What happens when the AI does not know
Your guest asks something TravelPal's AI does not know. It does not guess: it asks your team, and this is called escalation.
An illustrative chat with an invented guest and property.
How escalation works, in four steps
The question is held
TravelPal searches what it knows about your property, then searches again with different words. If it still has no answer, it holds the guest's exact question for your team. Only after that, it tells the guest your team was asked. So what the guest reads is already true.
Your team is told right away
As soon as a question is held, your team is told in two ways. A message and an email are sent at the same time. The alert shows the guest's name, the property and the question.
You answer once
The question waits as a card in your TravelPal inbox. One person answers it once. A second click, or another person answering the same card, changes nothing and sends nothing new.
Your answer appears in the chat
TravelPal writes a message from your answer. The message appears in the guest's chat. If you save your answer, it becomes property knowledge. Then the same question never has to come to your team again.
Held questions in your inbox, most important first
TravelPal sorts the list for you and shows every open item, oldest first.
Your team should not have to search for a booking. Each alert has enough detail to act:
A question is held. A request becomes a task before the AI confirms it:
Styled after TravelPal's answer panel, with an invented guest and property.
What happens when you answer a held question
You type your answer once. TravelPal does the rest.
TravelPal writes the message to the guest
TravelPal writes a plain message from your answer and speaks as “I”. The guest keeps talking with TravelPal in the same chat. Your team does not have to write there at 2 AM.
The question leaves your inbox
The question is marked answered and leaves your inbox. If a waiting guest asks the same open question again, TravelPal does not hold it twice.
Saved as property knowledge, if you choose
If you save it, your answer becomes knowledge written by your team. It is saved for that property, under a topic, with today's date. Your answer stays yours. Automatic learning is never allowed to replace it. Your answer is in your knowledge base with everything else the AI knows. To see how saved answers grow over time, read how the AI learns.
Guest requests are closed automatically
When your team marks a request done, or says no to it, the guest is told automatically. TravelPal writes that message too, as “I”. Nobody has to remember to write to the guest again.
Two promises are true at every step. The AI never promises how fast your team will answer. It never tells a guest that your team was asked unless it has held the question first.
Where escalation ends and stopping begins
Not every message the AI cannot answer becomes a held question. One kind of message stops the AI in that chat instead.
A question waits for you
The question is held, your team is told right away, and your answer appears in that chat.
A complaint stops the AI
A guest who complains needs a person, not an AI. A complaint or a request for money back is not a held question. It stops the AI completely in that chat, and a person on your team is alerted.
Guardrails are rules that limit the AI. Read the full list: when the AI stops. To see held questions in a whole team's day, read how TravelPal works for property management companies.
Read next:
TravelPal is an AI guest messaging and upsell tool for vacation rental property managers, connecting to 33 property management systems including Guesty and Hostaway, priced at $25 per unit per month. Escalation is part of that price, not sold separately.
Updated