A free, beginner-friendly guide to learning AI and using it well. Everyday work or a hard problem you're thinking through — this guide is built on one simple habit, designed to keep your judgment, agency, and skills firmly in your hands: Structure → Dialogue → Result. Whatever AI you choose — Claude, ChatGPT, or Gemini — the habit stays the same. You'll learn to structure the AI of your choice as a companion that thinks with you, not for you.
The whole guide is free to read right here — no sign-up, no paywall. The one-page SDR cheat sheet (PDF) is just below.
The guide starts below
This is a beginner's guide to using AI well — for the small, scattered work that fills a day, the harder thinking that takes longer, and the occasional decision that actually matters.
You don't need to be a programmer, a prompt engineer, or someone who memorizes perfect wording. What you need is a better working habit. That habit has a name — SDR: Structure, Dialogue, Result — and the whole guide builds around it.
But there's a more important idea underneath it. AI is more than a vending machine that hands you an answer when you press the right button. Throughout this guide it's called a tool, on purpose. Tools are things you learn, use, and put down. They don't replace judgment; they extend it.
A skilled user gets meaningfully different results than someone who types a question and accepts whatever comes back. That difference has a name too — shaping — and learning it is what this guide is for.
AI can help you think, organize, draft, compare, and explore. You still decide what fits, what's true enough to use, and what to do next.
By the end, you'll know how to shape AI to:
No matter how skillfully you shape it, three things stay yours: your judgment, your agency, and your responsibility for the outcome. AI is most useful when you stay in the conversation.
Tools like ChatGPT, Claude, and Gemini are large language models — LLMs. They work by predicting what words should come next, based on patterns in the huge amount of text they were trained on. That's the short version, and it's all you need here. The full technical story goes deeper than this guide needs, and it doesn't change how you use these tools well.
How to read this page: read the idea, try the prompt in your own words, add only the context the AI needs, ask one follow-up, then save the useful part. Start small, with one low-stakes task, and improve from there.
Vague questions get vague answers — and most beginner frustration starts right here.
Beginners often open with something like “Help me plan my life,” “Write something for me,” or “What should I do?” These feel clear in your head and land nearly empty on the page. You know what you mean. The AI only has the words.
So it fills the gaps. It doesn't know your goal, your audience, your timeline, your real constraints, the tone you want, or what “done” looks like for you. When those are missing, it guesses — and a guess dressed in confident prose is exactly what a generic answer is.
“Help me plan my week.”
“Help me plan my week so I can finish my taxes, exercise twice, and prepare for a Friday meeting. I work 9–5, have low energy after work, and don't want an overpacked schedule. Ask me three questions before you build the plan.”
Look at what the second version adds: a goal, a little context about your day, one real constraint, and a request for questions first. None of it is clever wording. It's just enough direction to work with.
And if you can't tell what direction to give, there's a question that does the work for you:
“What would help you give me a better answer?”
Ask that, and the AI tells you what's missing. Now you have the outline of the prompt you should have written in the first place.
The most reliable beginner habit is to tell the AI to ask you questions before it answers. That one instruction turns it from a guess-machine into something closer to a thinking partner. A pattern you can reuse:
“I want help with [task] so I can [goal]. Important context: [what matters]. Constraints: [time, budget, format, audience]. Before answering, ask me [number] questions if anything important is missing.”
For example:
“I want a simple morning routine so I can start the day with less stress. Before giving advice, ask me five questions about my schedule, my energy, and what's gotten in the way before.”
When an answer sounds useful but feels slightly off, it's usually running on an assumption you never gave it. Pull those into the open:
A few tells that the AI is filling gaps instead of answering you:
A vague prompt makes the AI guess. Fill each gap and watch how much it has to guess fall to nothing.
A vague prompt — the AI fills the gaps
Fill a gap — tell it, instead of letting it guess
It’s not filling in the blanks anymore — you are.
Structure → Dialogue → Result. The one habit the whole guide is built on — if you read one part, read this.
AI works best when you shape the conversation — steer it deliberately instead of taking the first thing it hands back. Three steps, every time:
Before you type, take ten seconds to think about five things:
That last one turns the AI from a guesser into a thinking partner.
