Travel’s AI Reckoning Has Arrived

0:22Seth Borko: The number of the week is$10 million. That is what Google paid for Spirit Airlines' rest in peace data in bankruptcy court. So that was 100 million emails, 3.4 million payroll records, 40 years of operational data, booking curves, flight behavior, in-flight purchase records. The whole kit and caboodle spirit is gone. But the data is going to train Google's AI. So this is going to be an AI-heavy, AI-fun-filled episode. We have spirit. We have AI investments in the travel industry that are actually paying off. We have gotten to that point in our, I don't know, what would you call it, Seth? Our evolution, I suppose, with AI.

1:11Seth Borko: And we're also going to show you some more fun with charts from the state of travel, too. So what do you think of that? What do you think of spirit data being used by Google to train the AI?

1:24Sarah Kopit: Feels cheap. I feel like they got a good price.

1:27Seth Borko: I was, you know, I was going to say, I'm like, is$10 million a lot for, you know, I'm just thinking. I thought the exact same thing.

1:35Sarah Kopit: It makes me think of, there's another viral story going around outside of the travel industry about AI companies acquiring rare books. And like, did you read this? Cutting the spines off and scanning them?

1:51Seth Borko: yeah yeah that sounds about right but like like rare books like like you can't like they're not widely in circulation you can't find them anywhere else yeah because they're running out of training

2:02Sarah Kopit: data and they basically you know trained the entire their entire billion dollar trillion dollar businesses have basically been trained on reddit so they have the personality of an edgy 16 year old and they've been trying to get um thoughts and words from before you know that existed outside of reddit and in the before times in the before times and they've turned to rare books and you know it doesn't even matter if it's a good book or not it's just how people talked or wrote in the 1950s 20s 30s 1880s 90s yeah and they've been scooping up all these books and the fastest cheapest way to process them apparently is you cut the spine off you run it through a scanner you throw them out um and then you train your model on it and so it's this idea of um ai companies just being so thirsty for this niche proprietary um hard to access data because the models as you alluded to the models are getting really good now and now you need the they're good at general purpose stuff but now they need specialty stuff to train them against and to execute against and that's what i assume this play is for i i feel like i mean i'm kind of a

3:15Seth Borko: claude power user but i feel like um claude doesn't have enough words yeah like he he he it uses the same words over and over again i've had to ban certain words like even with me i'm not

3:30Sarah Kopit: talking about an output i just i never want to hear the word load-bearing again that's a i was just sit with that, Sarah. That's a real load-bearing thought. Let me sit with that for a second.

3:41Seth Borko: Or that question is doing real work, or that statement is doing real work. I'm like, no longer say load-bearing with a hyphen or without. No longer say real work. Math. Like, the math doesn't work. It cannot say math. And now it's just substituted arithmetic for it. I'm like, Claude, that's cheating. So I don't know. Maybe the rare books will give it a... It reminds me I had a friend once who was Danish and he obviously spoke to me in English because I do not speak Danish. And he was a writer and I asked him if he liked writing in Danish or English more. And he's like, oh, English, definitely.

4:22Seth Borko: I said, why? He's like, well, there are more words. Like that was his real answer. There are more words in English than there are in Danish.

4:29Sarah Kopit: There is some writer. There is another multilingual writer who said something very. Yeah. it's so interesting because it's it's a language that we see we're we're kind of i guess writers by profession so this sort of stuff we can around it's a language that matches up like germanic latin romance like for a little bit of french um and so we have a lot of words for a lot of stuff that mean very specific yeah it's great like yeah um well maybe once gemini trains its model on spirits data it will start instead of saying that's low bear and we'll start saying that doesn't really fit in the assigned space uh we'll have to charge you extra for uh yeah i was gonna say this paragraph this paragraph costs extra yeah that's a it's not in the package if you only want

5:14Seth Borko: the bare bones price what what was their thing was it bare bones was it i think that's what it was and it went as a as a spirit flyer i i remember that it was the bare bones deal i think i think

5:26Sarah Kopit: that's what they called it on their website yeah so gemini will start giving you the the big uh the big front paragraph, the big front, big fronts.

