Auto-Agent is powered by IBM Watson - a computer system capable of answering questions posed in natural language, developed in IBM's DeepQA project research team. We’ve been training it specifically to be able to answer the kinds of questions that hoteliers are asked by guests when chatting online.
The more conversations that Auto-Agent has, the more it learns about how to answer future questions, and the better it gets at helping guests to book directly with your hotel.
When the Assistant recognises a question, it looks up the answer within a hotel’s FAQ responses, which are filled out by the hotel in advance. If there is a response available, the Assistant will respond, word for word, to the guest. It also has automated ‘flows’, like an availability enquiry, where it will gain additional information and try to resolves the enquiry without any hotel agent interaction. This is all in aid of reducing the number of questions the hotel staff have to reply to.
If the Assistant doesn’t know the answer to a question, it will ask the guest whether they would like to have a response from a human agent and, if so, the guest should leave their contact details so the hotel staff can respond at a time best suited - live chat is on the rise, but hotel staff are often too busy to answer immediately, so we try to gracefully transfer the conversation to email, which most hotel are best suited for.