Prompting 101
Turn vague requests into clear instructions that get you sharp, useful answers.
Be specific about what you want
A prompt is just the instruction you give the AI, and clearer instructions get better results. Vague requests like "write something about dogs" leave the model guessing. Instead, say what you actually need: "Write a friendly three-sentence intro about adopting senior dogs for an animal shelter newsletter." The more you pin down the topic, length, and purpose, the closer the first answer lands to what you wanted.
Give context and a role
The model does not know your situation unless you tell it. Add the background that matters: who the reader is, what tone you want, and any facts it should use. Asking it to take on a role helps too, such as "You are a patient math tutor explaining to a ten-year-old." That framing nudges the style and level of the response so it fits your audience instead of being generic.
Show the format you expect
If you want the answer in a particular shape, ask for it directly. Say "give me five bullet points," "reply in a short table," or "keep it under one hundred words." You can even paste a small example of the layout you have in mind. Spelling out the format saves you from reformatting later and makes the output easy to drop straight into your work.
Iterate instead of restarting
Your first prompt rarely has to be perfect. Treat the conversation as a back-and-forth: if the answer is close, just refine it with "make it shorter," "more formal," or "add an example." Building on the previous reply is usually faster than starting over. Small follow-up nudges are one of the easiest ways to steer the model toward exactly what you need.
Key takeaways
- Be specific about topic, length, and purpose so the model does not have to guess.
- Add context, a role, and the format you want to shape the response.
- Iterate with small follow-ups instead of starting over.
4 questions · pass at 60% to earn XP