Expedia today unveiled a rather smart, rather simple little bot for Facebook Messenger. The bot is designed to do one thing and do it well: It helps travelers book hotels. What’s more, the bot is extremely intuitive. As in easy-to-use. For the tech savvy, this may seem like a big “duh,” but simplicity matters given that bots are still in their infancy and Expedia’s customer base spans all walks of life: the company owns more than 200 travel booking sites from Orbitz to HomeAway to Trivago to Hotels.com.
While the bot took only about 5 weeks to develop, Expedia got a head start in making it so smart. “Our work with natural language processing and machine learning – the tools that helped get us here – have been going on for years,” said Tarran Street, senior public relations manager. “For us it’s all about experimentation to better understand travelers’ needs. The rise of mobile means people are starting to move away from the mouse and traditional keyboard, we want to be sure our products are available wherever travelers are searching.”
In a blog post on the company web site, David Fleischman, vice president of global product, described the bot’s use of natural language processing, “The bot operates on a structured conversation flow: it analyzes information provided and prompts the user to input other relevant data points to complete a search.”
Expedia’s bot for Facebook Messenger is a marriage of giants. The travel company booked more than $60 billion of travel for its customers across 282,000 hotels, 475 airlines and dozens of car rental companies. Facebook Messenger, meanwhile, has 900 million registered users; Expedia has 5.9 million followers on Facebook. It’s likely that booking hotels is just the beginning of a broader more comprehensive messaging strategy.
For now, the bot does not directly integrate with other Expedia services like Affirm for payments, or HomeAway for vacation home rentals, said Street. “It can, though some of that magic doesn’t happen directly within the bot. That part of the experience happens on an Expedia page vs. within the Messenger app directly.”
For instance, Street said, those travelers wanting to see vacation rental options from HomeAway can click on the “View All Hotels” option within the bot. They are then taken to an Expedia search result page with relevant properties, and can select a filter option.
Asked about incorporating services from partners like airport car service provider Wingz (Expedia invested$11 million in the company), Tarran said, “[They’re] not integrated into the Expedia bot directly yet; we’re still at early stages so we’re starting with hotels.”
Like other companies building bots, Expedia sees this initial effort as dipping a toe in the water. “For us it’s all about experimentation to better understand travelers’ needs. The rise of mobile means people are starting to move away from the mouse and traditional keyboard, we want to be sure our products are available wherever travelers are searching,” Street said. “Being involved in [new] technologies and learning how our customers interact with them gives us an opportunity to get their feedback and finesse our approach across other parts of our site.
But a simple toe in the water could soon lead to a cannonball-sized splash. Expedia booked more than $60 billion of travel for its customers across 282,000 hotels, 475 airlines and dozens of car rental companies during the last 12 months. Facebook Messenger, meanwhile, has 900 million registered users; Expedia has 5.9 million followers on Facebook. It’s not too hard to imagine Expedia going from booking hotels to more vertically-integrated bots on other platforms beyond Facebook Messenger. “Absolutely,” said Street. “In fact, we’re testing different communication engagements already. We are working with Alice on facilitating conversations directly between hoteliers and travelers. Right now it’s email based, but we’re looking at other delivery methods. And with the rise of messaging platforms around the world, we are open to expanding our work in this space.”
” We’re excited about how the work we’ve put into Messenger, and all the back-end work we’ve done with Natural Language Processing, could morph to other voice-activated search options. There’s a ton of innovation happening in this space,” said Street. “As we learn more from our initial tests, we’ll have a better sense of what we’ll do next.”