Travel’s AI Problem Is Bigger Than Most Companies Realize

Subscribe Apple Podcasts | Spotify | YouTube | RSS

Sarah Kopit and Adriana Lee break down why the travel industry may not actually be ready for AI despite the rush to adopt it.

They explore fragmented hotel data systems, why travelers still do not trust AI booking tools, and how companies are starting to optimize not just for customers — but for AI agents themselves.

Watch This Episode

AI Readiness Reality Check

The number of the week is 22. That is 22%, and it is the chair of hotel chains that actually have centralized data structures ready for AI. And that’s even though 78% are already using AI in some capacity.

So today, we are going to talk about all things about artificial intelligence.

0:48

Meet Adriana Lee: The Guest Co-Host

Seth is actually out this week. He is enjoying some well-deserved time off. So I have a new co-host with me today.

Adriana Lee, who is our travel tech and AI reporter, is subbing for Seth. And it’s kind of perfect timing, because we actually have a data and AI summit coming up next week. Adriana and I will both be there in New York City.

So if you haven’t got your tickets, this is my plug for that. We may plug it a few times in this podcast. And so Adriana, you’re relatively new to Skift.

How long have you been here now?

I joined. Yeah, no, my debut was at the Global Forum last year.

The Global Forum, I was going to say. Yeah, it was about a year ago. Right.

That’s right.

I broke my shoe and you loaned me yours. I’ll never forget that. Yes, I remember that.

I’d forgotten about that.

But yes, I think I tanked your appearance because you were running around going, where’s Adriana?

I need my shoes. And I felt immediately bad. But I will always remember that about you, Sarah.

I must say, I have no recollection of that.

So that’s probably a good thing. I mean, I remember giving you my shoes, but I don’t remember running around needing them. So there you go.

So we’re getting ready. We’re prepping. We’ve got a list coming out of at the top.

So I’m going to ask you, if you’re an AI operator in travel, why don’t you just give me your high-level overview, just a little snack, just a little bit of where travel is within the AI ecosystem right now, maybe compared to, we won’t compare them to

Well, please don’t.

I mean, the industry.

2:34

Hotels Data Fragmentation

Well, it depends which part of the industry you’re talking about. Right now, that 22 percent figure with hotels, I’m kind of skeptical. I mean, I would be shocked if it’s that much, honestly.

Yeah.

What I’m hearing, yeah.

Like they’re not doing even as well as that.

I don’t think so.

And I don’t have something concrete to pin that on, but it’s based on lots of conversations with lots of folks.

Yeah.

On exactly how fragmented and disjointed those systems are. And it’s like if you don’t have kind of the data structure necessary, then you can’t really fully take advantage of AI. So that 22% figure seems very optimistic to me.

But then when you think about even if everything went perfectly, and you’ve got your structures in place, and your data is clean or what passes between, and you know what an NCP server is, and you’re doing all the right things, even then.

Well, so okay, that was your number of the week.

3:40

Why Only Two Percent

My number of the week is two. And let me tell you why. Yeah, let me tell you why.

Because so Skift Research found that fewer than a quarter of travel companies are using GENAI widely, and only 2% say the same about a GENAI. And also, 2%, only 2% of young leisure travelers are willing to let AI book for them.

So if you’re talking about younger travelers that are gonna be eating with these systems, that presumably know it better than anyone else, right now, only 2%.

So even if you can cut everything perfectly in place, the return on investment is not there yet. But we’re not open.

Yeah. Agents are kind of tough, though. Seth and I, we talk about this every once in a while, and just going back to our research about this.

But his thought, and I will parrot him because I think he’s right as well, is that, especially for Americans, we only get 10 days of vacation a year, right? So, our vacation, not only, let’s just forget about the cost of the actual trip itself.

Let’s talk about the value in time of the investment that you’re about to make when you’re planning, or booking, or buying that trip because it is so scarce that time.

