Decision support, not a chatbot

NORA is not a chatbot. She is the layer of BEST that notices things — and she only speaks when noticing something changes a decision you are about to make.

The shape of everything she says

1. An observation — something true about your stay, your trip or the day.
2. Why it matters — the consequence for you specifically.
3. The basis — what the observation was derived from, so you can judge it.
4. A useful action — something you can actually do about it.

A real example: "There is one place to eat within walking distance of this stay, and it is the guesthouse itself. Basis: 2 mapped food options within 800 m, population 19. Consider arranging dinner when you book the room."

What NORA is not

• She is not a language model, today. Every observation is a rule applied to real data.
• She does not chat for the sake of it, and she does not appear on pages where she adds nothing.
• She does not make things up. If the underlying data is missing, she says so rather than guessing.
• She has no memory of you between sessions, and no profile beyond what you tell BEST directly.

The reason she is rule-based is deliberate. Fluent prose over thin data hides the thinness. Plain sentences over honest data do not.