KAYAK CEO on Why AI Won’t Replace Travel Search

0:01Hi, everybody. Welcome back to Skift Take Sessions, where we are live at the Skift Data and AI Summit here in London. And I'm sitting with Peer Bueller, who is the CEO of Kayak, but also not just the CEO. You were formerly the CFO, the COO, so you've been at Kayak for quite a while. Peer, thank you for joining me. Thank you very much for having me. Of course. Well, I'm excited to kind of geek out a little bit about data and AI, obviously. But one of the questions I want to ask with your new role being a CEO about the last eight months or so, I want to know what does your CFO instinct kind of kick in about the investment of AI now that you're in the CEO role at Kayak?

0:38How is that really like that shift and transition work for you? Yeah, it's been it's been a pretty wild change. I've been with Kayak for a very long time. I was very involved, as you said. I thought I knew exactly what this job entailed. Yeah. And I got it wrong. in a really good way. Okay. And having a blast. I'm involved in anything and everything. I like looking at numbers. I was never a hardcore accounting-centric CFO to begin with. Yeah. So I'm the same person, treating things the same way. Suddenly exposed to a lot more stuff though. What would you say is like one of the key things that you got wrong that you didn't expect?

1:18Probably the level of involvement. Okay. Um, I, I always thought of the CEO as the, the tiebreaker. Uh, I, I have a team of phenomenal people. They're really good at what they do, much better at what they do than I would be at what they do. Um, and if they don't agree in my mind, they would come to me and I would help decide, uh, who gets their way and who has to disagree and commit. Um, in reality though, I'm way more involved earlier on. And I'm at the table when we're having conversations about brand, about product, about development. I'm loving it, but it's really a level of involvement that I didn't see coming.

1:56Yeah, I was going to say probably a lot more meetings, maybe, give or take. Yeah, we're aggressively anti what we call stupid meetings, which is a cultural thing to get a grip on. It's a huge difference whether you cut back on the meetings and don't really add value. so you can devote and double down on the meetings that do. So maybe a little bit more meetings, but we still try to keep it light there. Yeah, I like it. Well, you guys launched Kayak.ai about 2025, I think May from what I remember, prior to which has just been launched, Lola at Booking Holdings, but also a lot of other AI models in the OTA space.

2:35I want to ask you just some key learnings because that was quite a while ago in the AI world, right? Like a lot has changed with the models. a lot has changed obviously probably with the data um where you know a lot of the traffic is coming from so i would love to learn from you key learnings with kayak.ai yeah there's there's a lot to share i'll try to just point out a couple of things kayak.ai is a url yeah it's a website um we did it that way to keep it out of the middle of the river okay we're running a big company we want to go full speed if we suddenly ask everybody to spend a day a week on ai we're just distracting everybody and not utilizing the people we should be utilizing full force.

3:15Yeah. So we tried to fence off a URL, a production environment, and a little group of people to try to move really quick. That's worked fantastically. I'm really proud that we did that and a great learning, a very positive learning. The next step, taking kayak.ai and pulling it into kayak.com to make it visible to the real masses of people is tricky. And I think we figured it out by now, but it was a bumpy road. And I think that's a learning, too, that you just got to be realistic about it. I would do it the same way if I do it again. You can speed up by keeping it separate. Integration is not a given.

3:55It's not a flip of a switch. And then third, probably the biggest learning, two years ago, three years ago, I thought, together with everybody else, soon enough, everything's going to be AI. Maybe everything's going to be agentic. and we'll just be sitting in a beach chair while our agents are making our lives work. The reality couldn't be more different. People love classic meta search interfaces. People need to look at options in order to create confidence for themselves so AI can play a huge role. AI is not replacing everything in meta search. So what we're doing right now is a bit of a hybrid user experience where you have classic meta search and an AI conversational interface side by side.

4:40So you can jump back and forth, which I think we got it right there. It'll be moving again, but that's a big one. Well, in the OTA landscape, I know Metasearch is kind of a subcategory, but would you say that Metasearch kind of has the upper advantage compared to maybe a bigger OTA that doesn't really compare quite across a wider scope? In a way, for sure. Yeah, especially through the AI. entry path. I think LLMs, big LLMs have educated us to become conversational with computers, to start searching the way I speak, expecting responses the way I would be spoken to. Metasearch has always done that a little bit more than OTAs that are even lower funnel, even more focused on just tell me the dates you want to travel, your destination, your origin, and I'll give you exactly the one you need.