5:36Seth Borko: Oh, how dorky is this? Okay. Let's talk about the serious business implications.

5:42Sarah Kopit: Yes. Yes. Go ahead. Sorry.

5:43Seth Borko: So it is interesting. So I'm trying to think how many other deals there have been out there for, you know, AI shops, tech companies, just simply buying data out of bankruptcy to, to train models, like similar to the, you know, the rare books that haven't gotten published, like, is I wonder how a how common that is right now. And we just happen to, you know, be very aware of this one, because we cover spirit so closely. Or if it's something that, like, it's going to be like the hot commodity coming out of the Delaware, coming out of the Delaware bankruptcy courts, it's like, that's going to be the most valuable asset that absolutely has data has always been valuable to be to be clear.

6:28Seth Borko: But now even now

6:29Sarah Kopit: even more so because yeah i think it has to be i feel like what used to it will be the new brand you know there was all these companies these like licensing companies that would buy these brands out of bankruptcy and actually there's an airline story to tell um because pan am pan american had its brand bought out of bankruptcy and you can now go and buy like there are some like hand like bags and some shirts and stuff.

6:55Seth Borko: I've seen those. Yeah. I've seen those. I've seen those Pan Am bags around.

7:00Sarah Kopit: Yeah, exactly. So like that used to be the most valuable thing when an airline or any business would think of it. The most valuable thing was the brand and you would pick up the brand and then, you know, make it into tchotchkes and sell it to boomers interested in nostalgia.

7:15Seth Borko: I was going to say the nostalgia play. Yeah.

7:18Sarah Kopit: The new version of that is probably going to be the data play. Yeah. And that's really, really interesting. I also feel like maybe I'm overstating the case. I have two thoughts on this. The first is I wonder if there are more of these deals happening, but because that are just happening with solvent companies and because they're solvent, they're not being reported. You know what I mean? Like just back in deals. And because you're not getting bought out of a bankruptcy court, no one knows. No one's the wiser. I wonder if that's happened. I suspect it is. I mean, we know that Reddit has signed content agreements.

7:57Sarah Kopit: I believe TripAdvisor wasn't, I don't know. I don't know if TripAdvisor has ever done anything like that. I could see them doing that.

8:03Seth Borko: The thing about travel, though, in both airlines, hospitality, you can imagine as well. I don't know. I was going to say more so than other types of businesses and other companies. there really is so much like ingrained human behavior like behavioral models like contained within that i mean i don't know well maybe it just makes it better and who cares but

8:31Sarah Kopit: i feel like how to be a human i'm bringing it back to my initial thought reaction but i feel like this is like a rare vintage of data this is the rare books of databases because i can figure out if I'm a good data scientist, I can probably figure out pretty quickly. If you're male or female, I can probably figure out pretty quickly by implication, what your age is. I can, I can go look through your Spotify or your YouTube browsing history. And I'll pretty quickly figure out based on the kind of stuff you're listening to the kind of stuff you're watching. Um, whether you are, you know, what your, your gender, your age, your income, it's not that hard to guess that based on zip codes probably even guess your politics but i don't know if you like a window or an ilc do i that's yeah privileged and confidential information and i certainly don't understand the uh operating fuel efficiency of an a321 neo uh and i certainly don't understand the route network of you know passenger travel flows between fort lauderdale and minneapolis this is the sort of stuff that we talk about that is highly both on the consumer side and on the operational side highly um wall specialized highly specialized niche data and we live in a world where that is increasingly the valuable stuff because you know new york city zip code a new york city zip code uh you know 40 you know 35 year old man in new york city i can pretty much guess like white man, New York City, 35, New York City zip code.

