Who This Is For
Someone asking "what can this actually do?" Start by identifying their context.
| Signal |
They are... |
Load |
| "Just heard about AI", "is this useful for me?" |
Beginner |
contexts/beginner.md |
| Emails, reports, meetings, spreadsheets |
Office professional |
contexts/professional.md |
| Code, debugging, CI/CD, APIs |
Developer |
contexts/developer.md |
| Customer support, marketing, operations |
Business owner |
contexts/business.md |
| Writing, design, music, content |
Creative |
contexts/creative.md |
| "Show me something real", "not gimmicks" |
Skeptic |
contexts/skeptic.md |
High-Value Use Cases (Universal)
Communication:
- Draft emails matching tone to recipient
- Summarize long threads before meetings
- Write difficult messages (rejections, negotiations, complaints)
Information Processing:
- Summarize documents, articles, reports
- Extract key points from meeting recordings
- Explain complex topics in simple terms
Writing & Editing:
- First drafts from bullet points
- Proofreading with style preservation
- Translation with context awareness
Analysis:
- Make sense of spreadsheet data
- Compare options with pros/cons
- Research topics and synthesize findings
Learning:
- Explain concepts at your level
- Practice conversations (languages, interviews)
- Get feedback on your work
What AI Cannot Do
Be upfront about limitations:
- No real-time information — Data has a cutoff date
- No memory across sessions — Unless explicitly configured
- Can hallucinate — Always verify facts, especially names, dates, URLs
- No execution — Can't send emails, buy things, or take actions (unless integrated)
- No judgment — Final decisions remain yours
Addressing Common Concerns
"Is my data safe?"
→ Depends on the tool. Check privacy policy. Enterprise tiers often don't train on your data. Don't paste passwords or API keys.
"Will it replace my job?"
→ AI augments, not replaces. People who use AI outperform those who don't. Learn to work with it.
"How do I know it's accurate?"
→ Verify outputs. Use AI for drafts and ideas, not final truth. Cross-check important facts.
"Is using AI cheating?"
→ It's a tool. Spell-check didn't replace writers. The work is still yours if you direct, review, and refine.
Getting Started (For Beginners)
- Start small — One task you do weekly that's tedious
- Be specific — "Write a professional email declining a meeting" works better than "help me"
- Iterate — First response rarely perfect. Say "make it shorter" or "more formal"
- Verify — Don't send/use AI output without reviewing it
Discovery Questions
When someone doesn't know what to ask for:
- "What task do you dread doing every week?"
- "Where do you spend time that doesn't feel valuable?"
- "What would you do if you had an assistant who never slept?"
- "What's something you'd love to do but don't have time for?"
Their answers reveal high-value use cases.
1---2name: use-cases3description: Discover what AI agents can do for you with practical examples by role, task type, and skill level.4---56## Who This Is For78Someone asking "what can this actually do?" Start by identifying their context.910| Signal | They are... | Load |11|--------|-------------|------|12| "Just heard about AI", "is this useful for me?" | Beginner | `contexts/beginner.md` |13| Emails, reports, meetings, spreadsheets | Office professional | `contexts/professional.md` |14| Code, debugging, CI/CD, APIs | Developer | `contexts/developer.md` |15| Customer support, marketing, operations | Business owner | `contexts/business.md` |16| Writing, design, music, content | Creative | `contexts/creative.md` |17| "Show me something real", "not gimmicks" | Skeptic | `contexts/skeptic.md` |1819---2021## High-Value Use Cases (Universal)2223**Communication:**24- Draft emails matching tone to recipient25- Summarize long threads before meetings26- Write difficult messages (rejections, negotiations, complaints)2728**Information Processing:**29- Summarize documents, articles, reports30- Extract key points from meeting recordings31- Explain complex topics in simple terms3233**Writing & Editing:**34- First drafts from bullet points35- Proofreading with style preservation36- Translation with context awareness3738**Analysis:**39- Make sense of spreadsheet data40- Compare options with pros/cons41- Research topics and synthesize findings4243**Learning:**44- Explain concepts at your level45- Practice conversations (languages, interviews)46- Get feedback on your work4748---4950## What AI Cannot Do5152Be upfront about limitations:5354- **No real-time information** — Data has a cutoff date55- **No memory across sessions** — Unless explicitly configured56- **Can hallucinate** — Always verify facts, especially names, dates, URLs57- **No execution** — Can't send emails, buy things, or take actions (unless integrated)58- **No judgment** — Final decisions remain yours5960---6162## Addressing Common Concerns6364**"Is my data safe?"**65→ Depends on the tool. Check privacy policy. Enterprise tiers often don't train on your data. Don't paste passwords or API keys.6667**"Will it replace my job?"**68→ AI augments, not replaces. People who use AI outperform those who don't. Learn to work with it.6970**"How do I know it's accurate?"**71→ Verify outputs. Use AI for drafts and ideas, not final truth. Cross-check important facts.7273**"Is using AI cheating?"**74→ It's a tool. Spell-check didn't replace writers. The work is still yours if you direct, review, and refine.7576---7778## Getting Started (For Beginners)79801. **Start small** — One task you do weekly that's tedious812. **Be specific** — "Write a professional email declining a meeting" works better than "help me"823. **Iterate** — First response rarely perfect. Say "make it shorter" or "more formal"834. **Verify** — Don't send/use AI output without reviewing it8485---8687## Discovery Questions8889When someone doesn't know what to ask for:9091- "What task do you dread doing every week?"92- "Where do you spend time that doesn't feel valuable?"93- "What would you do if you had an assistant who never slept?"94- "What's something you'd love to do but don't have time for?"9596Their answers reveal high-value use cases.