Unless you’re a power vacationer who’s doing it all the time, or maybe even a work travel.

I would say work travel is probably more of the use case that might let an agent just do the booking for you because it’s just like, well, where am I going this week? I don’t really care. I don’t really care that much.

If something goes a little bit wrong, whatever. But for a vacation, it’s not whatever. If it doesn’t do it right.

Some people save for years for that trip, and they don’t want to see it taint because the tech pipes weren’t really put in place properly.

And that is a big concern with AI, because for most consumers, they’re at a catch point with AI. It has to do with tools that hallucinated them constantly. And so it’s like now you’re going to trust them with a trip worth thousands of dollars.

And it doesn’t really matter that enterprise AI is different, that there are guardrails.

It’s like if that experience doesn’t immediately convey to the traveler that this is going to be different than the craziness you’re dealing with with your personal chatbot, then how do you fill the trust?

And I think that is the big reason why everyone’s focused on trust right now.

Yeah, I do think though, just from like we both were around at the dawn of the internet age, I only think it needs to work a couple of times for that trust to be built.

6:18

Trust Guarantees and Risk

Like once it is used and does work once, maybe twice, then everyone’s like, oh, well, okay, let’s go. Let’s use it, you know? You know, it’s like, I mean, do you remember when like, it’s like you’re going to put your credit card on the internet?

Yeah, I remember.

Yeah, your credit card. Are you really going to log in to your bank on a website? Wow, oh my god.

It’s funny now, but see, here’s the reason why we trust that so much more, because there are guarantees in place if something goes wrong or covered.

There are promises, there are policies that stand behind us, there’s legislation that stands behind us. We don’t have all of those things in place around us right now.

So you’re literally going on a wing and a prayer with some of this, and it’s not clear who’s going to stand behind that. It varies from platform to platform, and sometimes nobody will guarantee that for you.

And so then you have a problem, what do you do? And that’s if it’s a minor problem. If you have a major problem while you’re in trip, and you’re not clear which platform is going to stand by you and help you, that can be a really major problem.

So there is that. But I mean, so I got pushed back a little bit. Yes, if something goes right a couple of times, I don’t necessarily think it earns your trust, I think it earns your benefit of the doubt.

And so you’re willing to give it more chances because, oh, wow, I had a positive experience. But conversely, you have a couple of negative experiences. That’s enough to take the whole thing.

So I think you got to look at the whole scope of what that experience would entail. It’s got to be a track record of success. And I’m going to tell you one more thing, one more thing before we switch topics.

8:15

AI Search Fails in Practice

I know a quantum physicist who is obviously not as one does. Yeah. Well, I live in the Bay Area, so it’s lousy with scientists out here.

So this quantum physicist is not new to technology at all, very much a tech head, obviously. He told me that he still doesn’t trust AI. He keeps trying on this platform, that platform, this tool, whatever.

The problem is that when he does his own searches, the searches he gets from AI are different, half the time or more than half the time, don’t actually exist in the real world when he goes to try to book or does his own search to corroborate those

results. And so he’s doing A-B testing. The minute he starts to see that the results are actually real and that those results are discoverable, that he can corroborate it independently, then he’s willing to give that more of a chance.

And I think that is very much a scientist mindset, but I think a lot of consumers think that way too. So I just found that really interesting.

I tried it myself, actually, on some recent trips checking in Claude to see the flights that I found that met my needs actually paired there, and they didn’t. They showed me other things, but not the ones that I needed.

I had to find that out myself manually. So until that gets solved, I think we’re gonna have this issue. But I interrupted you, so please go on.

Oh, no, I was just thinking, maybe we wanna have your quantum physicist friend on the pod someday, because I think that Q Day, which is coming at some point, is so fascinating, has nothing to do with travel.

It just has to do with everything. It has to do with the world. It’s like the day that quantum physics gets, the computers get to the point, because of quantum physics, that everything becomes unencrypted.