10:11Sarah Kopit: I can guess a lot of information about you. I can't, I can't. The flights thing doesn't exist in that way. It doesn't correlate with other publicly available data sets in such a neat way.

10:23Seth Borko: It's so interesting. I, of course, immediately went to politics with that. I wonder if, while you're talking, while I'm talking here, Seth, I'm going to ask you to ask Claude, what political leaning is the window versus aisle versus middle seat what political lean in yeah yeah so like if you're if you are if you are a window seat person do you vote republican or democrat i want to know now now i want to know now i have to know and you know i guess in a few months we should ask gemini but uh yeah uh well what what what is your let's think about this

11:05Sarah Kopit: what's your take is is a window seat a republican seat or a democratic seat it's a little more closed off but you get to look at like you get you know a little you sleep privacy i'm gonna go

11:19Seth Borko: yeah i don't know this is my instinct i'm gonna go window is democratic and aisle and middle are

11:25Sarah Kopit: republican middle is libertarian i'm confident of that one that's that's definitely

11:33Seth Borko: are you taking both armrests if you're not if you're a libertarian are those yours the middle

11:37Sarah Kopit: seat is entitled the middle yeah yeah absolutely absolutely absolutely you are in the way it works yeah i think we can let's go with window is is democratic you like to look out the window at the clouds you're a dreamer window is middle seat is chaotic libertarian yeah uh aisle seat is Republican. It's more no-nonsense.

11:58Seth Borko: A lot of freedom.

12:00Sarah Kopit: You can get up.

12:01Seth Borko: Space.

12:02Sarah Kopit: Yeah, yeah, yeah. It's more space. Yeah. And the Green Party is on a train. It's on a train.

12:13Seth Borko: You're absolutely right. Also, you brought up Spotify. Did you do the Spotify thing at the end of the year where it told you how old you are based on your musical preferences?

12:25Sarah Kopit: i i'm embarrassed to admit i'm embarrassed to admit this but i don't i don't have a spotify account i i use i use apple music what does that say that's see that's that's an interesting data

12:34Seth Borko: set yeah you know and i'm a i'm a i'm an apple user like through and through and i don't even use apple music i was also said to be 78 years old based on my musical tastes that's what it gave 78 yeah so i think that means i have good good taste in music actually but but i think this is

12:54Sarah Kopit: the bigger the bigger thing these the this data is so valuable it's critical operational data it's going to help with things like i don't know like we're already seeing ai tools that are specific to the domain specific to the use case and even specific to aviation like predicting delays predicting routes predicting what routes to fly this is like critical stuff and it's valuable stuff i wonder if even gemini just bought it so that no one else could buy it there's a part of me that that thinks you just buy it but no one else can get it google has a history yeah that's

13:29Seth Borko: a that's a tried and true business strategy right there for with a lot of other applications so

13:34Sarah Kopit: sure maybe google by the way has has a history of acquiring these sort of things i mean they acquired um a much beloved flight plan in matrix tool i'm blanking on the name of it right now but it was like beloved by travel agents that have all of this data and they really bought it for the data to fuel their google flights so it was like this um it's called matrix ita i think matrix ITA.

14:02Seth Borko: Have you heard of that? Does it ring a bell? It does not. It's okay, though. You know, it's interesting. For all of you sci-fi nerds out there like me, one of the things that always strikes me whenever I'm reading, especially an older sci-fi novel, is just how correct they are. Like, if any of you go back and read 2001 A Space Odyssey, like, the iPad is in there. um it's it's like kind of spooky um but I just started watching this show uh which is kind of old now it's like six years old called devs have you ever heard of that I think it was an fx show and I I was surprised I'm no spoilers nobody right right to me with spoilers I'm only three episodes in but I had never heard of it which is surprising for me because I'm a just like rabid watcher of television, but also I really like science fiction.

