Write your first request to the AI assistant. Return ONLY JSON {"about_me":"...","goal":"...","format":"..."}. about_me - 2-3 real facts about you and your situation; goal - what you want to get right now; format - in what form is the answer convenient (list, table, steps).
The AI assistant works like a mirror: what it gives as input, it will reflect. By default, he doesn’t know anything about you and doesn’t remember between conversations. The mechanics are simple: the model predicts a useful answer based on your text. The more specific the input, the more precise the output. Three levers make a request work: context (who you are, what the situation is), goal (what should happen) and format (in what form to issue it). Insider: don’t ask for “help”, but “give me 3 options and mark which one you would choose” - this way the assistant takes part of the decision upon himself, rather than dumping the choice back. Second trick: end the request with the phrase “ask me 2 clarifying questions before answering” - the model will pull out what is missing instead of guessing. A typical beginner mistake: write one word 'planning' and wait for a miracle. The assistant will dispense the total water. Give him the material - your real situation - and you will get an answer tailored to you.