Use Cases

AI is best as a helpful teammate: it can draft, organize, and explain—fast. This page shows practical ways to use it across work and learning, with ready-to-copy prompts and simple safety rules.

Pick a category

Work

Use AI to turn messy thoughts into clear messages. It’s great for drafting and structuring, as long as you keep sensitive details out and verify anything factual.

Good for

  • Email drafts that match your tone (friendly, direct, formal).
  • Turning meeting notes into action items with owners and deadlines.
  • Creating first-pass slides or talk tracks for presentations.
  • Writing stakeholder updates that are short and scannable.

Be careful about

  • Don’t paste secrets, client data, or internal-only documents; anonymize details.
  • Verify names, dates, numbers, and commitments—AI can “sound right” but be wrong.
  • Watch for overconfident tone; ask it to flag assumptions explicitly.

Ready-to-use prompts

Rewrite an email (clear + kind)
You are my communication assistant.

Context: I’m emailing [audience] about [topic]. The goal is [goal].

Constraints:
- Keep it under [word count] words.
- Use a [tone] tone.
- Include a clear ask and a deadline: [deadline].
- Avoid jargon.

Output format:
1) Subject line
2) Email body
3) One-sentence follow-up I can send if they don’t reply

Draft to rewrite:
[PASTE DRAFT HERE]

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Meeting notes → actions
You are my project coordinator.

Input: These meeting notes are rough and possibly out of order:
[PASTE NOTES]

Task:
- Extract decisions, open questions, and action items.
- For each action item, suggest an owner role (not a person) and a realistic due date.

Output format:
- Decisions (bullets)
- Action items (table with: Item | Owner role | Due date | Risk/Blocker)
- Open questions (bullets)

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Study

AI can help you learn faster by summarizing, drilling, and explaining ideas in different ways. It works best when you provide the source content and ask for checks like “quiz me” or “show steps.”

Good for

  • Summaries that keep key definitions and examples.
  • Flashcards for spaced repetition.
  • Practice quizzes with explanations for wrong answers.
  • Tutoring-style explanations tailored to your level.

Be careful about

  • Don’t trust citations or “facts” without checking the original source.
  • For math/science, verify each step—small errors can cascade.
  • Avoid sharing personal student IDs, private grades, or exam materials that are not allowed.

Ready-to-use prompts

Notes → flashcards
You are my study coach.

Topic: [topic]
Audience level: [high school / undergrad / beginner / advanced]

Source notes:
[PASTE NOTES]

Create:
- 12 flashcards (front/back)
- Include 3 “concept” cards, 6 “definition” cards, 3 “example” cards
- For any formula, explain what each symbol means

Output format:
JSON array with fields: type, front, back

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Practice quiz (with feedback)
You are my tutor.

Topic: [topic]
Constraints: [time limit], focus on [subtopics], avoid [subtopics]

Create a 10-question quiz:
- 6 multiple choice
- 2 short answer
- 2 scenario questions

Then wait for my answers.
After I respond, grade them and explain:
- Why the correct answer is correct
- Why my answer is wrong (if wrong)
- What I should review next

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Programming

AI is great at narrowing down bugs, suggesting refactors, and generating tests—especially when you provide the error, expected behavior, and minimal code. Treat outputs as a starting point, not truth.

Good for

  • Debugging with a clear reproduction and error logs.
  • Code review checklists (naming, edge cases, performance).
  • Refactoring toward clearer functions and better types.
  • Writing unit tests and thinking through tricky cases.

Be careful about

  • Never paste API keys, tokens, private repo code, or customer data.
  • Watch for hallucinated APIs or outdated library usage—confirm with docs.
  • Validate security-sensitive changes (auth, payments, input validation) with extra care.

Ready-to-use prompts

Debug with a minimal repro
You are a senior software engineer.

Problem: [what is broken]
Expected behavior: [what should happen]
Actual behavior: [what happens]

Environment:
- Framework: [Next.js / React / Node]
- Language: [TypeScript]
- Relevant versions: [versions]

Here is the smallest code that reproduces the issue:
```ts
[PASTE MINIMAL CODE]
```

Error/Logs:
```
[PASTE ERROR]
```

Tasks:
1) List the top 3 likely root causes (with reasoning).
2) Suggest a fix for each.
3) Tell me what to log or test to confirm the real cause.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Write tests first (edge cases)
You are my testing assistant.

Function/component to test:
[PASTE CODE]

Requirements:
- Testing framework: [Jest/Vitest]
- Include edge cases and failure cases
- Prefer table-driven tests when possible

Output:
- Test file code
- Brief notes on what each test protects against

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Design / Creative

AI can brainstorm options quickly and help you improve clarity. It’s strongest when you give it constraints (voice, audience, style) and ask for multiple directions, not one perfect answer.

Good for

  • Brainstorming names, concepts, and campaign angles.
  • Copywriting variations for headlines and CTAs.
  • Drafting image prompts for different styles.
  • UI critique checklists (hierarchy, clarity, accessibility).

Be careful about

  • Be careful with copyrighted brands, artwork styles, or “clone this” requests.
  • Don’t present generated content as verified facts or official policy.
  • For UI critique, validate against real user needs and analytics—not only opinions.

Ready-to-use prompts

Brainstorm 10 options (with constraints)
You are a creative director.

Goal: Generate ideas for [project/topic].
Audience: [audience]
Constraints: [brand voice], must avoid [taboo words], must fit in [character limit].