14:59Seth Borko: But it debuted on March 5th, 2020. We were all a little bit preoccupied when it debuted. But anyway, it's very similar to Foundation, where it's talking about prediction models. And of course, that is what AI is, or predictive models, you know using math to basically see the future um and now that i kind of it's interesting because now that we all do use ai in our everyday lives not only in our everyday lives but in business it's like you know i had read foundation a long time ago and now watching this and now that that like mathematical prediction or predictive models using math to do that it's all becoming very real and i i kind of i understand it now more than i i did like when i read it the first time it just sounded more like magic and now it doesn't seem like magic anymore at all um and so yeah like when you said predicting you know with with the with the spirit data you know using it to actually do like real hardcore predictive work like it seems like kind of like a duh thing at this point well yeah i mean so this the company i'm thinking was called

16:23Sarah Kopit: ita it was ita software it was the leading independent airfare pricing and shopping engine it powered flight search for kayak or its travel aussie trip advisor and others so it was a real hardcore b2b pricing tool but it was like this like hack that in the travel agents and in the no travelers would use that there was a way to access it commercially ready and it was like It was like old school DOS kind of stuff. You couldn't really click around. You had to use commands and keywords. You had to know your airport code. So people really liked it. Google bought it. And ultimately, I believe, I can't necessarily prove it.

16:58Sarah Kopit: I think it's been proved, but I can't say for sure. But it became one data point, one input in the infrastructure build of what became Google Flights and their Google Flight Search, which is a really powerful tool. And so I wonder if that is what this is now. I mean, I don't wonder. I'm sure this is what this is now. Not sure is a difficult word to use, but I feel confident that this is probably the direction they're going is buying up these data providers to improve their flight search, to improve their flight prediction capabilities, to understand the kind of flying consumers, the kind of things and products they buy better.

17:32are no doubt as part of their push to take Google flights and their really powerful Google

17:39Sarah Kopit: travel search businesses like hotels and flights and modernize that for the Gemini AI world. And that seems to be where we're going. And I suspect 10 million will be cheap. I suspect that the story is when we write, when me and you write that final story, the final, you know, in the final estimation, that the ROI on that$10 million deal will be positive.

18:12Seth Borko: Yeah. Yeah. So it's interesting. We did another story this week. Adriana Lee wrote it also about AI and what is now. we've moved really fast through the stages of you know novelty application now we are truly in like operator like we are in usefulness stage i don't know if we're in profit profitability stage quite you know quite so much yet but we are in usefulness stage like there are a lot of um companies out there that are they're not just experimenting anymore they're not just you know using AI as a hope of something that will be good for them in the future. And I don't know, there was a lot of really interesting stuff that Adriana dug up out of earnings reports and other statements from companies.

19:13Seth Borko: What struck you about her reporting? What was your favorite like positive, actually useful bit of AI that is being used right now in the travel industry.

19:24Sarah Kopit: So I loved Adriana's story. I don't know if we can pull it up with the headline and the tear sheet and everything. Travel has started weighing AI's worth. She pulled out, I think, from five different travel tech booking platform distribution companies, Booking, Holdings, Airbnb, Expedia, Sabre, and Amadeus, and really did a great job of piecing together what they are actually saying and seeing. Can I share two that really stood out to me, Sarah? We can maybe discuss them. Do you have one that you want to talk about, too? Because I'll tell you what I find interesting.

20:01Seth Borko: I thought that Brian Chesky's like...

20:04Sarah Kopit: I was going to say that one. Let's talk about that one first.

20:06Seth Borko: Yeah, okay. Okay, you go first.

20:09Sarah Kopit: Well, Brian Chesky specifically called out... Well, it's so interesting because AI has, you're exactly right, has moved from this dog and pony show to this idea of, actually, it's really expensive. Actually, it's really expensive. And it's no longer, we're no longer doing it because when we just say we do AI, we get a huge stock boost. We now need to prove that we are using it and actually finding use cases and seeing ROI. And on their most recent earnings call, what did they say? They said nearly 45, they cited a customer service use case. This was something that we said, I mean, two or three years ago in our very first AI report in 2023, the very first use cases for AI would be, well, code in first, copilot code in first, and secondarily, customer service and chatbots.

