One Thing I Took Away From the Skift Forum

By Ryan Clarke

A singular thought stuck with me after the Skift Forum in NYC last month: the robots are here, and they’re inviting their friends. It was made abundantly clear, during those 2 days at the Lincoln Centre, that personalization through Artificial Intelligence and voice-based search will fundamentally change the way human beings search for, purchase, and experience travel.

The Forum, now in its third year, is essentially a “TED” for the travel industry. The speakers included founders and leaders from Priceline, Expedia, Google, Facebook, Marriott, TripAdvisor, Airbnb, Bunkhouse Group, Standard Hotels, Generator Hostels, among many others. The Skift team did an amazing job putting it together; you should all come next year.

There was much talk of recent consolidation in the hotel industry, the $30B elephant in the room, and of course, a few customary jabs at the OTAs. It occurred to me that this was just a room full of people trying to determine what the traveller really wants. To that end, almost everyone that crossed the stage expressed a desire to leverage machine learning to not only discover what the traveller wants, but in the actual delivery as well.

AI will affect everything from our use of planes, trains, automobiles, buses, to booking dinner reservations at a stranger’s apartment, day trips with a local, or an interesting place to work while you’re on the road.

In the past, I’ve been guilty of thinking of Artificial Intelligence as a new word for things that we already have. You know, like another, sexier word for algorithms. After all, we interact with machines and applications everyday that constantly make suggestions based on our supposed tastes and preferences. And the suggestions aren’t all that bad. Facebook is eerily good at suggesting a new friend the day after you meet them. I’m convinced it’s a location-based proximity trigger combined and cross-referenced with friends in common, but I digress.

Oliver Heckmann, VP Engineering at Google Travel, stated in his interview that the shift to AI and voice-based search over the next 10 years will be as big, if not bigger, than the shift to mobile we’ve experienced over the last 10 years. Google’s foray into hardware starts to make more sense.

This is way more than better ads and search results. AI and VBS are going to completely dismantle and rebuild the search function.

It’s difficult to wrap our heads around what exactly that will look like – but for the most part, it means interacting with our digital assistants (bots) via voice-based, unstructured and conversational queries.

For example, you could ask your assistant:

“I want to book a flight to Boston early Tuesday morning and fly back on Wednesday afternoon. I’ll need a rental car for the time that I’m there. The meeting is with ABC, so I’d like a hotel close to their office in the Back Bay. I’d like to schedule dinner with their VP, Jennifer, at the same place we went last time. Can you make reservations and send her an invite from my calendar? Oh and all expenses should be billed back to XYZ Co.”

This is a set of commands that has traditionally been transmitted to, and handled, by a human assistant – the logistics of which would probably consume the better part of their week. Soon, this will all be done by a intelligent, digital assistant, in seconds.

When broken down, the command string isn’t really that complicated.

  • Book a return flight to Boston on preferred airline
  • Reserve a rental car from preferred rental co. for 30 minutes after landing
  • Confirm ABC address and find a hotel from preferred chain within 10 minute walking radius from meeting address.
  • Make dinner reservations
  • Notify Jennifer @ ABC
  • Send all digital receipts during this time to special folder for this trip

What’s interesting is that the traveller may not see a single search result. In this respect, it’s important to consider the role of loyalty programs. The assistant would certainly know that the traveller was a preferred member at Delta, Avis, and Marriott, so there would be no real need to shop around for prices.



We already live in a world where airlines and hotels are using predictive pricing models and basic AI to fine-tune their profit margins.

Big companies have always had more money to invest in these systems, and as such, intelligent pricing automation has traditionally worked against the buyer. This created a market for tools that, in effect, work against the bots. These services that help the traveller find ways of tricking the machines and saving money.

But when we trade control for convenience – as human beings are wont to do – we will likely sacrifice the ability to compare prices fairly. The bots that work on our behalf are beholden to the laws of their world; laws written by Google, Facebook, Microsoft, Amazon – who ever gets there first.

These bots will presumably communicate and negotiate, on the traveller’s behalf, with their corporate counterparts. The question then is, will the bots work for the traveller or against them? And for how long? And can we ever be sure?

How will independent hotels compete with the chains and OTAs when the screen disappears?


Further Reading: Humanity and AI will be Inseparable

An optimistic interview with Mauela Velosa, Head of Machine Learning at Carnegie Mellon University. (The Verge)