5:32Yeah, 100%. Well, I want to ask you, because I think that's a good follow-up for this is that I want to ask, where do you think Kayak is leading in the booking holdings group? I know AI, obviously, you guys got early in May and now there's another product inside of the booking holdings category families. But where do you guys feel like you're really leading in the charge of the booking holdings family? Well, I got to say that we're a little bit of an odd duck in the group. And I really like it that way. I'm a big fan of the portfolio approach. I love the brands that we're sister brands with. Um, and I think we're being empowered to, to run independently for full speed and to create just the right level of coordination to make sure that we learn from one another, that we share learnings, that we're not in each other's way.

6:21That's really easy for kayak. We're the only meta search player. Um, we've done this for a quarter of a century. We're obsessed with what I've started to call the art of comparison. the visuals, the multiple options, the cross vertical from flight to hotel and back. All that stuff is quite unique to Kayak. So it's something we do. We're also small and fast. We can get a lot of stuff done quickly and give it a try, just like Lola, just like the smaller groups inside the portfolio and take learnings from that. Yeah. One of the questions I wanted to ask you was also a lot of the talk around AIs, around LLMs.

7:01but now you know muse has launched instinct has launched all these agents are actually becoming a real thing i feel like four months ago the conversation was agentic when it comes but now as seth has said on stage it's here do you think agentic is going to have any real impact in the way people not only just travel but like the way you guys operate are you guys going to be building for agentic rather than llm or do you see them as kind of one the same no i don't see them as one of the same and i think it should have a real impact if it doesn't yeah then wow the hype curve yeah proved again that it can go wild and the and disconnect from reality um but i do believe it's a big deal and i do believe agentic is real again i i don't believe in two years yeah everything's agentic at all okay i think there's a use case um there's a time and a space for it where it can really fit kayak will have to adapt and go with it we'll be building kayak to work for LLMs the way we've always made it work for search engines.

8:02We made it work for direct travelers and we'll make it work for agents in the future as well. Yeah, I love it. Well, I think that's really all I'm going to ask you today. I'm going to keep you short and sweet. Don't want to throw too many curveballs around the day. I know you're going to be on session today. So before your stage session today, for the viewers that aren't here in the room, can you give us a high level 30, 60 second half of like what you're going to go into? What's a main thing that they should take away from your session? in the room. Oh, that's good. Often with these kind of things, they try to tell me exactly what's going to happen like you did earlier today.

8:36Yeah, I tried to. And I always say that don't, just let's go. It'll work. So I don't know exactly because I've tried to fend off big detail. But we'll talk a lot about what Kayak's doing in the AI space, both on the consumer front, which is really exciting for people that love the meta search experience that want to create clarity in a world with overwhelming amount of options and man, can I bring a bag? Do I have a chair? When do I check in? This is hotel, does it have a gym? Like all that type of stuff is meta search sweet spot. AI is going to make that work better. On the corporate travel side, Kayak is doing a lot with AI where the use case is probably more powerful even than it is on the leisure travel side.

9:18I'm expecting a question about what's life worth after you take over from Steve? Yeah, yeah. Because that's what everybody asks and fair enough. Yeah, those are the big ones, I think. Yeah, I got to chat with Steve Hafner last year at our Skip Level Forum. Fun dude, interesting travel habits. So I'm curious, what's an unexpected travel habit? This will be my closing question for you. An unexpected travel habit that you wouldn't expect a travel company CEO to have. I'm a hardcore window person, whatever that is. um i oh i book hotels far away from places that i need to be why is that because i love walking okay um and i think it really helps me so i booked a hotel it's uh close to it's called it's close to chinatown soho like near soho square okay uh took me 40 minutes to walk here this morning that's i i aim for that because i think that keeps me very awake and sharp my biggest fear is a conference in a big hotel where you have to sleep in the hotel run into people all the time in the elevator that are at the conference as well I'm a bit of an introvert and I love my walks so that's probably the big one I'm the opposite I'm like 8 minute walk that's like the max I could do so that's where I'm at today so I'm very different you see this is why meta search matters because we're all who we are we have our own preferences so whatever an LLM is going to tell both of us is good for us, for one of us, it's going to be wrong.

10:51Yeah. And hopefully not a 40 minute walk if you're not a long walker, but if you are, you know what to do. Uh, well, Pierre, I really appreciate the, uh, the time and the conversation you guys gave us today. And, uh, I'm excited to see your session. So for anyone who's missed it, check it out on skiff.com later, but, uh, I guess. Thank you so much for having me. Really good fun. Thank you. Bye.

11:19Thank you.