Your first 15 minutes

Do not take an answer on faith. Test it on a task you know. These short exercises from the book's introduction help you see, in fifteen minutes, where AI is useful and where it can be confidently wrong. Checked: 2026-06.

Open any assistant from the list and try these five short prompts. Write as you would to a friend.

Five tests

Explain something. "Explain [a difficult topic from my work] in plain language, as if I were 15." Is it clear? Did the simplification distort the point?
Make a draft. "Write a short, polite draft email to a client about moving a meeting." Is it a useful start? What would you change for your situation?
Review something. Paste your own text, without personal data: "Find the weak spots in this text and suggest three improvements." AI is often more useful as a critic than as an author.
Test for an error. Ask about something you know for certain: a fact from your field, a date, or a number. Check the answer. You will see that "confident" does not mean "correct."
Ask for options, not an answer. "Give me three approaches to [my task], with the pros and cons of each. I will choose." This is the main move: let AI widen the choice, then make the decision yourself.

What to notice

AI is good at first drafts and options. It is less reliable for accuracy and the final decision.
The smoother the answer sounds, the more carefully you should check it. Smooth writing can make you lower your guard.
The point is not to let it do your work. The point is to reach a result faster while you remain the person who owns the decision.

Your first real result

The tests show what AI can do. Now use it on a task you have kept putting off: a procedure you have not had time to write, a presentation you keep delaying, or a spreadsheet that has been waiting for a review. Pick something that has sat on your list because you never get around to it.

You can get a usable first draft in one evening and decide whether this way of working is worth keeping. You still check it and finish it yourself.

Next: safety, including what not to put in a chat.

Version: 2026-06-20.