Sabre Corporation has launched a pilot program of an AI-powered chatbot with a US travel agency.
Travel Services International USA (TSI) and its customers will trial the chatbot, which will be available via the TSI app or on Facebook Messenger.
The bot, named Ella, has been designed to answer frequently asked questions and save agents time so they can focus on more complex inquiries. It will be reviewed to see how customers interract with the bot.
“We are excited to start pilot testing for the bot,” said Chad Callaghan, head of the Sabre Studios incubation team. “We believe the self-service convenience the bot offers will improve travellers’ ability to resolve routine support requests. At the same time, we are happy to help our travel agency customers ensure agents can focus on supporting more complex traveller requests. We will be interested to track how travellers learn about and interact with the bot.”
TSI will test a white-label version of the chatbot with its corporate travellers, who can use the chatbot via Facebook Messenger to answer questions related to existing flight reservations.
Travellers can ask trip-related questions, make voluntary changes to existing flight reservations, request a special meal, select and pay for a seat, and request an emailed copy of an itinerary.
The chatbot, leverages Microsoft’s Bot Framework and Microsoft Cognitive Services, will divert to a live travel agent if it is unable to fulfill a request.
“We’ve named the TSI bot, Ella,” noted Claire LeBuhn, vice president, support services at TSI USA. “Ella represents TSI’s heritage of innovation and commitment to exploring new technologies and tools that can help our corporate customers and their busy employees. Ella supports reservations originated by TSI. We are eager to understand how corporate travellers engage with Ella and the kinds of emerging conversational interfaces that she represents.”
Throughout the test, Sabre and TSI will evaluate travellers’ preferences for how often and when they engage the chatbot and when they are likely to divert to a live agent.
The companies expect to gain a better understanding of the best way to engage travellers with chatbots and solve some of the most common friction points associated with chatbots, such as how travellers will find the bot.