All of the protections that we have everywhere, on every little machine with a password. I mean, when I used to work in crypto, like we talked about this a lot too. Like it’s going to break the blockchain.

So I was just reading an article about Q Day coming up in just a few years. Some people think it’s coming. So, you know, buckle up.

Here we go.

10:34

Mews and SiteMinder Integration

All right.

So let’s talk about Muse too.

You had an exclusive this morning. Tell us a little bit about what they’re up to.

So Muse is really interesting. I’m sure our audience knows exactly who they are, but for anyone who doesn’t, I mean, they’re basically, they deal with hotel operations.

There are a platform that in some ways, you can think of almost like as a Shopify for hotel management. Like there are different, it’s a big old platform. There are pieces of it that people can take advantage of.

They frame it as a modular sort of structure. So what they did with SiteMinder is basically, they had this partnership to directly integrate. So it is a native integration.

It’s not like an app or plugin that you download and it’s optional. It’s actually, it’s right there baked into the operating system.

So if you want to take advantage of their distribution products and their features, you can do that right in the same place where you’re managing your hotel.

So they have, I think in the overlap, I think SiteMinder has like over 50,000 clients, but I mean, the ones that also use Muse, that overlaps not huge, but it’s about 3,000 or so.

So this rollout is going to hit them first, but that is not where they want to end. They see this as a real differentiating feature.

And it grabbed my attention because there’s so much, we were talking earlier at the top of this segment that, you know, there’s so much fragmentation to any siloed system.

So when you see something like this happened, where they’re coming together and trying to create more of a unified situation, and it’s not just user experience. I mean, these systems share data.

So Muse gets to see a little bit more about what’s happening in the distribution end. SiteMinder gets to see more of what’s happening on the operation and where revenue is getting derived and what’s affecting availability and etc.

So because these systems can communicate, they both have a clearer picture of what’s going on.

And so that is a sort of kind of cross-pollination of data that presumably would be very helpful in terms of preparing for AI, because that’s what they want to see. They want to get a bigger picture, a better picture of a hotel.

And that’s going to help the agents find them and recommend them.

Yeah. And Muse’s CEO, Richard Volter is going to, he’s actually closing our data and AI summit next week. He’ll be talking with Raffa on stage.

So get your tickets, plug, plug. Plug, plug, plug.

Come and see us. And Richard is a great speaker.

Yeah. He’s a really interesting dude.

Yeah.

He’s super fun to listen to, I don’t know, a pine on any and all things that are going on in the travel world.

I agree. I like to throw open questions at him and just sit back and let my hair get blown back and down. Like, wow, that’s interesting.

The point of view. Yeah, exactly. And if you can catch them, you should because it’s a fascinating perspective.

13:35

Business to Agent Era

And so I also, you know, we talked a lot about agents up at the top here, but I’m also really curious to kind of pick your brain a little bit about this idea that in Skift also, we had, you know, we had another story about this by you from, you know,

a few days ago about how, about the so-called B to A. So in the, you know, for most of our audience, I, you know, I don’t know how many travelers come here and listen to us.

I’m going to just assume that, that, that we are all sufficiently B to B, B to B fans in all of our, in all of our wonkiness. But, you know, we talk about B to B, business to business. We talk about B to C, business to consumer.

But now there is this kind of word out there in the zeitgeist, B to A, which stands for?

Business to agents.

Agents.

Yes. Agents. Tell us about this.

Yes.

So if you’re thinking about what appeals to people, what appeals to customers, you got your B to C locked in.

If you’re thinking about what appeals to business partners or potential clients, then that’s where B to B comes in. But if you’re thinking about what appeals to AI, what appeals to these agents, that’s B to A right there. It’s not a new term.

I mean, it’s been around for a little while, at least a couple of years, but it’s kind of been the sort of niche abstract thing that was really hard to kind of get your mind around.

And the reason for that is because the technology wasn’t really there for this to become meaningful yet. Well, technology is starting to get there. And so this term hit the Expedia Explore stage, so.

