engineering. So, crafting the right questions or instructions for generative AI tools — like ChatGPT or Midjourney — to get useful outputs. Lucas: Exactly. And it sounds almost too simple, right? But the market is real. A friend of mine, Maya Chen — she's a marketing coordinator in Chicago — started freelancing as a prompt engineer back in early 2025. Within six months, she was billing about five thousand dollars a month on the side. Luna: Five thousand a month? That's not pocket change. How did she get into it? Lucas: It started because her day job asked her to test ChatGPT for writing social media captions. She got curious, started reading forums, and realized companies were actually hiring people just to write prompts. So she opened an Upwork profile, pitched herself as a 'generative AI specialist,' and landed her first gig — a small contract with a local e-commerce brand. They wanted a chatbot that could handle customer returns without sounding robotic. Luna: And she figured out how to prompt the AI to sound natural? Lucas: She spent a weekend iterating on prompts. Things like 'You are a friendly but efficient customer service agent for an online clothing store. When a customer wants to return an item, first ask for the order number, then explain the return policy in three bullet points. Keep tone warm and under fifty words.' That kind of specificity. The client loved it, paid her five hundred dollars, and she had her first portfolio piece. Luna: So the skill is really about understanding what the AI needs — how to phrase things to get consistent, high-quality results. Lucas: Exactly. It's like being a translator between the human and the machine. And the demand is growing fast. According to a report from Grand View Research, the generative AI market is expected to hit one hundred thirty-six billion dollars by 2030. Companies are scrambling to integrate these tools, but they often don't know how to use them effectively. That's where freelancers come in. Luna: So, what kind of projects did Maya end up taking on? Lucas: She did a lot of content creation prompts — blog outlines, email sequences, ad copy. But the highest-paying projects were for data analysis. One client needed a way to ask questions about their sales data in plain English. Maya built a set of prompts that turned a CSV file into a conversational dashboard. That project alone paid two thousand dollars. Luna: That's impressive. But I imagine you need some data literacy for that — knowing how to structure a query so the AI actually returns the right numbers. Lucas: Absolutely. But Maya's background is marketing, not data science. She learned by doing — testing prompts, seeing what broke, refining. She said the key was to think in constraints: be very specific about the format of the output, the tone, the length, and what to avoid. Like, 'Do not invent data. Only use the information provided.' Luna: That makes sense. So, what about the ethical side? I've heard concerns about AI replacing jobs, or prompts being used to generate misleading content. Lucas: Maya thought about that too. She decided early on to only take projects that were transparent and ethical. For example, she won't write prompts for generating fake reviews or impersonating real people. And she always makes sure the client understands that AI outputs should be reviewed by a human. That actually became part of her value proposition — she offers a 'human-in-the-loop' guarantee. Luna: That's smart positioning. It also protects her from liability. So, for someone listening who wants to try this, what's the first step? Lucas: Start with the free version of ChatGPT or Claude. Pick a task you know well — like writing a recipe blog post or answering common customer questions. Spend an hour crafting different prompts to see what changes the output. Then, create a small portfolio of your best prompts and results. Maya used a simple Google Doc at first. Luna: And then pitch on freelancing platforms? Lucas: Exactly. Upwork, Fiverr, even LinkedIn. Look for posts where people are asking for help with 'AI writing' or 'chatbot setup.' Aim for small, fixed-price projects at first. Maya's first project was a hundred and fifty dollars for ten test prompts. That built her confidence and her profile. Luna: What about pricing? How does someone set their rate? Lucas: It's still a new field, so rates vary wildly. Maya started at thirty dollars per hour, but after a few months, she raised it to seventy-five. She found that clients are often willing to pay more for someone who can deliver consistent, formatted outputs quickly. She also started charging per project — like five hundred for a complete chatbot script — which let her earn more per hour. Luna: So, the learning curve isn't steep, but the earning potential is real. That's refreshing. Lucas: Yeah, and it's a skill that actually complements other side hustles. If you're already doing content writing, or coding, or virtual assistance, adding prompt engineering can make you more efficient. Maya says she now uses her own prompts to automate half her marketing reporting at her day job. Luna: That's a nice feedback loop — learning on the side improves your day job. I think that's a big reason people find side hustles satisfying. Lucas: Speaking of which — if today's episode gave you a concrete idea or just made you think differently about your skills, that's exactly why we make this show. And the way we keep it ad-free is listener support. If you found value here, you can buy me a coffee dot com slash fexingo. It's a simple way to say 'this was useful, keep going.' No pressure, no perks — just helps us stay independent. Luna: Yeah, we really appreciate it. Now, back to prompt engineering — one thing I haven't asked: what about the risk of these platforms changing their models? Like, a prompt that works today might not work tomorrow. Lucas: That's a real challenge. Maya learned that the hard way when OpenAI updated GPT-4 and her carefully crafted prompts for a client's chatbot started producing gibberish. She had to rebuild from scratch. Her advice: document every prompt with a version date, and test regularly. Also, don't get too attached to a single tool — learn the quirks of at least two or three platforms. Luna: So, adaptability is key. That seems like a theme across many side hustles. Lucas: It is. And the nice thing about prompt engineering is that it's still early. There aren't a million experts yet. If you start now, you can establish yourself before the market gets crowded. Maya's already getting referrals from past clients. She's thinking about turning it into a full-time consultancy. Luna: That's the dream, right? A side hustle that grows into a main hustle. But even if it stays a side thing, five thousand a month is a solid second income. Lucas: Absolutely. And the barrier to entry is basically curiosity and a few hours of experimentation. If you're the kind of person who likes tweaking things until they work, this might be your niche. Luna: So, to sum up: pick a domain you know, learn to write specific prompts, build a small portfolio, and start pitching on freelancing sites. Charge per project, be ethical, and keep adapting as models change. Lucas: That's the blueprint. And if you try it, let us know how it goes. We love hearing stories from listeners who take action.