Deliver 10 distinct options.
For each option include:
- The idea
- Why it might work
- One risk or downside

Output format:
Numbered list.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

UI critique (actionable)
You are a senior product designer.

Screen/context: [describe the screen and user goal]
Users: [who the users are]
Constraints: [platform], must support [accessibility requirement], primary metric is [metric].

Here’s the UI text/content:
[PASTE COPY]

Task:
- Identify the top 5 clarity or hierarchy issues.
- Suggest specific fixes (rewrite, layout changes, component changes).
- Provide a quick accessibility check (contrast, labels, focus order) as a checklist.

Output format:
Sections: Issues, Fixes, Accessibility checklist.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Productivity

AI helps when you feel stuck: it can break down vague goals into small next steps. It’s also great for planning templates—just keep the plan realistic and measurable.

Good for

  • Weekly planning with a time budget and priorities.
  • Checklists for recurring tasks (launches, travel, renewals).
  • Habit tracking ideas and gentle reminders.
  • Decision matrices when choices feel emotionally noisy.

Be careful about

  • Don’t let the plan become busywork—ask for the smallest next step.
  • Watch for over-optimization; real life needs slack time.
  • If it suggests medical/legal/financial advice, treat it as general info and verify.

Ready-to-use prompts

Plan my week (realistic)
You are my productivity coach.

My priorities this week:
1) [priority 1]
2) [priority 2]
3) [priority 3]

Constraints:
- I have [hours] hours available.
- I need buffers for [meetings/commute/family].
- I want 2 deep-work blocks of [duration] each.

Output format:
- A simple weekly plan (Mon–Fri)
- Daily top 3 tasks
- One “if I fall behind” recovery plan

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Decision matrix (simple and honest)
Help me choose between these options: [option A], [option B], [option C].

Context: [what decision is for]
Constraints: [budget/time], must-have: [must-haves]

Create a decision matrix:
- Criteria (5–7) with weights (sum to 100)
- Score each option (1–10) with short justification
- Final recommendation

Also ask me 3 clarifying questions if anything important is missing.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Customer Support / Operations

AI is useful for turning repeated questions into consistent answers and SOPs. It can also rewrite tone (calm, helpful, firm) so customers feel respected—while you stay on policy.

Good for

  • Drafting FAQ articles and help-center responses.
  • Creating internal SOP checklists for new teammates.
  • Rewriting replies to match brand tone and reduce back-and-forth.
  • Summarizing ticket themes and proposing improvements.

Be careful about

  • Never include private customer data; redact names, IDs, and order details.
  • Don’t let it invent policy—always align with your real rules and approvals.
  • For refunds/charges/legal wording, verify with the source policy before sending.

Ready-to-use prompts

FAQ draft (from real tickets)
You are a support knowledge base writer.

Product: [product]
Audience: [new users / power users]
Tone: [friendly, concise]

Here are common questions (sanitized):
[PASTE QUESTIONS]

Create one FAQ article:
- Short intro
- 6–10 Q&As
- Each answer: 2–5 sentences, includes a clear next step

Output format:
Markdown with headings.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

Rewrite a reply (tone + policy-safe)
You are my customer support assistant.

Policy constraints (must follow):
[PASTE POLICY BULLETS]

Customer message (sanitized):
[PASTE MESSAGE]

Task:
- Write a reply in a [tone] tone.
- Be empathetic but clear about what we can/can’t do.
- Ask at most 2 questions.

Output format:
One message ready to send.

Tip: Replace placeholders like [topic] and keep any sensitive info anonymized.

How to get better results

When your prompt feels “meh”, don’t write more—write clearer. Use this simple structure and you’ll get more consistent outputs.

  1. 1Role: Tell the AI who it should act as (e.g., “senior engineer”, “study coach”).
  2. 2Context: Share what's happening and what success looks like.
  3. 3Constraints: Add limits (length, tone, tools, deadlines, must/avoid).
  4. 4Output format: Ask for a structure (bullets, table, JSON, steps).
  5. 5Examples: Provide a small example or a “good vs bad” reference if you can.
Before
Summarize this and tell me what to do next:
[PASTE TEXT]
After
You are my executive assistant.

Context: This is a project update about [topic]. I need to brief [audience] in 2 minutes.

Constraints:
- Keep it under 120 words.
- Highlight risks, blockers, and decisions needed.
- If facts are unclear, write “Unverified” and list what to confirm.

Output format:
- 3-bullet summary
- Risks (bullets)
- Next actions (table: Action | Owner role | Due date)

Text:
[PASTE TEXT]
Notice how the “after” prompt makes the job, limits, and output format explicit—this reduces guesswork and improves consistency.

Safety & privacy quick rules

These rules keep you out of trouble while still getting strong results.

  • Don't share secrets: passwords, API keys, private links, or confidential documents.
  • Anonymize sensitive data: remove names, emails, IDs, and exact financial figures when possible.
  • Verify important facts: dates, numbers, policies, and anything you would bet your job on.
  • Ask it to label uncertainty: “state assumptions” and “flag what needs confirmation.”
  • Prefer primary sources: paste the reference text you want it to use, instead of asking it to “look it up.”
  • Be careful with regulated topics (medical/legal/financial): treat outputs as general guidance, not professional advice.
  • Keep an audit trail: save the final prompt and output for work that needs accountability.