Expedia Plans to Use Artificial Intelligence for Customer Service

How does one of the largest players in global travel view the burgeoning field of artificial intelligence?

As the tech world salivates over its game-changing potential, Expedia Inc. CEO Dara Khosrowshahi said the company plans to first use artificial intelligence for customer service rather than for something like trip-planning.

“For me to be able to message Expedia on [Facebook] Messenger or text … ‘cancel my hotel booking in New York next week,’ it is a much more delightful experience than calling or even my getting on the website etc.,” Khosrowshahi said at Skift Global Forum in New York City last month.

While leisure travelers might take on average perhaps just one or two vacations annually, artificial intelligence isn’t practical at this point for travel research and planning because artificial intelligence needs huge amounts of data to work, Khosrowshahi argued.

“We are going to take it on the service side,” Khosrowshahi told the audience last month in a session with online travel founders. “I think the planning experience and build for AI [artificial intelligence] has a real challenge because you are going to get a lot of null results. The AI isn’t going to understand what you say.”

Get Your Tickets to Skift Global Forum: Europe Now

Jay Walker, who founded Priceline.com and is currently CEO of business travel startup Upside, agreed that artificial intelligence makes the most sense in leisure travel to bolster customer support but he sees it as particularly beneficial in business travel.

“Business travel is a completely different set of problems and there AI might be able to be much more narrow and focused because a business traveler is doing something more repetitive, more predictive and in many ways more commercially driven from a result than leisure travel, which is much more discovery-based, more opportunistic,” Walker said.

Another plus on the side of business versus leisure travel is that business travelers aren’t as focused on price, Walker said, adding they expect high-quality service. “My time is money.”

Khosrowshahi agreed with Walker about the applicability of artificial intelligence for business travel, and the Expedia CEO made it clear that customer service for leisure travelers would likely merely be an entry point for Expedia’s tapping of artificial intelligence.

“Start with small use cases and start from there,” Khosrowshahi said.

Building A Better Travel Agent?

Artificial intelligence basically is the science of analyzing data to understand and predict consumer behavior and to make recommendations. It has similarities to and overlaps with concepts such as machine learning and Big Data. The terms are increasingly being used interchangeably in some quarters.

In the travel industry, one example would have artificial intelligence replicating great travel agents, offered Altimeter Capital CEO Brad Gerstner, a former Orbitz board member and ex-co-CEO of cruise and vacation seller National Leisure Group.

Gerstner stated that startups such as venture-backed Lola are trying to combine artificial intelligence, messaging and human travel agents but he predicted that a revival of the travel agent sector isn’t in the offing. Many travel agencies, on the other hand, state that there is a renaissance going on in their sector.

“This idea that we’re going backwards instead of forwards, that we are going to re-engage all these human offline agents to do this, I think, personally, is unlikely,” Gerstner said. “Simply because, I think, the power of Big Data. [Expedia] knows my preferences, they know how often I book, they own my calendar. Their ability to do predictive and suggestive search is incredible. And, by the way, they employ a lot of travel agents.”

Gerstner said startups will play a role in the evolution of artificial intelligence in travel but large companies such as Google and Expedia will drive the trend.

“When you start thinking about machine learning, Big Data, AI — whatever you want to call it – … the advantage is to the largest player because they have all the data,” Gerstner said.

A new study published today by Amadeus with the London School of Economics backs that theory, positing that “the growing sophistication of assistants and virtual reality will change consumer behaviour and shift greater power to those players who control the technology.”

Lola Is Supplementing Its Data

Contacted by Skift yesterday, Paul English, co-founder and CEO of Lola, agreed that it is easier to make personalized recommendations for frequent — rather than sporadic — travelers whether the trips are for business or leisure because there is more data to work with.

But Lola is adapting to this challenge by buying data and will soon allow travelers to import  their travel histories from players such as Expedia, English said.

“Given that Lola is a new company, we don’t have millions of our own transactions for our machine learning system to use so we’re supplanting this with data we are buying,” English said. “By allowing the traveler to import travel records, and by using heuristics I’ve developed based on my prior 10 years in the travel industry serving both business and leisure travelers.”

Among his ventures, English co-founded travel metasearch company Kayak.

Unlike Booking.com and Airbnb, Lola isn’t asking its customers — yet — whether their trip is for business or leisure.

“Lola — like most modern travel companies — does not require the user to tell us which occasion is business versus leisure, but instead we look at search parameters (number of travelers, days-out between booking event and travel date, number of days of trip, Saturday night stayover or not, whether they are traveling to a ‘business city’) to help guess the occasion, and to then personalize the results by occasion,” English said. “This is important since the hotel you want for business (where your company pays) is usually different than the hotel you want for leisure. We also look at number of guests, and make different recommendations for one traveler versus two (maybe romantic) versus three or more (maybe family with kids, thus, maybe a hotel with a pool).”

Along with artificial intelligence, members of the online travel reunion panel at the Skift Global Forum opined that voice-based search is another new frontier that the travel industry and others will increasingly grapple with in the next few years.

Here’s the video showing online travel founders at the Skift Global Forum September 27. The discussion about artificial intelligence begins in earnest around 26:20. [The rest of the video is a great listen, too.]

Get Your Tickets to Skift Global Forum: Europe Now