Opinion2 min read
You do not need to learn prompt engineering
You need to learn to brief. It is the same skill you use with a colleague.
"Prompt engineering" sounds like a technical job. It is not something you need to study before AI is useful at work.
What you need is simpler. You need to brief the AI the way you would brief a new colleague. Most people already know how to do that.
What a good brief has
Think of the last time you handed work to someone. A good hand-off had four parts:
- The goal. What the work is for.
- The context. What the person needs to know.
- An example. What good looks like.
- The format. How you want to get it back.
A good prompt has the same four parts. Here is a weak one:
Write an email about the delay.
And a stronger one:
Write a short email to a customer whose order is two days late. The goal is to keep their trust. They have ordered from us three times before. Keep it under 120 words, polite and direct, and offer one clear next step.
The second one is not clever. It is clear. That is the whole trick.
Why the "engineering" idea sticks
When AI tools first appeared, wording seemed to matter a great deal. Small changes gave very different answers. People shared special phrases as if they were spells.
Tools have improved since then. They follow plain instructions much better than they did. The special phrases matter less. Clear thinking matters more.
What to do instead
Next time you open an AI tool, do this:
- Say what the work is for, in one sentence.
- Paste in the background it needs.
- Show one example of an answer you like, if you have one.
- Say how long it should be and what shape it should take.
Then read the answer like you would read a colleague's draft. Ask: is this right? What is missing? Tell the AI what to change, and let it try again.
Where a bit of skill still helps
There is room to get better. Saving a brief that works means you do not start from nothing next time. That is a skill: a saved set of instructions that teaches AI to do one job your way. Learning to break a big task into steps helps too.
None of this needs a course. It needs practice and the habit of checking what you get back.
The point
If you can write a clear message to a colleague, you can use AI well. Start there. Learn more only when a real task needs it.
Want more practical notes like this? Read the latest on the News page or browse the issue archive.