AI chat assistants like ChatGPT, Claude, Gemini and Microsoft Copilot can feel a little intimidating at first. What do you type? Will you “break” it? Here’s the reassuring truth: you can’t break it, and talking to an AI assistant is a lot like giving instructions to a helpful new colleague. The clearer you are, the better the help you get.
Step 1: Pick one assistant to start
All of the big assistants have free versions, and any of them is a fine place to begin:
- ChatGPT (OpenAI): a popular all-rounder
- Claude (Anthropic): strong with long documents and careful writing
- Gemini (Google): handy if you use Gmail and Google Docs
- Microsoft Copilot: built into Windows and Microsoft 365
Choose one, create an account, and stick with it for a couple of weeks while you learn. You can always try others later. (See our tool pages for current free and paid plans.)
Step 2: Understand what a “prompt” is
A prompt is simply what you type (or say) to the AI. It can be a question, a request or a set of instructions. There’s no special code or magic words. Plain, everyday language works best.
Step 3: Use the “four-part” recipe
The companies that make these tools publish advice on writing good prompts, and it comes down to a few common ideas: be clear and specific, give context and say what you want the result to look like (OpenAI; Anthropic). An easy way to remember it:
- Task: What do you want? (“Write an email…”)
- Context: What’s the background? (“…to my landlord about a leaking kitchen tap that’s been broken for two weeks…”)
- Tone or role: How should it sound? (“…polite but firm…”)
- Format: What should the result look like? (“…under 150 words, with a clear subject line.”)
Before and after
Vague prompt:
Write an email to my landlord.
Clear prompt:
Write a polite but firm email to my landlord asking them to fix a leaking kitchen tap. It’s been leaking for two weeks and I reported it once already on May 3. Keep it under 150 words and include a subject line.
The second version gives the AI everything it needs, so you’ll spend less time fixing the result.
Step 4: Have a conversation
You don’t have to get it perfect the first time. The real power is in the follow-up. Try:
- “Make it shorter.”
- “Make it sound friendlier.”
- “Give me three other options.”
- “Explain that like I’m new to the topic.”
- “Put that in a table.”
- “What questions should I ask a professional about this?”
Treat it like a back-and-forth with an assistant who’s happy to revise as many times as you like.
Helpful prompt starters
Here are some everyday prompts to try:
- Explaining: “Explain what a 401(k) is in simple terms, with an example.”
- Summarizing: “Summarize this article in five bullet points: [paste text].”
- Planning: “Create a 7-day meal plan for two people on a budget, with a shopping list.”
- Writing: “Help me write a thank-you note to my child’s teacher.”
- Learning: “Quiz me with 5 questions on basic Spanish greetings.”
- Deciding: “Help me compare two job offers. Ask me questions first.”
That last one is a great trick: ask the AI to ask you questions first. It helps it understand what you really need.
Step 5: Give it examples
If you want something in a particular style, show it. For example: “Here’s a social post I liked: [example]. Write three more like it about our weekend bakery special.” Both OpenAI and Anthropic recommend examples as one of the most effective ways to guide the result (OpenAI; Anthropic).
Common beginner mistakes (and easy fixes)
- Being too vague. Add who, what, why and how long.
- Accepting the first answer. Ask for changes, more options or simpler wording.
- Trusting every fact. AI can make things up. Double-check names, numbers, dates and quotes, especially for health, money or legal topics.
- Sharing private information. Don’t paste passwords, ID numbers or other people’s personal details. Use placeholders like “[my address]”.
- Giving up too soon. If an answer misses the mark, rephrase and try again. That’s normal, not failure.
Good to know
- Free versions have limits. You may hit a daily cap or get a less powerful model. That’s fine for learning.
- Some assistants can search the web, which helps with recent news, but still check the links.
- You can usually delete chats or turn off chat history in the settings.
- Voice mode on phone apps lets you just talk, which many beginners find easier.
Practice for 10 minutes
Try this today:
- Ask your assistant to explain something you’ve always wondered about.
- Ask a follow-up: “Can you give me a real-life example?”
- Ask it to turn the answer into a short checklist.
- Check one fact from the answer on a trusted website.
That’s it. You’ve just practiced the core skills: asking clearly, refining and verifying.
Want to go further?
- Google AI Essentials: a beginner course on using AI tools at work
- Elements of AI: a free introduction to how AI works
- DeepLearning.AI prompt engineering short course: a free, slightly more technical next step
The takeaway
Good prompting isn’t a technical skill. It’s clear communication. Say what you want, give some background, describe the result you’d like, and keep the conversation going. With a little practice, an AI assistant can become one of the most useful tools you have.
Sources
- OpenAI, “Prompt engineering” guide. https://platform.openai.com/docs/guides/prompt-engineering
- Anthropic, “Prompt engineering overview.” https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview