AI Prompting Playbook

Prompt Smarter.
Get Better Results.

Generative AI is powerful, but the quality of its response depends heavily on how we communicate with it. This playbook explores five practical principles for creating clearer, more useful, and more reliable prompts.

Explore the Playbook ↓

From Vague Requests to Strategic Prompts

Effective prompting is not simply asking AI a question. It means providing direction, defining the desired output, showing examples, evaluating the response, and breaking complex work into manageable steps.

01

Give Direction

Tell the AI its role, audience, purpose, and context.

A vague request forces the AI to guess what you want. Providing context gives the model a clearer target and helps produce a response that fits your actual situation.

✕ Weak Prompt

"Explain budgeting."

✓ Improved Prompt

"Act as a financial literacy tutor. Explain personal budgeting to a first-year college student using simple language and a realistic student example."

Why it works: The improved prompt establishes a role, audience, topic, and context, reducing guesswork and making the response more relevant.
02

Specify Format

Tell the AI exactly what the final response should look like.

Even when the topic is clear, the AI may return the information in an inconvenient format. Specify the structure, length, and organization you need.

✕ Weak Prompt

"Give me information about cybersecurity."

✓ Improved Prompt

"Explain five common cybersecurity threats in a table with three columns: threat, description, and one prevention method. Keep each description under 30 words."

Why it works: The AI knows the exact number of items, structure, categories, and length expected in the response.
03

Provide Examples

Show the AI what a successful response looks like.

Examples communicate style and structure more clearly than vague adjectives such as "professional" or "interesting."

✕ Weak Prompt

"Write a discussion question about this reading."

✓ Improved Prompt

"Write a discussion question about this reading in the style of this example: 'How does the author's definition of fairness shape their argument?'"

Why it works: The example gives the AI a concrete model to follow and reduces uncertainty about the desired style.
04

Evaluate Quality

Never assume the first AI response is automatically correct.

AI can generate convincing responses that still contain weak reasoning, missing context, or incorrect information. Define what quality means and review the result.

✕ Weak Prompt

"Is this thesis statement good?"

✓ Improved Prompt

"Evaluate this thesis statement for clarity, originality, and arguability on a scale of 1–5. Explain each score, identify weak wording, and suggest one stronger revision."

Why it works: The improved prompt establishes measurable criteria. Important claims should still be verified using reliable sources instead of relying blindly on AI.
05

Divide the Labor

Break complex work into smaller, manageable steps.

Large requests can produce shallow or disorganized answers. Separating the work into stages makes each response easier to evaluate while keeping the user responsible for important decisions.

✕ Weak Prompt

"Write my 10-page paper on renewable energy."

✓ Improved Prompt

"Help me develop a paper on renewable energy policy in four steps. First, suggest five possible research questions. After I choose one, help me create an outline. Then suggest types of sources I should research. I will write the draft myself and ask you to review it for clarity."

Why it works: The task becomes a sequence instead of one massive request. AI supports brainstorming and organization while the student remains responsible for research, judgment, and writing.

The Prompt Formula

Before pressing Enter, think through these five elements.

01 Direction
02 Format
03 Examples
04 Quality
05 Labor

Test Your Prompt

Check each item before submitting an important prompt.

Prompt readiness: 0%

Key Takeaways

  1. Be specific. Clear direction beats vague wording.
  2. Design the output. Tell the AI what shape the response should take.
  3. Show, don't just tell. Examples communicate expectations clearly.
  4. Stay skeptical. Verify important information instead of blindly trusting AI.
  5. Stay in charge. Use AI to support your thinking, not replace it.