Prompting Ain’t Plug-and-Play: Why AI Still Needs an Artist

A deeper dive into prompt engineering as a creative discipline

*WARNING* — Controversial take, but true:
Critics often say AI outputs feel soulless. That they’re cold. Generic. That no real artist would create something like that.

And many times… they’re right.

But that’s not the model’s fault. It’s the prompt.

Because the spark isn’t in the machine — it’s in the idea… it’s in YOU. In the way YOU shape it. It’s in the human behind the keys.

Prompting is where the soul shows up.

That’s the power of being a creative director in the age of AI. You’re not just asking for results. You’re guiding them. You’re making judgment calls. You’re embedding taste.

We’ve been remixing influence from other’s work for centuries. AI just compresses it.

When used with intention, AI becomes your collaborator. Your translator. Your amplifier.

There’s a reason your results don’t come out the way you expected.

We keep hearing how easy it is to use AI. Just type what you want and out it comes, right?

Until it doesn’t. Until it sounds off. Or flat. Or just weirdly not you.

That’s because prompting isn’t just a button press. It’s a skill.

And the further you go in your creative work, the more that skill looks less like coding and more like directing.

This article is for anyone who’s ever typed a prompt and thought: why doesn’t this sound right?

We’ll break down what good prompts are made of, explain the jargon along the way, and show how creative prompting is becoming its own craft.

Prompting Is a Creative Skillset

Think of a prompt like a director’s note. You’re not just asking the actor to say the line. You’re setting mood. Context. Intention.

AI works the same way. If you want meaningful output, you need meaningful guidance.

Prompt vs. Completion
A prompt is your input. The completion is the AI’s output.
Better input = better result.

Great prompting isn’t just about what you say. It’s about how you say it.

Four Elements of a Strong Prompt

1. Intentionality

Know your goal. Are you drafting from scratch? Editing something? Exploring a new tone? Get clear.

2. Specificity

Don’t say “write a story.” Say: “Write a 2-paragraph story about a retired rodeo clown who misses the spotlight.”

3. Structure

Structure gives your AI output a skeleton to follow — the “bones” of your piece — so you can come in later and fill out the flesh. Use bullets or numbered steps to guide it clearly.

Pro Tip: Break your prompt into bullet points or ordered steps for better clarity.

Example:

Write a blog post that:
- Introduces the problem
- Explains why AI is a good solution
- Gives 3 use cases
- Ends with a casual call to action

4. Tone Control

Tell it the vibe you want.

“Write this in the voice of a charming but slightly bitter country songwriter.”

Note: I would avoid naming specific artists or celebrities in your prompts. Not only is it a legal gray area, but IMO, that’s where your idea loses its soul. Instead of saying “make this sound like Jimmy Buffett,” describe the vibe musically: “tropical-country with mellow guitars, vibraphone, and laid-back vocals.”

What the Heck Is a Delimiter?

What’s a Delimiter?
A delimiter is just a boundary. It tells the AI where sections begin and end.

Example: use """ or --- to separate instructions from examples.

This keeps your prompt clean and avoids confusion.

Delimiters help your structure stay readable — and more importantly, understandable to the model.

Write a poem in the voice of a cowboy reflecting on lost love.
"""
Example:
The fire’s gone cold where we once sat tight,
Now it’s just ashes in the pale moonlight.
"""

The triple quotes tell the model that everything between them is a reference or example — not part of the instruction.

Before → After: Prompt Upgrades

Let’s take a vague prompt and make it better.

“Write a blog post about AI.”

Too broad. No direction.

“Write a 3-paragraph blog post in a casual, witty tone explaining how AI can help musicians write lyrics. Include an example chorus.”

Clear, scoped, with an example. That’s a solid prompt.

Prompting Is Iterative

You won’t get it perfect the first time. And that’s fine.

Prompting is creative drafting. You try something. You edit. You rerun. You refine.

You don’t send the first sketch to the printer — why would you use the first prompt?

Try rewriting one of your last prompts using the tools above — you might be surprised how much closer it sounds to you.

Tooling, Hacks & Habits

A few tricks from regular prompt writers:

• System Messages: Use them to set persistent context.

“You are a blunt music critic. Be honest and detailed.”

• Save What Works: Build a prompt notebook, spreadsheet, or use a tool like PromptLocker.
• Use Examples: Show the AI what “good” looks like. Paste in samples with delimiters.
• Prompt-Tune in SLMs: For the truly committed, tune a small model to learn your prompt patterns.

Final Thought

Prompting isn’t about stealing ideas or cutting corners. It’s about translating creative intent into structured action.

The model is your instrument. And the prompt? That’s your sheet music. You are the origin of the idea.

When you prompt with clarity, context, and conviction — the results don’t just echo your words. They echo your voice.

Because when the prompt sounds like you?

So will the output.