Home Using neurolist how the AI breakdown works (and how to tune it)

how the AI breakdown works (and how to tune it)

Last updated on Jun 23, 2026

When you add a task, the AI receives:

  • The task title you typed.
  • Your estimate of how long it'll take and how hard it feels (you set these on the "confirm information" screen).
  • The bullets stored in your memory (see → "what neurolist knows about you").
  • Some context about how you've broken down similar tasks before.

It returns a list of small, concrete subtasks. Each one is sized to be small enough that you don't bounce off it before starting.

Editing the breakdown

The AI's output is a starting point, not a verdict. On the breakdown screen you can:

  • Tap a step to edit it.
  • Tap the X on a step to remove it.
  • Tap the + icon at the bottom of the list to add a new step.
  • Reorder steps by long-pressing and dragging.
  • Undo / redo with the controls at the bottom of the screen.

Your edits also help future breakdowns improve — neurolist learns from how you adjust the AI's suggestions.

Making future breakdowns better

  • Be more specific in the task title. "Clean kitchen" vs. "clean the stovetop and wipe the counter" — the second gets a sharper breakdown.
  • Tune the time and difficulty estimate. The AI uses these to decide how many steps to generate.

The AI's output is guesswork

The AI uses OpenAI's models to generate plans. AI output is best-effort and can be wrong. Use your own judgment — especially for anything important (legal, financial, medical). neurolist is not responsible for what the AI suggests.

If a breakdown feels off → see "the AI gave a weird or wrong breakdown — what to do".

Still stuck? [email protected].