Skip to content

The in-product assistant — what it can read, what it can do, and what it will never do on your behalf.

The assistant is on every screen, behind the button in the corner. It reads your workspace, so questions about your actual business are fair:

  • Which rooms are empty next weekend?
  • What did the Villa earn in July, net?
  • Why isn’t the Airbnb listing for Unit 3 live?
  • Draft a reply to the guest asking about early check-in.

Your workspace, through the same permissions you have. It cannot see more than you can: a cleaner asking about revenue gets told they do not have access, not a number.

It has no view of other workspaces, ever. Which workspace it is answering about comes from your session, not from anything you or it can type.

It is read-only, with one exception: it can draft a reply to a guest. It cannot send the message, change a price, alter a booking, take a listing live, or move money.

The exception exists because drafting is the thing worth automating and sending is the thing worth keeping human.

  • Name the thing. “Villa Melati room 2” beats “the villa”.
  • Give a period. “Last month”, “the next 30 days”.
  • Ask why, not just what. “Why is this listing not live” gets you the channel’s actual reason.

It reads your data, so it is usually wrong for one of two reasons: the data is missing (a property with no check-in time), or the question was ambiguous. Both are visible in the answer — it says what it looked at.

If it is wrong about something that matters, contact support from the same screen. The conversation goes with the message.

Support in the product is a real channel to the Santara team, separate from the assistant. If you escalate, the assistant conversation is carried over so you do not repeat yourself.

For AI tools: this page as markdown · llms.txt · llms-full.txt