The first answer is a draft, not a verdict. Read it, then ask one clear follow-up:
You don't need a clever follow-up. You need a clear one.
Don't save the whole conversation. Pull out the part you'll actually use:
If the useful part is buried, ask: “Pull out the next three actions.”
Walk through the method once. Each step adds a line to a single prompt — by the end you’ve built one complete, copyable ask.
Build the core of your prompt — pick a piece for each part.
Goal
Context
Constraint
Now set up how it works with you — this gets added to the prompt.
Dialogue
Last, add the form you want back — this finishes the prompt.
Result
That’s one complete prompt — Structure, Dialogue, and Result, all in a single ask.
Want the whole method on a single printable page? We'll email you the free SDR cheat sheet (PDF).
Get the cheat sheet (PDF)AI can help you think — it can't replace your thinking. A confident tone is never the same as a correct answer.
AI tools can organize messy ideas, draft language, compare options, explain concepts, and suggest next steps. That makes them useful — and it makes them easy to over-trust.
A clean answer is not the same as a correct answer. A confident tone is not the same as evidence. A detailed plan is not the same as a plan that fits your real life.
AI doesn't know your goals, your values, your responsibilities, or what's actually at stake for you. So the judgment stays with you: you decide what matters, what's true enough to use, what needs checking, and what to do next.
Type a question, read what comes back, move on. The answer feels generic — because the question was generic. You've quietly handed over the thinking.
Give it direction. Push back when something doesn't fit. Refine until it helps, then pull out the part you'll use. The thinking stays yours.
AI writes in a smooth, confident style, and that polish can feel like authority. It's one of the biggest beginner traps — and it matters most when the answer touches something with real consequences: your health, money, legal standing, safety, job, family, a major purchase, or private information. For those, use AI to organize your thinking, prepare questions, and compare trade-offs — then take the decision to a real expert. There's a full section on high-consequence topics later.
Before you act on something AI gave you, run five quick questions:
Drop either of these in when an answer looks useful but you want to pressure-test it:
“Before I use this, list the assumptions you're making. Then tell me what I should verify from another source.”
“Treat this as a first draft, not a final answer. What could be wrong, missing, or oversimplified?”
An AI gives you a confident answer. It’s exactly the kind worth checking — so what would you write back to make it prove it?
Question asked
Does taking short breaks actually improve productivity?
The AI’s answer
A 2011 study in the journal Cognition found that working in exactly 52-minute blocks boosts performance by 43%.
Your move — pick the reply that gets it to back this up
You came in wanting a better way to work with AI, and now you have one. You saw why vague questions get vague answers. You picked up SDR — Structure, Dialogue, Result — as a habit you can run on anything, from a quick email to a real decision. And you learned to keep your judgment in the loop, so a confident answer never gets mistaken for a correct one.
If one idea sticks, let it be shaping. The difference between a generic answer and a useful one usually comes down to whether you gave the AI direction and kept refining it, or took the first thing it handed back. Your judgment, your agency, and your responsibility for the result stay with you — the AI helps you think, and the thinking still belongs to you.
This guide stays tool-neutral on purpose — SDR works the same whether you use Claude, ChatGPT, or Gemini. What comes next goes deeper. Companion guides take one tool at a time and walk through its setup and the features worth knowing, starting with a guide to Claude and more to follow. Deeper guides on specific ways to use AI are in the works too, for the times a habit isn’t enough and you want a full worked example.
More guides are on the way. Grab the free SDR cheat sheet (PDF), and we’ll send a note when the next one ships.
Get the cheat sheet (PDF)This guide is free, and I’d like to keep it that way. If it helped you get more out of AI, a tip keeps the next one coming — no pressure, and thank you for reading.
Enjoying the guide? Consider supporting my work :)
Tip on Ko-fiThe whole method on a single printable page — Structure → Dialogue → Result, the follow-up moves, and how to match your care to the stakes. Read the guide once, then keep this by your desk.
We just sent you a confirmation email. Click the button inside and the cheat sheet will be on its way. If it doesn't show up in a few minutes, check your spam folder.