Guides
Prompt Engineering Basics
Prompt engineering sounds like a job title. It’s mostly just: be specific about what you want.
Here are the principles that actually move the needle.
The four elements of a good prompt
- Role — who should the AI be?
- Task — what exactly do you want?
- Context — what should it know?
- Format — how should it respond?
You don’t need all four every time. But thinking through each one before you write your prompt will make it dramatically more effective.
Example: before and after
Before (vague):
Write a summary of this article.
After (specific):
Act as a copy editor. Summarize this article in 5 bullet points for
a busy executive who won't read the full piece. Focus on decisions
or actions implied by the content.
[article text]
The second version specifies role, task, context, and format. The output will be more useful on the first try.
The most important principle: iterate
Treat the first response as a starting point. Say what you’d change, and ask again:
- “Make this more concise.”
- “The tone is too formal — loosen it up.”
- “Add a specific example to point 3.”
Good prompting is a conversation, not a one-shot command.
When you’re stuck
If you’re not getting what you want, try one of these:
- Add a format constraint: “Answer in 3 bullet points” or “Respond in 2 sentences”
- Specify the audience: “Explain this to someone with no technical background”
- Ask for options: “Give me 3 different versions of this”
- Ask it to ask you: “What questions do you need answered to write the best version of this?”
The last one is underrated. Claude will tell you exactly what context it needs.