Practical use cases for AI are delivering concrete, powerful business results, from reducing customer service costs to reliably helping deliver what a customer wants, when they want it and where. It’s time to get AI-savvy to keep your competitive edge. Join Kayak’s chief scientist and others at this VB Live event to learn how.

Register here for free.

“Data science is the core of Kayak,” says Matthias Keller, chief scientist at travel booking giant Kayak.  Since its inception, Kayak has used machine learning to improve the experience on their sites behind the scenes. Machine learning powers everything from sorting recommendations to uncovering the best prices at any given moment to recommending the best flights for your buck. But now they’re entering a new domain with natural language processing and computer vision.

With NLP, the company can process hundreds of thousands of user-generated hotel and flight reviews, condense them into a few snippets that capture the general essence of them all, and sort each item into the buckets users are searching for. NLP is also powering Kayak’s march into AI integration with chat bots and assistants such as Amazon Alexa, Google Assistant, Slack, Facebook Messenger, and more.

It happened organically, Keller says.

“Our project didn’t even start with saying, ‘We want to build an Alexa skill,'” he explains. “Our project started with, ‘We want to be able to understand natural language in the travel domain.'”

It began with building a piece of software that was able to understand conversational requests like, ‘I want a flight to San Francisco on Sunday for three days.’ And the team wasn’t even initially very sure where to put this type of system. They learned quickly that it wouldn’t be a good fit for web pages, but soon found a great fit in Slack and then Amazon Alexa. And it’s getting more sophisticated every day.

“We’re trying to make the assistant smarter, more personalized, more transactional, allowing you to book real trips, knowing who you are, knowing what you’re up to,” Keller says. “When you say, I have to be at this meeting, in a perfect world it would look in your calendar and know where that meeting is and find you a hotel next to the meeting.”

The vision is to ensure that the variety of ways a user can interact with Kayak become perfectly synchronized — from searching the websites to getting on the mobile app to using Alexa or Google Assistant or Facebook Messenger bots.

“All these devices should be able to understand context and know everything about what you’ve been doing,” Keller says.

For example, you start to search for a flight on your browser, and then you should be able to continue the conversation with Alexa while you wash dishes, checking prices and daydreaming about Hawaii. These connections are all intended, ideally, to create a personalized experience that keeps the user coming back for more — the perpetual challenge of getting the right information packed up in the right format at the right time, helping the user get done exactly what the user wants to get done.

The tricky part is ensuring that the responses measure up, Keller says. There’s the eternal issue of the consumer’s limited attention span — you can’t keep up an endless conversation about things that the user doesn’t want to have or that are out of context.

“We’re realistic here,” Keller says. “You’re not going to use our Alexa skill because you want to have a conversation. You’re going to use it because you want to get something done.”

Today, it’s finding out if your flight is on time, or checking prices; in the future, it’ll be asking Alexa to rebook your flights and reserve a car for you.

They’re not stopping there, Keller says. AI opens up the possibilities of information crowdsourcing — for instance, ways to assess the wait times at airports from our users in the Kayak app, to algorithms that help them understand which 10 different ways a traveler can get from one side of the airport to the other, and which is most effective in what situation and when.

“We want to do fundamentally new things in how we get and process this data to get actionable results to users that can help them tremendously during travel,” he adds.

To learn more about AI use cases that can transform your business, and how to get started, join this VB Live event featuring Forrester Principal Analyst Brandon Purcell and Kayak Chief Scientist Matthias Keller.

Don’t miss out!

Register now for free.

You’ll learn:

  • What technologies fall under the AI umbrella
  • How companies like Kayak use AI to understand customers and personalize experiences
  • How to identify the right AI use case for your business
  • Common challenges firms face when implementing AI
  • Why AI’s time in the sun has finally come, and why it’s here to stay


  • Brandon Purcell, Principal Analyst, Forrester
  • Matthias Keller, Chief Scientist, Kayak
  • Martine van der Lee, Director of Social Media, KLM
  • Pradeep Elankumaran, CEO and Co-Founder of Farmstead
  • Rachael Brownell, Moderator, VentureBeat