Your First AI Image.
Step by step, from blank prompt to finished visual. No experience needed.
After this lesson you'll know
- How to write your first image prompt from scratch
- The basic anatomy of a good prompt: subject, style, details, mood
- How to iterate quickly when the first result is not quite right
- How to save, download, and use your generated images
Let's use ChatGPT for this walkthrough. It is the simplest starting point.
Open chatgpt.com and sign in (a free account works for a few image generations). If you prefer another tool from Lesson 2, the principles are identical — only the interface changes. We are using ChatGPT here because you can type your prompt in natural language, just like having a conversation.
Your first prompt should be one clear sentence.
Resist the urge to write a novel. Start with something direct and visual. Here is your first prompt:
Look at what you got. Then ask: what would make it better?
Your first image will appear in seconds. It will probably look good — maybe even great. But now is when the real skill begins. Look at the image carefully. Is the dog the right breed? Is the sunset the color you imagined? Are there enough wildflowers? Is the mood right?
This is the core loop of AI image creation: generate, evaluate, refine. Every professional who uses these tools does exactly this. The magic is not in the first result. It is in knowing how to guide the next one.
Now layer in specifics to steer the image closer to your vision.
Take your original prompt and add details. Each detail gives the AI more information to work with:
Every great prompt has four parts.
Subject: What is in the image? (A golden retriever puppy)
Setting: Where is it? (In a field of purple and yellow wildflowers)
Mood/Lighting: How does it feel? (Golden hour, soft warm lighting)
Style: What does it look like? (Photograph style, shallow depth of field)
You do not need all four every time, but including more of them gives the AI a clearer picture of what you want. Think of it as painting with words — the more specific your words, the more specific the result.
The third or fourth version is usually the one you love.
Do not expect perfection on attempt one. Professional designers using these tools typically generate 3-10 variations before they find the one that works. Each iteration teaches you something about how the AI interprets language. Maybe "cinematic" gives you a different mood than "dramatic." Maybe "watercolor" gives you something softer than "painted." This is how you build your prompt vocabulary.
Try it now
Create three images using the same subject but different styles. Start with:
- "A cat sitting on a windowsill, photograph"
- "A cat sitting on a windowsill, oil painting style"
- "A cat sitting on a windowsill, Studio Ghibli anime style"
Notice how the same subject transforms completely based on the style keyword. Save all three — you just created your first intentional variation set. That is a real skill.
Download everything you like. Build a personal library from day one.
In ChatGPT, click the image to expand it, then use the download button. In Midjourney, upscale the image first, then save it. Create a folder on your computer called "AI Visuals" and start organizing by project or style. This library will become incredibly valuable as you develop your eye for what works.
The Four Prompt Elements
Prompt Anatomy
Tap one on the left, then its match on the right