Everyone’s Talking About AI, and I Don’t Know Where to Start - How to Use AI for Everyday Tasks, Check Its Answers, and Keep Thinking for Yourself ebook

Chapter 1 - You Don’t Need to Know How AI Works to Know What It Can Help You With

One of the first mistakes people make is assuming that before they start using AI, they first have to “understand” it. As a result, they watch videos, read articles, learn new terms, and after a week discover that the subject is even bigger than it seemed at the beginning. It is a little like preparing to cook scrambled eggs by studying the engineering behind modern induction cooktops. The knowledge may be interesting, but breakfast still is not ready.

For everyday use, a very simple mental model is enough at first. The AI you communicate with through text is a tool that can analyze material you provide and generate responses based on patterns, information, and the context it receives. It is not a human being, it does not have common sense that works exactly like yours, and it should not automatically be promoted to expert status simply because it writes smoothly. At the same time, it can perform many tasks involving text, ideas, information organization, and option analysis extremely quickly.

The easiest way to begin is not by asking, “What can AI do?”, but by looking at a normal day. Think about the things that make you sigh, the ones you keep postponing, or the ones you mechanically repeat for the hundredth time. Maybe you regularly write very similar emails. Maybe you receive long documents from which you only need five important facts. Maybe every trip begins with creating a packing list from scratch, even though for years you have been taking roughly the same pair of trousers and exactly the same charging cable.

Those are good places to experiment because the risk is low and the result is easy to judge. If AI creates a poor packing list for a weekend away, you can add your toothbrush and remind yourself why humans are still needed in the process. If it reduces a four-hundred-word email to one hundred and twenty words, you can immediately see whether the meaning survived. If it organizes your meeting notes, you can compare the summary with what you actually wrote down.

Start With Tasks, Not Technology

Imagine Marta, who hears about AI almost every day but still does not really know what she is supposed to do with it. Marta does not code, she does not design rockets, and she has no plans to launch a startup with friends in a garage she does not own. What she does have is a Monday routine in which she spends forty minutes organizing notes after her team meeting. That is her starting point.

Instead of asking AI, “How can you help me at work?”, Marta can paste in her own notes and write: “Organize these notes into a short summary. Separate decisions, tasks, owners, and deadlines. Do not add information that is not in the text. If something is missing, mark it as not established.” Suddenly, AI is no longer an abstract technological revolution. It is a way to turn forty minutes of tedious work into a few minutes spent checking the result.

That leads to one of the most important principles in this book: the best first uses of AI are usually ordinary ones. Do not try to create an automated system for managing your entire life on day one. Find a task that is repetitive, text-based, time-consuming, or irritating, while still being simple enough that you can judge the quality of the answer yourself. If your choice is between “help me shorten this email” and “decide whether I should change careers,” the email is definitely the better first experiment.

You can divide tasks into three simple groups. The first contains tasks where AI can help quite freely because any mistake will be easy to spot and correct. The second contains tasks where the answer matters and needs checking. The third contains situations where the consequences are serious enough that AI should at most help you prepare questions or organize information, but should not be the only basis for a decision.

The first group might include title ideas, shortening your own text, preparing a shopping list, generating questions for a meeting, or organizing a cleaning plan. The second might include product comparisons, travel information, interpretation of offer terms, or document summaries. The third includes health questions, legal matters, serious financial decisions, and other situations where a mistake could cost far more than a forgotten bottle of shampoo.

The point is not to create a formal policy for every conversation with AI. The point is to develop one simple reflex: before using the answer, ask yourself what happens if it is wrong. If the answer is “not much, I’ll fix it,” you can be fairly relaxed. If the answer is “I could lose money, harm my health, or make an important decision based on false information,” the level of checking should rise immediately.

Seven Ordinary Things AI Can Actually Help With

Many of the most useful applications of AI fall into a few broad types of work. AI is very good at transforming text, which includes shortening, simplifying, changing tone, organizing, or translating material you already have. It can help create a first draft when you know what you want to say but do not know how to begin. It can also generate ideas, compare options according to criteria you provide, explain a difficult subject in simpler language, build checklists, and identify questions you have not yet thought to ask.

Suppose you receive a message from your child’s school that runs for nine long paragraphs, contains three dates, two exceptions, and one sentence written as though the author was paid by the word. You can ask AI: “Summarize this message in five points. List deadlines, things to prepare, and actions required from the parent separately. Do not omit exceptions.” You should still glance back at the original, especially if there is a deadline or payment involved, but instead of fighting through the text a second time, you now have a map.

Another example is a complaint. You know what happened, you have the purchase date and order number, but by the third sentence you are beginning to sound like someone preparing to demand a parliamentary inquiry. You can describe the facts and ask AI to prepare a calm, specific version of the message. Then you read it, correct the details, and decide whether you really want to send the phrase “I kindly request that this matter be reconsidered,” or whether your emotions would prefer a slightly stronger verb.

AI can also help when the problem is not too little information but too much of it. You have three internet plans, five coffee machines, or eight ideas for a weekend away, and after an hour they all begin to look the same. Instead of asking, “Which one is best?”, give your criteria and ask for a comparison. “Reliability, easy cleaning, and low noise matter most to me. Price matters less. Compare these options using those criteria and tell me what information is missing for a fair choice.”