Oh, and you were there with Dennis last week.

Yes, correct.

Yes.

That’s right.

That’s right. So they were talking about Vita A, and it’s all about what appeals to agents. And so if anyone out there is curious what that means, basically you might have the best travel business in the world.

But I mean, if a lot of travelers are going to AI tools and they’re having those agents go out and find results for them, if those agents can’t find you, then you become invisible in those surfaces.

And it’s not even a reflection of how good your business is or what you offer. It’s just that they can’t find you.

So there’s a tremendous amount of investment that these companies are making to make sure that their businesses, their offerings, their inventory is available and searchable, and not even just available, but what is actively going to draw them in

there. And so a lot of that has to go to answer engine optimization. You might have heard of this, AEO. That’s part of it.

There’s a ton of things. So now we’re building what some very smart people think is not even a new phase of the internet, which was built for people. It’s not even an escalation of that.

It’s almost like a branch. This is now a new internet that’s starting to be built for agents. And so yeah, the mindset is just crazy when you think about it.

It’s funny, I was just reading last night.

There was a DigiDay piece about the economist. And pivoting a little bit just to journalism, but the piece was about how the economist is now basically bifurcating their audience and doing exactly what you’re talking about.

I don’t know if it was necessarily with agents per se in the way that we’re thinking about AI agents, but basically, they will be writing one piece of content for their human readers, and they will be writing an entirely different piece of content,

optimized just for AI. No human being will read that particular thing. I mean, I think the idea is to form a loop, so the agents will feed the human, but it’s for discoverability and just how do you do that now?

That is what the economist’s plan is as of right now, is to do these two completely separate pieces of work, each optimized for its own audience, one a computer and one a human being.

Whether there’s, and this is what the Digiday author noted, whether there’s money on the other side of that remains to be seen.

It is, yeah, in the travel world, right? So that sort of mentality is what everyone is trying to figure out.

And it is interesting, because I think Ariane Gorin said that less than 1.5% of their traffic comes from some of these optimization techniques. So it’s really, really small, but they’re such a huge company.

Even that tiny, tiny percentage is actually meaningful. And so they’re investing in this heavily.

And here’s the thing, what’s really hard to balance is you can’t completely go all in on agent and not build experiences, interfaces, content for humans as well. So now is that bifurcation you’re talking about, right?

And what was hilarious was OpenAI was on that stage. And, and they were saying-

Who was there from OpenAI, do you remember?

Oh gosh, I’m going to mess up her name. You know what, let me, let me dig that up so I make sure I don’t mess that up and I’ll get back to you.

But yeah, she, so basically, OpenAI, OpenAI was saying that they built this shopping experience and the people didn’t want it. And I thought it was such a blunt, interesting admission.

It’s like, so you’ve got travel trying to figure out how to appeal to AI, and then you’ve got AI trying to not script the human experience. So it’s like these two things are sort of moving in different directions.

Some of the things I actually kind of love about Silicon Valley, they’re like fail fast mentality. And it’s like, yeah, if people don’t want it, we can’t make money doing it, like we’re just going to scrap it and move on. That’s right.

It’s a very specific kind of attitude towards tech, and it’s very American as well. Something that I also have learned. I have a friend.

Now I’m just going to opine. I have a friend who is German and he lives here, and he’s a scientist as well. He’s one of those, like very close to the quantum physics wonks.

And he said one of the things that he just loves the most about Americans is about how we are just willing. Not even willing, like it’s a good thing.

You go out there, you build your thing, and it doesn’t work, and you tried it, you’re like, okay, it failed, I’m going to go try something else. And people are like, oh, good for you. And he’s like, that is so American.

That is so not German. It’s like, you know, your thing failed and your mother is like crushed and, you know, can’t show up at Christmas. But you’re right.

