Voice note - to structure

A voice note is a stream of thought: I said it on the go before I forgot. Value appears when the raw flow is transformed into a structure - a list of tasks, a plan, theses. Mechanics: first, speech becomes text (transcript), then the model reads this text and arranges it in the form you specified. The key is to specify the form explicitly. If you simply say “do it beautifully”, you will get a random result. If you say “take out a list of tasks from the note, for each - what to do and the deadline, if the deadline is named,” you will get what you need. Pro trick: end the voice note itself with a short built-in command (meta-instruction) - say at the end “now break this down into tasks and ideas separately.” The model will hear both the note and the command in one pass. The second technique: ask the AI ​​to mark gaps - “where there is not enough data, ask a question” so that the structure honestly shows the holes, and does not invent. A typical mistake is to expect the AI ​​to guess the required format; he executes form well, but he doesn't read minds. Describe the target structure in your own words - and the note will turn into action.