Could a robot book a cheaper itinerary than you?News / OnlineBy Sean O'Neil | August 15, 2016Share This article was originally published on Artificial intelligence (A.I.) isn't a dream from the distant future. Machine learning is a method of data analysis that allows computers to identify and adapt to patterns in user behavior. This form of A.I. is already used to recommend new music, detect credit card fraud, and increasingly, help speed up travel planning.NB: This is a guest viewpoint from Dan Ruch, CEO of Rocketrip.A.I. (in essence, a robot) could help businesses reduce their travel costs by millions of dollars. But that will only happen if new technologies override the existing "default settings" that subtly influence travelers' decisions.Airlines, hotels, car rental companies, and online travel sites incorporate preset options strategically throughout the booking process. These default settings have dual purposes.The first is to increase conversion rate: travel sites strip away as many barriers as possible that might discourage a shopper from turning into a customer.The second is to maximize the revenue per transaction: options are arranged based on a shopper’s likely price sensitivity. Business travelers are notoriously price insensitive, and therefore more affected by how booking tools are designed.Let’s look at a few examples. When you search for hotels on Expedia or Skyscanner, the default setting ranks results by “most popular.”Conversely, flight and car rental search results are sorted from lowest to highest price The default settings assume that you are less price-sensitive about hotels and more price-sensitive about flights and cars.On other sites, you can search by criteria of your choice. Kayak, for one, let’s you search for hotels within a given price range, but by default, the site still organizes your results by “Recommended.”Why?Well, why would a commission-based business remind you to put price ahead of popularity? Starting with luxurious, expensive hotels shapes your expectations of what’s reasonable. If the first four options are $500 per night or more, the $300 hotels seem like good deals.Note that Kayak places “search by price” under a not-obvious drop-down menu, and Skyscanner, too, places its hotel sorting options in an inconspicuous location.Default settings are not designed to produce the least expensive itinerary. In these examples, you can see how designers push and pull the levers of human psychology.Artificially intelligent travel planning apps don’t have those levers. That’s why AI could become an ally in reducing corporate travel costs. It could make saving the “default setting.”However, this frugal AI would have to overcome certain hurdles.First is understanding price in its correct context. An AI-enabled app can easily find you the cheapest itinerary for a business trip from New York to Chicago, but you might hate the result. Why did it put you in a middle seat when aisles and windows were open? Why did it choose a hotel so far from the client you intend to visit?Instead of selecting the cheapest options, travel AI could select options a certain percentage below the average price for a given itinerary. The default savings rate for AI could be 30% below the average, and travelers could toggle that number up or down to find their happy medium between cost and comfort.Second, AI apps can mimic users’ booking decisions, but may not understand the underlying reason for those decisions.For instance, would an algorithm recognize that you always book hotels near public transit stations? Or will it detect a different pattern? Maybe all those hotels have fitness rooms.What might be an obvious choice to human beings will not be apparent to many algorithms. For trickier patterns, AI users might have to set dozens of booking rules manually.That’s the third hurdle: travelers will have to invest time in “training” their AI. Coaching an app like Pandora is easy – it’s thumbs up or thumbs down. Travel booking will require more thought and effort.Will business travelers churn or stick it out? That will depend on their incentives. How badly do they want to save?AI booking tools like Pana, HelloGbye, WayBlazer, and Hello Hipmunk, have gotten buzz for their potential to make booking business travel more convenient. With thoughtful implementation, this new technology can make business travel less expensive, too.NB: This is a guest viewpoint from Dan Ruch, CEO of Rocketrip.Image by Maridav/BigStockPhoto.