21:04Sarah Kopit: I am a little disappointed that we're still in chatbot use cases. And yet, customer service is a big deal and a huge cost service and a huge differentiator. So what did Brian say? He gave some stats. Nearly 45 % of issues that start with their assistants are now resolved without a human. And I think he gave a stat. I don't know if it's in this article, but I think the stat was something like their cost per customer service resolution had gone down by something like 16 % to service a customer service. complaint. So they were really proving out the ROI and the value of their AI implementation. That was what I found quite interesting.

21:48Sarah Kopit: But it was still on the cost side of things. But what stood out to you? I mean, what stood out to me was, I mean, it was what he said, but it was also

21:57Seth Borko: what we mentioned about the stock price. I mean, the stock price jumped 17 % after the 2Q earnings. I mean, I know there are other things going on there. I mean, Airbnb had like a pretty knockout.

22:09Sarah Kopit: Blockbuster. Yeah.

22:11Seth Borko: But, you know, it's just, it's one of those things. Like I remember Brian Chesky was the first person that I really talked about AI with. I interviewed him. It was three years ago. And, you know, when he said, I mean, I remember him saying and me at the time thinking, oh, that's cool. You know, how just how impactful it was going to be for customer service because. You know, of the. Let's just say thousand predicaments that Airbnb's customers are in. And he's like, you would think there are millions, but there aren't there are like, you know, it's a far. smaller number. Like people generally have the same problems.

23:00Seth Borko: You think that your problem is special. Generally, it's not like it's it's not so unusual. But, you know, a customer service rep can't search through those like thousand use cases. But he's like, hey, I can. And it can do it instantly. And it can give our human agent the, you know, top three successful remedies that have worked in the past. And so, you know, that made, for me, like that's not a chatbot. That is a completely different thing. That is assisting the human. Because I still think chatbots are bad at this. I still think they're bad at this.

23:42Sarah Kopit: By the way, I was going to say, and I love Brian Chesky and he's a great operator, but saying, oh, you think your customer service is special, but actually it really falls to my book. It's such an online booking platform thing to say. It's such a tech company. It's such a tech response to a customer service thing as opposed to what a luxury hotel group. I mean, they're not.

24:01Seth Borko: I mean, as opposed to like how a thousand times the GM of the four seasons.

24:07Sarah Kopit: Yeah, you're right. He gets the same five. But anyways, points aside.

24:10Seth Borko: Yeah.

24:11Sarah Kopit: Yeah. Here's the data. The company said its AI assistant helped drive customer support costs per book and down about 16 % year on year in the second quarter and nearly 45 % of issues. that start with the assistant are resolved without a human. So he is seeing significant. I mean, he's obviously right. I mean, the tech, as a human, as a customer, I want the luxury GM approach to all my problems, but he's obviously right that at the scale of an online platform and a tech platform, they do kind of boil down to the same general things and you can help humans answer those better.

24:45Seth Borko: You know, it is interesting for somebody like me, a traveler like me, who I'd say probably you too, Seth, who, you know, you book through your OTA, you book an Airbnb, you know, there's a lot of mix between online and real life in our own personal travel. But it is always interesting, like every once in a while when we get the opportunity to stay at like a very like five-star or above hotel and just how different it is when something does go wrong. You know, like, I forgot my charger. Oh, let me go out and buy you one at the store.

25:23Sarah Kopit: Well, I think that's why the startup advice in the beginning of the startup is do things that don't scale. Because luxury does things that don't scale and impresses customers. And I would also say that's why luxury has full service hotels have much worse margins than Airbnb's margins.

25:43Seth Borko: Absolutely. Without question.