It’s interesting to hear when people, and it’s so prevalent mostly on tech stages when they’re like, yeah, it didn’t work. Didn’t work. We’re pivoting.

Okay, so I’m inspired by your story.

So I think I’m going to butcher her name, but I think it was, I think, I’m not sure if it’s Mayhak or Mahak Sharma. Okay. Okay.

I was very reluctant to scrub people’s names, so I was going to kind of hang back, but your story inspired me, so I’m not afraid to follow my faith. Mahak Sharma, she runs Product Partnerships over at OpenAI.

Awesome.

OpenAI, which I think has a new CMO right now too. So they’re really trying to kind of like solidify for their B2B moves.

21:00

Expedia Explore Takeaways

What was the most interesting thing that you heard other than this at Expedia Explorer last week?

What struck you the most?

Well, this story is not out yet, so okay, if you’re tuning in, here’s a little preview. I actually got to sit down with Tabea Matrine, who is the Chief AI and Data Officer. He started in December and they gave me all the time with him.

So I got to chat with him, try to understand his thinking a little bit more and how he’s guiding the organization. And really, he comes from some deep experience at companies you’ve heard of running AI.

And so he’s bringing all of that into Expedia, and it is for sure shaping the way he’s running their AI work. So that was a really interesting conversation. I can’t wait for that story to run on the site.

Yeah, I can’t wait to read it.

Yeah.

Yeah. So for me, that was a standout moment of the show.

Yeah.

21:54

Top AI Operators Insights

And we’ve also got a list coming out by the time this podcast airs, which we will have published. And it’s all on about the top, I think we have 13 people this time who we found.

It’s the top folks in AI, but it’s very specifically limited to operators, the people who are actually building the stuff. It’s not like the consumer, and pretty much all enterprise, not consumer facing.

They’re the ones that are really doing the plumbing of this in the back end. And you had an integral part of putting this list together and talking to a lot of the folks who were on it.

Did anybody tell you anything that you either didn’t know or didn’t expect to hear? Any surprises from that?

I think the only surprise, and I don’t know if this really counts, but I do think that when people have such weighty responsibilities, who they are as people, they bring that to the job too.

So it’s not always about the technical knowledge that you have. But for example, what struck me is Chavi is energetic, but also well-spoken, well-considered. He is very thoughtful about the way he’s approaching what he’s building.

And as I’ve mentioned, Chavi is on that list of the top leaders list. So look from there too. When I spoke with Airbnb’s Aldal, who runs data over there, he is energetic.

I think he was a million of a million things, and they gave me a few minutes with him, and the energy just came through. But he struck me as a guy who’s balancing 50 plates spinning on sticks.

And so I loved the energy, but it was very much like, oh my gosh, there’s so much going on. So that struck me. And Gail, she at Amadeus, he’s dealing with data governance.

I feel like, well, number one, it’s great to see a woman in that role. Number two, governance is not talked about enough. It really matters in shaping how much data, what kind of data, what type of guardrails are you going to put around your AI?

How much access to your business systems are you going to give it? These things matter. And so when you’ve got a huge organization like Amadeus, there’s a lot that can go wrong and screw up for a lot of businesses and a lot of people.

So what she does is so integral. So I’m excited to see the full list, but these are people that stood out to me.

Yeah. And I will say the best piece of journalism, or at least the most interesting piece of journalism, other than everything that we produce here at Skift that I have read this year, is Ronan Farrow’s piece on Sam Altman of OpenAI.

And that was all about Sam Altman as a person. That had nothing to do with the technology.

And I think your point is very well taken, because these folks, like the people that are on this list, the people that we talk about in our pieces with AI, they are, I mean, we talk about the computer overlords, like they are the computer overlords,

right? Like they are the ones who are creating the machines and the systems and the processes that really are going to shape our lives, our children’s lives, you know, not to be incredibly melodramatic.

But, you know, we are at this, at this, you know, beginning stage. It’s hard to believe we’re only, what, three, three years, three years in, for almost four years in, from ChatGPT’s launch. And just everything that’s changed since then.