That final sentence is especially valuable. AI is very willing to give an answer even when it does not have enough information. If you learn to ask, “What don’t you know?” or “What information is missing?”, you reduce the risk of receiving an elegant solution built on a foundation of cotton wool. It is one of those small changes that makes the conversation much more useful.

Don’t Ask Only, “What Should I Do?”

When people discover a tool that answers almost instantly, it is easy to move from “help me think” to “think for me.” The difference looks small on the page but can be enormous in practice. If you always ask AI what to choose, what to reply, what to buy, and how to react, you may slowly lose touch with your own criteria. The answers will become faster, but your decisions will not necessarily become better.

A better approach is to use AI to widen your field of view. Instead of asking, “Which hotel should I choose?”, you can write: “Compare these three hotels based on beach distance, room quality, evening quiet, and convenience for someone who wants to spend part of the trip sightseeing. Show the trade-offs. Do not choose for me.” You get material for a decision, but the steering wheel remains on your side.

The same applies at work. Instead of asking, “Should I accept this project?”, try: “Help me analyze this decision. List potential benefits, costs, risks, information I am missing, and questions I should ask before answering.” This form pushes the tool to support your thinking process rather than pretend it understands your priorities better than you do.

You can think of AI as someone sitting across the table with a whiteboard. It is fast, has many ideas, and may notice things you missed. But it does not automatically know your life, relationships, budget, history, risk tolerance, or the fact that when a hotel promises an “energetic atmosphere in the evenings,” what you hear is “karaoke until two in the morning.” Context matters.

Your First Prompt Does Not Have to Be Perfect

Many people get stuck on the word “prompt,” as though it were a secret access code to a control center. A prompt is simply the instruction or question you give the AI. The better you describe the task, the better your chances of getting a useful result, but that does not mean every question needs to become a technical document. Conversation has one major advantage: you can improve it as you go.

A good starting point usually contains four things: what you have, what you need, what limits matter, and what form the result should take. For example: “Tomorrow I have a conversation with my building manager about recurring heating problems. I want to prepare calmly and specifically. Based on the notes below, create a list of six questions and a short summary of the facts. Do not add laws or information I did not provide.”

This prompt will not win a literary prize, and that is exactly right. It is clear. AI knows what it is receiving, what it should do, what the context is, and what it should not invent. If the result is too long, you say, “Reduce the summary to five sentences.” If the questions are too soft, you say, “Make them more specific while keeping them polite.”

Conversation with AI is iterative. The first answer is often a draft, not a bronze monument. You can improve it, narrow it, expand it, and challenge it. That is one of the biggest advantages of this tool over a blank page: you do not have to know everything at the beginning.

The Five-Part Method

If you want a simple framework, remember five elements: situation, task, criteria, format, and control. First, briefly describe the situation. Then say what you want AI to do. Next, provide the most important criteria or limits, specify the form of the answer, and finally add a control instruction, such as asking it to mark uncertain information or not invent missing data.

Suppose you want to plan a weekend in a city. Instead of “Plan my weekend,” you might write: “My partner and I are spending Saturday and Sunday in the city. We like walking, good coffee, and quiet restaurants, and we are not interested in nightlife. Create a relaxed two-day plan with no more than three main activities per day. Leave time to rest and mark which details, such as opening hours or prices, I should check before going.” It is still a simple prompt, but the result has a much better chance of being useful.

The control element is especially important when facts are involved. You can write, “Separate information taken from the provided text from your own suggestions,” “Do not guess missing values,” “If you are uncertain, say so,” “Mark the parts that require checking,” or “Give me the questions I should verify in an official source.” None of this makes AI suddenly infallible, but it helps establish the right mode of collaboration.

Don’t Hand Over Data You Don’t Need to Hand Over

While experimenting, it is worth building one more habit: before pasting material into an AI system, ask whether you really need to provide every piece of information contained in it. If you want help improving a letter, you often do not need to include your full address, document number, your child’s details, customer number, or confidential company information. You can replace them with placeholders such as “[CONTRACT NUMBER]” or “[COMPANY NAME].” AI needs context to do the task, but it does not always need your entire filing cabinet.

The same applies to other people’s information. The fact that you technically can paste an entire exchange with an employee, customer, or doctor does not automatically mean you should. Privacy rules depend on the tool, settings, and environment you are using, so it is worth knowing what you are working with. A useful starting rule is simple: if the task can still be done after removing identifying information, remove it.

There is no need to become paranoid about this. It is ordinary information hygiene, similar to the idea that your payment card number is not an excellent addition to a group chat. AI is a tool, and tools work best when convenience is combined with basic boundaries.

Your First Experiment Should Be Small

After finishing this chapter, do not create a folder called “AI TRANSFORMATION 2026” and start redesigning your entire life. Choose one thing you would normally do within the next two days. Ideally, it should take between ten and thirty minutes, contain no especially sensitive information, and produce a result you can judge yourself.