25:46Sarah Kopit: Can I point out another one that I thought was interesting from this article, which is, again, is a great roundup. So we're talking about chatbots, about customer support, about ROI and, you know, companies seeing payoffs from AI in a cost bucket. they're able to do the same or more with less, which is, of course, an ROI of its own. But there's another form of ROI that comes from more revenue, not just lower costs. And here it's interesting, and we're kind of going to mix messages here. This is the final frontier. Are you not just reducing cost to AI, but driving more revenue upside through AI?

26:30Sarah Kopit: And we're getting somewhat mixed messages. So Expedia Group, Ariane Gorin, the CEO, she said, she gave a quote, and she was talking about their AI adoption. She said, more broadly, our agentic technology stack is allowing us to design and ship products faster, ultimately unlocking new capabilities for both travelers and partners. So she's talking about not just saving money, but building new things. Expedia then gave a quote or a stat. I'm paraphrasing. This is not a direct quote. But Expedia said along the lines of AI-powered recommendations and AI, just generally AI-infused products and experiences generally yield 60 % more information about traveler intent.

27:21So what she was saying is like, it goes back to this idea about buying Spirit Airlines' data.

27:28Sarah Kopit: she's saying is that we are, when we, when people interact with AI, the keyword searches are longer, they spend more time there. This is not exactly, but I'm assuming, you know, you ask these long queries, you have these full conversations, the AI is able to answer them, synthesize them, because it is a more conversational, more human-like search, but you're able to capture the data from that more human-like conversation in ways that you previously, a structured form, a structured data only had a handful of dropdowns in a form, your customer clicked off from this, URL at this time and they spent this much time on your page.

28:01Sarah Kopit: That's data, but that's not as useful as the data of like, oh, this person had a whole conversation and this person had the, you know, so she's saying we are able to get more intent data out of our AI infused products, which presumably they will then go on to monetize down the line. Presumably Google will use the spirit data to do something similar and is already doing something similar with its own search data. That was an interesting tidbit to me.

28:29Seth Borko: I don't think I don't have any worry about companies not being able to make money. Surely that surely like that's coming and is probably coming sooner than we think.

28:40Sarah Kopit: I'm worried about it. AI is expensive. I am very worried about it. This this is why we're getting these stats, because Wall Street is pushing back on the CFOs, telling them you are spending a ridiculous amount of money on AI to do effectively the same thing you've already been doing. you're doing more for less.

29:00Seth Borko: It's short-sighted, though. I don't know. I think it's short-sighted. I mean, it's typical Wall Street, meaning they want their trade right now. But, I mean, it's classic R &D play to me. You know, this is what OTAs and other very online companies have to do right now in order to, like, grow into the future. It is interesting, though, about what is not working yet. And Expedia talked about, you know, they talked about that, like, also with the natural language search and just about how that is not driving the conversions right now. It's, I mean, I wonder if people are going in there and just, like, having a chat.

29:43Seth Borko: Well, I'm... And it's costing them money.

29:45Sarah Kopit: And so to round out the big three, Booking Holdings said the same thing. They said that room nights, that AI models, large language models, remain below 1 % of room nights. So less than 1%. I actually don't know how they're measuring. There's some sort of conversion metric they must be measuring. But the quote was something like AI is less than 1 % of room nights. So I don't know if there's probably some sort of conversion factor there in terms of actual large language models. We search for stuff, and then you come through a Google ad or you type in Booking.com, and it's technically an AI-sponsored booking, but the actual technical attribution attributes it to a direct site or to TripAdvisor or to Google or whatever.

