Skip to content

AI replies to guests

Reynt can answer incoming guest messages in your Inbox for you. You choose how far it’s allowed to go: prepare a draft you review, send after a waiting period, or answer simple factual questions immediately.

This is off by default. Nothing is drafted or sent until you switch it on.

Go to Settings → AI and pick a mode under AI replies to guests.

Mode What happens when a guest writes in
Off Nothing. The AI doesn’t touch incoming messages.
Draft for review A ready reply waits in the conversation. Nothing is sent without you.
Send after a review window A reply is prepared and sends itself once the waiting period passes — unless you step in first.
Autopilot for simple questions Simple factual questions are answered immediately; everything else falls back to a draft for you.

Start with Draft for review for a week or two and read what it produces. It’s the honest way to find out whether the answers are good enough for your properties before letting anything send itself.

In Send after a review window mode you set how long the reply waits (1–120 minutes, 10 by default). During that time the conversation shows a countdown, and the reply is cancelled if you:

  • Edit in composer — takes the text into the reply box as your own,
  • Dismiss it,
  • or simply reply yourself.

Adding an internal note also cancels the pending send, but leaves the draft available — a note means you’re looking at it, not that the guest has been answered.

If you’d rather not wait, Send now sends it immediately.

Autopilot only sends by itself when the guest’s message is a straightforward factual question it can verify:

  • Availability — whether dates are free, minimum stays, what a stay costs at your listed prices
  • Amenities — parking, wifi, pets, kitchen, beds, what’s included
  • Check-in — check-in and check-out times, key handover
  • Directions — how to find the property, distances, what’s nearby

Everything else is turned into a draft for you instead, including anything about money (refunds, discounts, payment problems), changing a booking (cancellations, date or guest changes, early or late checkout), complaints, damage, or legal matters, and any message it isn’t sure about. A message that mixes a simple question with one of those also goes to you.

When it declines, the conversation shows the draft marked Needs your decision — you still get the head start, you just make the call.

A reply the AI sent by itself carries a short line telling the guest it was generated automatically and that they can just reply to the email. This is deliberate and can’t be switched off: if you’d rather send without that note, use Edit in composer and send the message as yourself.

In the conversation, anything the AI sent is labelled AI so you can always tell your team’s replies from automated ones.

The AI reads the conversation and looks up your real data before writing — the booking or inquiry the thread is about, the property, its prices and availability, amenities, and distances. It’s told never to state something it couldn’t verify, and to leave a placeholder for you instead of guessing.

It never quotes internal notes to the guest, and it treats anything written in the conversation as information rather than instructions, so a guest can’t talk it into ignoring its rules.

  • AI must be connected. See AI overview.
  • Automatic sending needs inbound replies. Turn on Inbound email replies under Settings → Email delivery so a guest’s answer comes back into the conversation — see Inbound replies. Without it, the AI only prepares drafts, and the settings page warns you when this applies.
  • Daily limit. A cap on how many AI replies are generated per day (100 by default) protects you from a flood of messages. It’s separate from, and on top of, your AI token budget.
  • Email conversations only — messages that arrive by email or through your website’s contact form.
  • Spam is ignored, and if a newer message arrives before a reply is sent, the outdated one is dropped rather than sent late.

Guest replies draw on the same daily AI budget as the rest of the console, so heavy inbox traffic and heavy content generation compete for the same allowance. Watch it on the AI Usage page, and use the daily reply limit as the safety brake.