The technology is here. Yes. Though of course artificial intelligence will always evolve. That’s inherent: If computers can “think” in their own way, the foundational tenet of AI, they thus will learn. The more practical question to ask, though, is whether AI has arrived in the corporate managed travel arena. The answer: in small ways, sure, but AI won’t truly change the managed travel game until suppliers develop solutions that solve for travelers’ and travel managers’ problems and until business travelers and corporate travel managers come to trust the results.
Yes, AI Is Here for Booking, but It Won’t Look the Same by 2038
The current reality of AI for managed travel booking, according to Microsoft travel technology manager Steve Clagg: Booking tool developers are consolidating, in one place, data from multiple sources, and they’re angling to build the best interface with which “to throw all that content at the user.” That includes determining the most valuable options to present to the traveler. “Simple AI can remove the irrelevant ones from thousands of choices,” Clagg said. “No one is going to pick, like, Seattle to Chicago with two hops in between just because it’s like $50 cheaper. Why is that listed?”
Next up will be customizing those remaining, logical options for each traveler—really knowing the traveler and predicting his or her needs—plus delivering helpful, customized information to that traveler in the right place and time. But traveler preferences and sort order are not enough, he said. Clagg eagerly anticipates intelligence that accounts not only for his documented preferences but also for his past behavior, as well as information Big Data can glean about his type, what Microsoft’s travel program has been calling a persona. “It’s really only three choices a traveler is going to be weighing from” those thousands of flight options, he said. “You can get to those by understanding the traveler better.”
And within 20 years? An AI system will be able to “delight the traveler” by anticipating the traveler’s needs in advance—from recognizing that a person will want to take a trip to helping rearrange their schedule to making suggestions for things to do along the way. At this point, Clagg said, booking tools will be no more.
“Trust is key here,” said Microsoft travel technology manager Steve Clagg. His enthusiasm for AI lies largely in its potential to anticipate the traveler’s needs: recognizing an email conversation about getting together as the intent to travel, proactively searching for and proposing a minimal number of tailored flights, arranging ground transportation, helping to reschedule conflicts, suggesting a restaurant that suits the traveler’s taste the way Netflix recommends movies and so on. That’s stuff that personal assistants already do but that AI someday will do better as it integrates and analyzes data feeds. There are calendars but also booking history and expense history, which can reveal a traveler’s preferred way to get to the airport. AI could know that a traveler loves meatballs and assimilate that with Italian restaurants coworkers seem to like and with the restaurants whose meatballs the locals praise.
That’s all nice in theory, but travelers have to see it to believe it. “You trust your personal assistant because of past performance and capability,” Clagg said, “from repeatedly demonstrated value.” For AI, “that isn’t going to happen overnight.” Rather, the value of AI as a travel assistant will increase as travelers use it, as will their dependency on it. That will prompt them to use it more, and with that performance will improve. And so begins the cycle. Somewhere in there, trust will have been established, Clagg said.
United Technologies corporate travel associate director Colleen Kearney bears witness to that self-fulfilling cycle around hotels’ AI offerings. “We hear from travelers, ‘I got a notice I could check in on my phone. They knew what I wanted: away from the elevator or king-size, nonsmoking. It was very easy.’ There are comments like that,” she said, “that make that of more and more interest to our travelers.” She explained why: The AI in play is personalized and makes the traveler feel cared for, even though there’s no human interaction.
Kearney also has come to trust AI biometrics after seeing it in action herself. She attended a JetBlue event at Boston Logan International Airport, where the carrier has debuted facial-recognition for boarding. However, because it matches the passenger’s face to a passport photo on file, it’s limited to international flights. She wonders whether the technology would work with other forms of ID. “It’s a limited application in my estimation right now,” she said. Still, she described it as cool, easy and successful a high percentage of the time.
On the flip side, negative experiences can sow doubt. FINRA corporate travel services manager Carol McDowell has witnessed the frustration on her husband’s face as he talks in circles via their cable company’s online chat service. “It seems right now that it actually takes him longer using the chat box than if I call someone,” she said. “Sometimes with travel, you need that interaction and you need that personalization. You need someone to talk it through with you because you’re not really sure what you need unless you know the right questions to ask.”
PSG corporate meeting and travel manager Jennifer Brown is similarly skeptical of chatbots, at least as far as their facility for self-booking. “Anything that makes the booking process easier for the traveler I would fully support, but it has to be consistent,” she said, citing the fact that the Siri voice-activated assistant on her iPhone doesn’t always understand her. “That’s the one thing I’m a little skeptical about with AI. … I don’t know if it’s been fully perfected yet.