30:46Sarah Kopit: But what they are saying is that in their actual attribution data, they see an incredibly small, incredibly small, 1 % or less share of actual bookends where they can confidently attribute it to an AI. That was a little shocking to me. And that speaks to your point that people do seem to be coming in and having a chat. People seem to be using this. And I had a contact of mine in the industry who's working on a Gentic AI tool in TravelTech. messaged me and we were having a sidebar conversation about this said how do you square the fact that our survey data shows that two-thirds of travelers more are using the i-a-plan trip and booking.com says they get less than one percent of room nights that is a huge i mean drive a truck through that gap yeah well i don't know make it it does make sense to me though because

31:36Seth Borko: it's just we as humans have not yet it is too soon to be it is too soon for us to be using it to like spend our money unless you're like you know way ahead of the curve like you know but it's not too soon to be researching like I think even my parents use it to research now like everybody is using it now like we have come we like I remember we've even had this podcast where we you know used to say things like well but you know my parents aren't doing it well now they are oh they definitely are yeah they definitely know here we are just like just like my mom also knows how to text now has for a while every day so it was so funny my father-in-law we're i was in san

32:21Sarah Kopit: francisco with him for a trip and um he used ai to plan his whole san francisco itinerary but but he printed it out and he came in with like like nav class yeah came with a printed handout of an AI itinerary. And he said, oh, I've got my itinerary right here. And I thought he was going to take out his phone. He said, oh, I planned my trip in chat to you. I got a piece of paper. And it was a printout of his itinerary. So some of the old ways never die, even though we are modernizers.

32:52Seth Borko: So we were talking about itineraries and about how we, you know, basically we've kind of come full circle and about how on this very podcast, you know, we haven't been doing this podcast for too long. We used to talk a lot about how, well, some people are using AI, like the youth is using, the youths are using AI, but like our parents are not. But like now my parents very much are. Like everybody is for research, but I don't know if they're going to give them any money yet. You know, like that's, like you're going to go, you're going to do your research in your AI tool. you're going to go off platform in order to spend your money that just kind of still makes sense to

33:39Sarah Kopit: me or do it you dismiss the wall street analyst as short-sighted but if no one's going to spend money with ai the wall street analysts are right if no one's spending money with your ai and everyone is investing in ai then someone is not going to make someone is losing money in fact a

33:54Seth Borko: lot of people just losing money i just think we're not there yet seth i think it's a timing thing that's my and it's happening it's happening faster i mean i don't even know if it's fast sooner than we think or faster than we think i think it just simply is happening fast we're just not there in august of 2026 well and that's why this great piece from adriana gives us the updates

34:17Sarah Kopit: and frankly that pressure is why companies are giving us these updates we we push as as journalists and researchers we push every single day for numbers and stats behind um behind ai usage and sometimes we get them but nothing nothing lights a fire quite like quite like your investor base saying what did you just spend billions of dollars on and what are we getting how much

34:43Seth Borko: exactly right do we have any charts about this stuff has anybody made a chart about about such

34:50Sarah Kopit: things about such things we got a lot of charts i mean i will share and i've shared this one before one of my one of my favorite um charts which is about ai comfort and so this is in our um that's what we were just talking about really to travel commerce section and um this was about net comfort score i think i've shared this one before but we might as well share it again we asked a bunch of travelers to talk about um a bunch of different channels and a bunch of different ways to purchase travel and we said how comfortable are you on a scale of one through five are you do you love this channel are you super comfortable with it are you uh not so not so comfortable are you somewhat comfortable are you neutral do you not care or maybe are you very uncomfortable or somewhat uncomfortable and we created uh we called it a net i'm using the word comfort off.

35:45Sarah Kopit: But we created a net comfort score where we took the percentage of those who said four or five, I love this channel, subtracted the share who said one or two, I hate this channel, and got a net score. We did this for a bunch of different channels, direct package tours, old school travel agents, online travel agents, and loyalty portals like Chase American Express. We also asked about AI tools. The only one of those channels with a negative net comfort score, where more people were uncomfortable with the tool than were comfortable using the tool to book, was AI. Every other channel had a positive one.

