Your First 5 Minutes
No setup, no jargon. In five minutes you'll get one genuinely useful result and understand why it worked.
- Open Claude and ask your first real question (no setup needed)
- Write a prompt the way a sharp colleague would want it: goal, length, audience, tone
- Refine an answer with plain-language follow-ups (iterative refinement)
- Adopt the one safety habit that matters: verify before you trust
:::tip Add a screenshot here This page is the beginner floor — it deserves annotated screenshots or a short GIF. If you can contribute visuals, please open a PR. :::
The 5-minute path
- Go to claude.ai and sign in (the free tier is enough to start). You can also use Claude on the mobile app or by voice — see Which Claude Should I Use? for the full map.
- Don't 'search.' Give context and say what you want: name the goal, length, audience, tone, and what 'done' looks like. That's 80% of prompting, and it works on every AI.
- You don't have to get it perfect first try. Reply in plain language — 'make it shorter and more casual,' 'add a sentence about the free Q&A.' This back-and-forth is iterative refinement.
- Prove to yourself it's general-purpose: summarize an article, explain a concept simply, or plan a quick dinner from the ingredients you have.
- AI can sound confident and still be wrong. For anything that matters, check it. Verify facts, numbers, dates, quotes, and citations.
Minute 2, up close — the single biggest beginner upgrade
The biggest first-day win: give context and say what you want. Compare:
❌ "marketing email"
✅ "Write a friendly 120-word email inviting existing customers to a webinar on June 30 about our new analytics dashboard. Warm but not salesy. End with a clear RSVP call to action."
The second prompt names the goal, length, audience, tone, and what 'done' looks like. That's 80% of prompting, and it works on every AI. (More in Prompting Basics.)
Your first prompt to try
Write a friendly 120-word email inviting existing customers to a webinar on June 30 about our new analytics dashboard. Warm but not salesy. End with a clear RSVP call to action.
Minute 3, up close — refine in plain language
You don't have to get it perfect first try. Just reply like you would to a colleague:
- "Make it shorter and more casual."
- "Add a sentence about the free Q&A at the end."
- "Give me 3 subject-line options."
This back-and-forth — iterative refinement — beats agonizing over one perfect prompt.
Minute 4, up close — try a second, different task
Pick one to prove it's general-purpose:
- "Summarise this article in 5 bullet points: [paste text]."
- "Explain compound interest to me like I'm 12, with one example."
- "I have chicken, rice, and spinach. Give me a 20-minute dinner."
Minute 5, up close — the one safety habit
:::warning Verify before you trust AI can sound confident and still be wrong — especially about facts, numbers, dates, quotes, and citations. For anything that matters, check it. This isn't a flaw you can prompt away; it's how the tool works. See Hallucinations & How to Reduce Them. :::
- Context beats keywords: state the goal, length, audience, tone, and what 'done' looks like
- Iterative refinement (plain-language follow-ups) beats chasing one perfect prompt
- Claude is general-purpose — try a second, unrelated task to feel it
- Always verify anything that matters; confidence is not correctness
Check yourself
Check yourself
0/3You're off the ground 🎉
Where to next:
- Make every answer better → Prompting Basics
- Pick a structured plan → Learning Paths
- Use AI for real work → Everyday Productivity
- Something felt off? → Troubleshooting: "Why did Claude do that?"