1. Use Role-Based Prompts for More Targeted Insights

🔴 Mistake: Asking AI to generate responses without defining a specific role or expertise level.

👉 Example of a Weak Prompt:
"Write a report on AI in education."

👉 Why This Fails:

• AI has no guidance on perspective—should it write from the standpoint of a teacher, policymaker, or EdTech entrepreneur?
• The depth of analysis is unclear—is this for a beginner’s overview or an expert-level policy brief?
• The output may lack real-world application because there’s no context on how the report will be used.

✅ How to Fix It:
Use role-based prompting to guide AI toward the most relevant response.

👉 Improved Prompt Example:
"As an education policy advisor, draft a 1,500-word report on the impact of AI in secondary education. Include policy recommendations, ethical concerns, and examples from African and Asian contexts."

2. Structure Prompts for Step-by-Step Responses

🔴 Mistake: Expecting AI to generate a complex, multi-step process in a single unstructured response.

👉 Example of a Weak Prompt:
"How do I integrate AI into my consulting practice?"

👉 Why This Fails:

• AI lacks clarity on where to start—does the consultant need help with client communication, research, or business strategy?
• The response may be overwhelmingly broad, making it difficult to implement.

✅ How to Fix It:
Break the request into clear, step-by-step instructions.

👉 Improved Prompt Example:
"List the first five steps an education consultant should take to integrate AI into their workflow. Include specific AI tools, practical applications, and potential challenges."

3. Use Constraints to Get More Relevant AI Responses

🔴 Mistake: Giving AI too much freedom, resulting in outputs that are either too long, too complex, or irrelevant.

👉 Example of a Weak Prompt:
"Summarize the latest trends in AI for education."

👉 Why This Fails:

• AI might generate an unfocused summary covering many unrelated aspects (e.g., AI in assessments, AI in student engagement, AI in curriculum development).
• The level of detail is unclear—should the summary be bullet points, a one-paragraph overview, or a full-page report?

✅ How to Fix It:
Use word limits, formatting constraints, and specific focus areas to refine AI outputs.

👉 Improved Prompt Example:
"Summarize the top three AI trends in education for 2024 in 100 words. Focus on adaptive learning, AI-driven student support, and teacher AI-assisted grading."

4. Provide Reference Text for Improved Accuracy

🔴 Mistake: Relying on AI to generate content without providing it with high-quality reference materials.

👉 Example of a Weak Prompt:
"Explain the impact of AI on teacher training."

👉 Why This Fails:

• AI may fabricate information or pull from outdated knowledge.
• The response may lack alignment with recent studies or established frameworks.

✅ How to Fix It:
Feed AI relevant reference text and ask it to generate responses based

on that source material.

👉 Improved Prompt Example:
"Based on this UNESCO report on AI in teacher training (text provided below), summarize its key recommendations in 300 words for an audience of policymakers."

5. Use Iterative Refinement for Improved AI Outputs

🔴 Mistake: Using AI once and accepting the first draft, rather than iterating and refining for a better result.

👉 Example of a Weak Approach:

• Asking AI to write an entire funding proposal in one attempt and using it without reviewing or improving clarity.

✅ How to Fix It:

• Use sequential prompts to improve AI outputs.
• Ask AI to critique and refine its own response.

👉 Example of an Effective Iterative Process:
1️⃣ First Prompt: "Draft a one-page project proposal for an AI-based literacy program for rural schools."
2️⃣ Follow-up Prompt: "Make the proposal more concise and highlight key impact metrics."
3️⃣ Final Prompt: "Rephrase the conclusion to make it more persuasive for donors."