When a user re-opens a Brain chat or navigates back to a previous prompt, the AI should return the stored response rather than regenerating it from scratch (as it happens now). Brain already remembers recent prompts in the session, so persisting the paired answers is a natural extension. Benefits: faster UX (instant recall vs. waiting for generation), lower compute cost, and (most importantly) consistent answers when revisiting the same question.