Before I ever had a paying client, I made money with AI in a completely different way: by generating images and video with a tool called Seedance 2.0, using a custom prompting assistant I got access to through a course I’d joined. I had no design background, no Photoshop skills, and no idea what I was doing at first. I just started typing descriptions and seeing what came back.
That’s really the whole secret to this. You don’t need to know how to draw. You need to know how to describe.
What “generating an image” actually means
You type a description of what you want to see — this is called a prompt — and an AI model creates an image based on it. No sketching, no software to learn. The skill that matters here isn’t artistic, it’s descriptive.
Step 1: Pick a tool to start with
There are several good options for beginners:
- DALL-E (built into ChatGPT Plus) — easiest if you’re already using ChatGPT
- Midjourney — known for the most polished, artistic results, but has a slightly steeper learning curve
- Free options — several tools let you generate a handful of images for free before asking you to pay, which is a good way to test the waters
You don’t need to pick “the best one.” Pick the one that’s easiest for you to access right now, and start there.
Step 2: Write your first prompt
A good beginner prompt has three parts: subject, style, and detail.
Weak prompt: “A cat”
Better prompt: “A fluffy orange cat sitting on a windowsill, soft morning light, photorealistic style”
Notice what changed: you told it what, how it should look, and the lighting/mood. That’s the difference between a flat, generic result and something that actually looks intentional.
Step 3: Learn from what comes back
Your first image almost never comes out exactly how you imagined it — and that’s normal, not a failure. Look at what’s off, and adjust your next prompt:
- Wrong style? Add a specific reference (“in the style of a magazine editorial photo”).
- Composition feels off? Describe the camera angle (“close-up shot”, “wide angle”, “from above”).
- Too plain? Add mood and lighting words (“golden hour”, “dramatic shadows”, “soft and dreamy”).
This back-and-forth is the actual skill. Nobody writes a perfect prompt on the first try — they refine it.
A tool that genuinely leveled me up
This is the part of my own story I mentioned at the start: the custom prompting assistant I use for Seedance 2.0 (a tool built specifically to help you write image and video prompts that actually work) came from a course called AiMUSE Alchemy. It’s a paid course, so I want to be upfront about that, but it’s the single thing that took my results from “okay” to genuinely good, faster than trial and error ever would have. If you want to skip months of guessing, AiMUSE Alchemy is where I’d point you.
(This is an affiliate link — if you join through it, I may earn a small commission at no extra cost to you.)
What to try this week
Generate five images. Not one — five. Use the same basic idea each time, but change one thing per attempt (the lighting, the angle, the style). This is the fastest way to actually feel how prompts affect results, instead of just reading about it.
Look what’s possible
Just to show you how far this can go once you get comfortable — here are a few AI-generated images to give you a sense of the range, from portraits to full concept art.
Next up: What Is an AI Avatar? Beginner’s Guide — from still images to a face that can actually talk.
Leave a Reply