The first time I heard the term prompt engineering I thought it was an engineering topic. It seemed technical, full of parameters and syntax. Something that required knowing how to code. I was completely wrong.
Prompt engineering is communication. It is the art of explaining what you want in a way that a machine understands and produces what you need. And that is exactly what designers do every day. Every creative brief we write is a prompt for a human. The only difference is that now the executor is an AI.
The creative brief is a prompt
Think about how you explain a project to a junior designer. You do not say make me something pretty. You say: I need a landing page for a fintech targeting millennials, the tone is professional but approachable, colors should convey trust, the call to action must be clear and prominent. That is a prompt. It has subject, style, context, constraints, and objective.
As I explained in my article about prompts, the structure is the same. What changes is the receiver. And it turns out machines need the same clarity as humans, just with less tolerance for ambiguity.
Techniques that work
Chain of thought: instead of asking for the final result, I ask for intermediate steps. Not design me a dashboard. First analyze what data a retail investor needs. Then prioritize by usage frequency. Then suggest a visual hierarchy. AI produces better results when it thinks before executing. Just like a designer.
Few-shot prompting: giving examples. If I want a specific style in Midjourney, I do not describe it with adjectives. I show references. The equivalent of a mood board. AI learns from example more than from abstract description.
Negative prompting: saying what you do not want is as important as saying what you do want. Not generic, not corporate, not stock photo feeling. These constraints are filters that guide AI in the right direction.
Progressive iteration: I never expect the first prompt to be perfect. I generate, evaluate, adjust. It is the same iteration process we use in design. The difference is each cycle takes seconds instead of hours.
Why designers have an advantage
We think visually. We can imagine the result before describing it. That gives us an enormous advantage because we know how to evaluate whether what AI generated is close to what we had in mind. A programmer using AI to generate images often accepts the first result because they lack a trained visual criterion.
We understand hierarchies. We know what is most important and what is secondary. We can structure a prompt with the same logic we structure an interface. Essentials first, then details.
We handle constraints. Our entire career is working within limitations. Screen sizes, brand guidelines, accessibility requirements. Prompts are exactly that: communicating a vision within technical constraints.
Prompt engineering is not a new skill for designers. It is a natural extension of what we already know how to do.