Look, I’ve said this before, I have covered technology for a lot of years, and I’ve never ever seen anything like this.

The speed of development, the speed of adoption, this is really unique. I’m gonna steal something, though, that a former editor of mine told me once, that I’ve adopted. And he said, look, we cover tech, but we don’t really care about the tech.

What we care about is people. We care about the people that are being shaped by this tech, that are using this tech, that are building this tech. Everything boils down to the people.

So when you approach a tech story, yeah, you can make it dry and technical, sure. And there are a million places for that, if that’s what you want to read. But we’re going to tell you why it matters.

And it has everything to do with people and effects. There are winners and losers being carved out because of these technologies too, and that matters as well. So I love that thinking.

I’ve adopted it for myself. I’ve absolutely stolen it.

And it has been really helpful as I cover travel tech and AI for Skift, because I always remember there are people at the end of this thing, whether that’s going to be the travel advisor or that traveler themselves, that executive that is putting

their stake on this thing. And it all boils down to people.

So you just led me right in there so easily.

27:06

Winners Losers and Wrap

We’re going to do our winners and losers now. This is how Seth and I always end the program, always talk about our winners and losers of the week. They do not always need to be travel related, and I’m going to go with my winner this week.

And it is not travel related, although I guess it could be maybe live tourism related. I am talking about the New York Knicks, baby. Like here I am in New York City, New York Knicks are going to do that.

They’re going all the way. It hasn’t happened since 1999. New York City is all ablaze in blue and orange right now.

So Jalen Brunton and the Knicks, they are my winners of this week.

All right. Well, I will light a candle. I’m an ex-New Yorker myself.

I’m in the Bay Area now, but I mean, man, I used to live in Chelsea, Williamsburg, all around the place. So I’m right there with you. So okay, let me back into this a little bit.

I think the big winner right now is Airbnb, just because they are so active and so much in the narrative at the moment. And that I think makes the loser of the week. Everyone who owns the pieces of the trip that Airbnb is now moving into.

So Airbnb, they were pushing into home sharing. They start from home sharing into like hotels, car rentals, airport pickups, luggage store, you name it, right? And so even like grocery delivery, I think.

And so they’re turning themselves into like this one stop travel stack, rather than just a place to rest your head. So I think the people who lose in that story, it’s not just one company, but it’s like a whole set of incumbents.

Now, there’s a caveat. I mean, obviously, everything depends on execution, what actually happens in the numbers, etc. But I think they’re staking a claim to something interesting and kind of large.

I do the newsletter, the prompt for Skift, and I think I stole my headline for that from you, Sarah. Everything everywhere all at once.

All of the lines between travel are starting to blur, and this is just one of the most visible examples to happen recently.

Yeah. While you were over with Expedia, I was in San Francisco. We traded spots.

That’s right.

That’s right.

Yeah, with Airbnb, I got to hear all about it. Brian Chesky is a super interesting guy, and their offices are just beautiful.

Just beautiful. I think today we had a story about their latest move, which is like they’re investing in WeRoad. So I think Dennis has that story today.

It was like a $50 million Series C round. And if you don’t know, WeRoad is a European tour and experience platform. So Airbnb gets a 10 percent stake.

They hired their CEO to run the hotel’s business, their Virginian hotel business. So it’s like, it’s worth keeping an eye on how that pans out. But we’ll plug for Dennis’ story as well.

So we’re full of plugs this segment. Data AI Summit. Yeah, top AI leaders.

Yep.

We got all sorts of stuff for you. So check all of it out. And that is it for this week, everybody.

Thank you so much for tuning in. I hope to see some of you at the event. If you’re there, come over and say hi, introduce yourself.

I love hearing from, Steph and I both do, love hearing from listeners and watchers. So until next week, everybody.

Thanks for having me. See you then. Bye.

Bye.