The Complete Guide to Writing Prompts | Tips and Examples for Using ChatGPT
A practical guide to writing clearer ChatGPT prompts with context, constraints, examples, and an explicit output format.
2 min read
An AI response that misses the point often reflects missing context or an unclear instruction.
The key to mastering AI lies in how you write your prompts (instructions).
using ChatGPT as an example, I’ve summarized tips for designing effective prompts. Even beginners can put these into practice starting today.
Feature and specification statements were checked on July 31, 2026 against Prompt engineering. The article distinguishes what the official source confirms from editorial organization and recommendations.
What Is a Prompt?
A prompt is an “instruction” you give to an AI.
The quality of the response you receive can vary greatly depending on how you phrase your question or request.
3 Basic Steps for Writing Prompts
- Clarify Your Objective
- “Please translate this English email”
- “Please write some sample code in Python” → Stating the goal first makes it easier for the AI to understand the direction of the response.
- Specify Conditions and Assumptions
- “In a casual style”
- “Summarize it in 200 characters or less”
- “As a bulleted list, not a table” → If you leave things vague, you’re likely to get an answer that falls short of your expectations, so specifying constraints is key.
- Specify the output format
- “In Markdown format”
- “With YAML frontmatter”
- “Explain each step” → To get practical results, clearly specify the format.
Common Mistakes and Examples of Improvements
-
Example of a Mistake: “Write a blog post”
→ The content and tone are unclear and too abstract. -
Improved Example: “Write a blog post about ‘digital detox’ in Markdown format, aimed at beginners. Include front matter with the title, date, and tags, and structure it as Introduction → Headings → Summary.”
→ By specifying concrete conditions, you’ll end up with an article that’s ready to use right away.
Advanced Technique: Assigning Roles
You can further improve the accuracy of AI responses by assigning it a “role.”
- “You are a professional editor. Please improve the headings in this blog post.”
- “You are a front-end engineer. Please provide a code example using Next.js.”
→ Just this one phrase completely changes the direction of the response.
Summary
- The quality of the prompt determines the AI’s answer
- Specifying the goal, conditions, and format together improves accuracy
- Specifying a role helps elicit professional-quality output
In the next post, I plan to explain “Differences in Prompt Thinking Modes (GPT-5’s Instant/Balanced/Thinking).”
👉 Recommended Related Articles
Primary sources checked
Important claims should also link to the relevant source in the article body.
- Prompt engineeringOpenAI API Documentation · official-documentation · Checked: 2026-07-31