Effective Prompt Writing

Summary

Text-based generative AI systems such as Google Gemini and M365 Copilot process user inputs (prompts) using natural language processing algorithms and generate predictive text based on what they have learned from training data. These tools can be used for a variety of tasks, such as content creation, summarization, question answering, coding, translation, and creative writing.

AI systems are not “intelligent” in the human sense; they operate based on statistical patterns. Therefore, outputs should always be reviewed by the user before being used directly.

Writing Effective Prompts – Principles, Explanations, and Practical Examples

The output of a text-based generative AI tool largely depends on the quality of the prompt provided. Clear, contextual, specific, and well-structured prompts yield more accurate, reliable, and useful responses. Below are key principles of effective prompt writing, explained with examples:

1. Be Clear and Specific

AI tools cannot infer or guess meaning like humans. The clearer and more direct the prompt, the more accurate and consistent the result will be.

✅ Good Example:
"Write a 200-word paragraph explaining Pavlov’s classical conditioning theory in simple language for undergraduate psychology students."

❌ Poor Example:
"Write about Pavlov."
→ Too general and lacks context; the answer may be vague, irrelevant, or overly broad.

2. Provide Context

Without knowing who the content is for, where it will be used, and why, AI may not generate ideal output. Providing context helps produce more relevant content.

✅ Good Example:
"Simplify the following text for a high school magazine audience: [text]"
"Write a case-based introductory paragraph for 3rd-year medical students studying clinical communication."

❌ Poor Example:
"Simplify the text."
→ Missing information about audience, purpose, and difficulty level.

3. Be Specific and Directive

Don’t just say what you want the AI to write about—also explain how it should write. Indicate the type of content, target audience, length, and structure.

✅ Good Example:
"Create a 250-word academic 'statement of purpose' highlighting research interests for a graduate school application."
"Write a brief and formal email reply addressed to digital marketing professionals."

❌ Poor Example:
"Write something for me."
→ The content type, tone, goal, and length are unclear.

4. Specify the Structure

Clarifying the desired format (e.g., bullet points, headings, paragraph structure, introduction-body-conclusion) results in cleaner, more usable content.

✅ Good Example:
"Write a short informational piece on the benefits of smart cities, with three separate paragraphs under different headings."
"Create a bulleted list of suggestions for improving cybersecurity awareness in universities."

❌ Poor Example:
"Write about cybersecurity."
→ The output may lack structure and be hard to apply directly.

5. Set Boundaries on Unwanted Content

Sometimes it's important to say not just what you want—but also what you don't. Limiting content types can improve quality and appropriateness.

✅ Good Example:
"Write a paragraph in academic tone, based on scholarly sources, with no exaggeration or personal opinions."
"Create a neutral introduction to AI applications, avoiding dramatic or speculative language."

❌ Poor Example:
"Write something creative about AI."
→ May result in off-topic or overly imaginative content.

6. Indicate Tone, Length, and Language

If a specific tone is required (e.g., academic, formal, conversational, humorous, technical), include that in the prompt. Also, clearly state word count if needed.

✅ Good Example:
"Write a 300-word academic-style analysis of the use of AI in higher education."
"Generate a two-paragraph blog post in an energetic, conversational tone for a young audience."

❌ Poor Example:
"Write an analysis."
→ Without target audience or tone, the content may be unfocused.

7. Use an Iterative Approach (Refine as You Go)

Your first prompt may not yield the perfect result. You can refine the response by rephrasing parts or giving further instructions.

Example Workflow:

Initial Prompt: "Write a short explanation of quantum computers."
Refinement: "Now rewrite that explanation using simpler terms, suitable for high school students."
Expansion: "Divide the explanation into three sections with subheadings: 1) Key Concepts, 2) Use Cases, 3) Current Developments."

This step-by-step approach helps align the content with your needs.

8. Develop Content Through Connected and Layered Prompts

Rather than asking for large outputs in one step, guide the AI through smaller, linked prompts for better accuracy.