How This Side Hustler Makes 8000 Monthly as a Freelance AI Prompt Engineer

Lucas: So you've probably heard the term 'prompt engineer' floating around — it sounds like one of those job titles that didn't exist three years ago. Luna: Right. And honestly, I used to think it was just another buzzword. But then I met someone who's making eight thousand dollars a month doing exactly that as a side hustle. Lucas: That's actually what we're digging into today. Her name is Sarah — she was a middle school English teacher before she started freelancing as an AI prompt engineer. She now works with about six clients, mostly small law firms and medical coding companies. Luna: Wait — medical coding? That's pretty niche. How does prompt engineering apply there? Lucas: Great question. So Sarah builds prompts that help these companies automate parts of their workflow. For the law firms, she writes prompts that summarize depositions or extract clauses from contracts. For the medical coding side, she crafts prompts that translate doctor's notes into standardized billing codes. Luna: That's fascinating because — let's be honest — most people think prompt engineering is just asking ChatGPT to write a poem or a haiku. Lucas: Exactly. The real skill is understanding how to structure a query so the model produces consistent, accurate, and useful output every time. It's less about the AI and more about the logic and domain knowledge you bring. Luna: Sarah's background as a teacher — that seems oddly relevant. Teachers are trained to break down complex instructions for different audiences. Lucas: That's exactly what she told me. She said teaching taught her how to be precise with language, anticipate misunderstandings, and iterate on explanations. Those are the core skills of prompt engineering. She started by taking a few online courses — nothing expensive, some free YouTube tutorials — and then offered to do free prompt audits for small businesses. Luna: A free audit — so she'd look at how they were currently using AI and suggest better prompts? Lucas: Exactly. She'd say, 'Show me what you're asking your AI tool, and I'll show you how to get better results.' That landed her first three clients. One was a real estate agent who was using ChatGPT to write listing descriptions but getting generic copy. Sarah rewrote the prompt to include the property's unique features, the neighborhood vibe, and the target buyer's emotional triggers. The agent's listings started getting more inquiries. Luna: That's a concrete win. So how did she scale from there? Luna: She raised her rates. She started at fifty dollars an hour, then moved to seventy-five, then to a hundred. Now she does monthly retainers — three thousand dollars for a set number of prompt iterations and updates. With two or three retainers plus some project work, she hits eight thousand. Lucas: And she's picky about clients. She only works with businesses where the prompt has a measurable impact — like turnaround time saved or error rate reduced. That makes it easy to justify the cost. Luna: You know, this episode is exactly the kind of practical breakdown that makes this show valuable. If today was actually useful to you, the way we keep these conversations ad-free is through listener support — buy me a coffee dot com slash fexingo. Lucas: Yeah, it's a small way to keep the lights on and helps us keep bringing stories like Sarah's. And we really appreciate it. Luna: Absolutely. So back to Sarah — she also told me that one of the biggest misconceptions is that you need to be a coder. Lucas: Right. She doesn't write a single line of code. It's all natural language. But she does need to understand the model's limitations — like token limits, temperature settings, and how different models handle ambiguity. Luna: That's interesting — the technical part is more about knowing which parameters to tweak. Lucas: Exactly. For example, when she's working on a medical coding prompt, she sets the temperature very low — close to zero — because you want deterministic output, not creative guesses. But when she's writing prompts for marketing copy, she might set it a bit higher to get variety. Luna: So the same skill set applied to different domains with different constraints. Lucas: Exactly. And she says the key is to test relentlessly. She maintains a library of hundreds of test cases for each client. Every time she updates a prompt, she runs it against the test cases to make sure it doesn't break anything. Luna: That sounds like software testing, but with words. Lucas: It really is. And that's actually a great entry point — if you're detail-oriented and enjoy debugging why a sentence gets misinterpreted, this could be a natural fit. Luna: What about tools? Does she use specific platforms? Lucas: She uses a mix — ChatGPT for most text tasks, Claude for longer documents, and Midjourney for image prompts when a client needs visual content. She also uses a tool called PromptLayer to track version history and performance. Luna: Version control for prompts — that's becoming its own little ecosystem. Lucas: It really is. And she says the demand is only growing. In the first half of 2026, she's seen more requests from healthcare and legal — industries that are late adopters of AI but now realize they need experts to make it work reliably. Luna: So for someone listening who wants to start — what's the first step? Lucas: Sarah's advice: pick one domain you already know something about — it could be fitness, real estate, education, whatever — and start building a portfolio of prompts for common tasks in that field. Then offer to help a small business for free in exchange for a testimonial. Luna: That's the same playbook as any freelance niche — specialize, prove value, then charge. Lucas: Exactly. And don't underestimate the power of having a niche. General prompt engineers are a dime a dozen. But someone who can write prompts specifically for, say, dental practice management software? That's rare and valuable. Luna: Sarah sounds like she's really carved out a smart position for herself. Lucas: She has. And she's now thinking about creating a course to teach other teachers how to transition into prompt engineering. She sees it as a way to leverage their existing skills without starting from scratch. Luna: That's a great mission. And it really underscores the idea that the best side hustles often come from combining an existing strength with a new market need. Lucas: Exactly. AI isn't replacing jobs — it's creating new ones for people who can communicate with it effectively. And that's a skill almost anyone can learn.