There’s Trust & Then There’s Faith
At some point, though, people just have to have a little faith and jump in, even if they don’t understand the nuances of how AI does what it does. “I don’t understand how I can throw a penny in a river and it drops to the bottom of the river and [yet] boats with cargo on it that weigh like a kazillion jillion times more than anything I could ever account for can cross the ocean,” said MassMutual travel services director Kelly Taylor. “But I know it works. There are just people who are awesome at that and that’s their niche … whether it be making a tanker float or making AI work.”
Kearney, who looks forward to AI’s ability to sift through reams of data and predict spend and traveler behavior, also has faith, at least enough to try it out. “I don’t have any reason to doubt it at this point,” she said of the results AI-enhanced predictive analytics might spit out. “We would have to see what they come up with and if it sounds about right. We would do some verification, at least initially.”
If there’s risk in not fully grasping how AI does its magic, it’s worth the reward in Kearney’s eyes. “There are ways that information is somehow being collected in the cloud, and they’re bringing information and data together and bringing insights that we wouldn’t normally have. We have all of these data sources that are hanging out there, and we’re manually pulling from each one of them and trying to lay them over one another and bring insights ourselves.” She dreams of a reality in which she can ask an AI-enhanced database how much a particular business unit will spend on business class flights for a certain time period and the system “suddenly brings a report up and it puts it in a very nice pie chart and I don’t have to sit here for hours inputting data from different sources into a spreadsheet and then creating a nice pie chart.” She added: “It’s a way for us to get our hands on more immediate data, then move forward with predictive anticipation of what our travelers are going to do, how much we’re going to be spending versus just looking at historical” data.
On the traveler side, Taylor believes, widespread adoption of AI is inevitable anyway, another reason to embrace it. She bases that assumption both on younger travelers’ consumer habits and on older executives’ business decisions. When Taylor’s 24-year-old daughter is planning a trip, for example, Taylor wonders whether her daughter doesn’t want to talk to people for advice. Instead, her daughter visits TripAdvisor and seeks out other online resources. “She can plan everything without talking to anyone,” Taylor said, “and it works.” That lays out in front of Taylor’s eyes a future in which machines can facilitate travel without human help.
In contrast, most MassMutual EVPs and C-suiters still avail themselves of concierge-level travel management company services; they like the high-touch interaction, she said. But behold! Those same decision makers are investing the company’s capital in AI. They know AI is the future, she said. MassMutual data scientists are applying algorithms to the company’s data to identify sales leads—customers that just experienced a life change like buying a house, for example—and the company also has launched Haven Life, which walks people through the process of buying insurance without human interaction. Taylor has no doubt the company will at some point turn the personalization AI capabilities it’s developing to internal operations like T&E.
Where the Big Corporate Travel AI Opportunity Lies: Personalization
Like the travel program at MassMutual, Microsoft’s travel department will benefit from its larger company’s predilection for AI-enhanced technologies. The company is investing a lot in AI cloud services, said Clagg. “Over the next five years, we’re going to see some amazing stuff come out of this company.” It’s all about using data to enhance the user experience. For eight years, Microsoft has been working on embedding AI capabilities into Microsoft Exchange, on which the Outlook email program and Office 365 software are built. Clagg points out that email is unstructured data, meaning it requires a human brain to interpret and analyze the content and determine a course of action. There’s a “ripe opportunity to understand content and semantics to turn it into structured content that can be automated and utilized,” he said. Booking confirmations from suppliers, for example, can be added to a calendar. Expanding that capability to travel assistance—blocking out time in your calendar for an airport security checkpoint, for example—is coming to fruition now.
Taylor still thinks most AI personalization innovations will come from the supplier side. At the moment, concierge-level TMC services are the gold standard of personalization—telling a hotel how to stock a room on behalf of an executive, for example—but those are available at a steep price and thus only for a company’s VIP travelers. AI, though, could automate those preferences for anyone; a hotel could know, automatically, that Taylor prefers sparkling water and down pillows. And it’s worth it for suppliers to build out such technology, she said, for the upselling opportunity. If a flight is close to selling out, for example, the airline could approach Taylor and offer her one of the last remaining window seats for $30—because they know she prefers the window.
“I’m pretty optimistic,” she said. “The data is out there. People are working very hard behind the scenes to come up with the algorithms, to come up with ways to improve not just personalization but also upselling because that’s a win for them."
So What Will Travel Managers’ Role Be?
“They’re automating everything. What do they need me for?” Taylor wonders. There’s time, though, to figure out how to make herself needed. “The concern is not for today or tomorrow, but … I have to keep growing as a person so that I’m not doing something that could eventually be automated and making myself obsolete. … I need to make sure that I’m providing value.” According to Brown, though, her value lies in her humanity. She looks forward to the improvements she believes AI will bring to her travelers’ lives, but she said, “People are always going to want someone to go to if they have a problem.”