36:25Sarah Kopit: It won't surprise you to learn that direct was the most, was what people trusted the most. and so there is a big gap one of the pushbacks i get when i've shared data like this publicly people say actually say exactly what you said sarah which is but that's today what will it look like tomorrow and that is that is a big question what will look like tomorrow but i think that is why one of the the push points is going to be how will ai integrate into the platforms that we are more comfortable about with. One of the great debates happening in this sector today is who will own the bookend in an agentic AI world?

37:06Sarah Kopit: Will Google or Claude or ChatGPT own the bookend? Or will, and in doing so, will they disintermediate Expedia, Bookend, and others?

37:18Seth Borko: Yeah, the OTAs. Yeah.

37:19Sarah Kopit: And one of my answers to that is that you need a trust layer. we know that people don't really trust ai so if you can sort of um like sort of like like our our our guest host milo the golden doodle when you need to when when milo needs his medicine if we can sneak it in with a little bit of cheese you need you need the the cheese trust to swallow your ai medicine with so to speak in a very tortured metaphor that is i was just gonna say

37:48Seth Borko: that is a fantastic metaphor. Everyone understands this. At least everyone with a dog.

37:54Sarah Kopit: And so that's why I am more bullish than many on the future of online travel agencies. When people say, I don't understand how an OTA can survive in an AI world. Why do I even need an intermediary? I say, well, there are decades of trust and relationships. And by the way, trust with consumers and trust with suppliers. There are decades of relationships. embedded in these brands and these platforms, and they are working to integrate AI into them. So it's like, can the trusted platform build AI before the AI platform builds trust, if that makes any sense? It's a bit of a race. And so that's one of my favorite charts.

38:35Sarah Kopit: And I think it speaks to this dynamic. Adriano's article also speaks it, and we highlighted the big three travel agencies, but Adriano smartly also included the global distribution systems, GDSs, which is a real nerdy, real distribution, real proper B2B travel business without consumer brands. And they are working incredibly hard on launching AI tools and MCPs. I think Adriana also cites Sabre as making quite a lot of progress on their connectivity. They've acquired some AI tools. The point is that Sabre, you know, they've been acquiring companies and they've been growing and building it. And so it is sort of a race where like, can these global distribution systems, which are the plumbing behind the ecosystem, can they integrate AI?

39:29Sarah Kopit: You know, it's just it's just really interesting. It's a really interesting dynamic. and we are starting to see the tidbits of ROI. Although I have to say, I don't really think we're there yet.

39:43Seth Borko: Shall we do some winners and losers this week?

39:45Sarah Kopit: I would love to do some winners and losers this week. Do you have any losers or winners that you'd like to highlight? Because I have a winner. Let's do your winner first. Well, we can't end on losers. We learned that lesson from that previous, it's too much of a bummer. I would like to call out a winner I'd like to shout out a listener Anna who is a client of ours on Skiff Research who told me she is a regular listener of the podcast and get this Sarah I said do you like the podcast? Is it useful to you? And she said well I usually listen to it while I'm cooking and all of my recipes that I've cooked while listening to you have turned out really good so the podcast must be working it i'll take it man yeah i'll take it too i made my day so thank you anna winner of the week for being a listener and any of our fans who are out there regular listeners i i see a lot of you sometimes in the comments and those in conversations i think we saw a friend of of the skiff podcast friend of skiff sarah dandashi and she commented we did the whole thing about um influencers and and that and iron boards and this one influencer samantha brown who i i only just discovered and sarah said she knows samantha brown and talks to her about iron boards so another my my whole winner section is just audience shout outs if you said you like the podcast if i saw it or i talked to you about it i just want to say thank you for for being such

41:13Seth Borko: a great community i tried to do this podcast on an ironing board just before we started here there was not enough room for the laptop and the mic people who use the ironing boards what are we doing about this?

41:25Sarah Kopit: How do you get a mic into the iron board setup? That's the question.

41:28Seth Borko: Yeah, exactly. How do you do that? So listeners, please do let me know. Let's just end on winners this week. Let's not do losers. I think let's say thank you to our listeners. Thank you to our viewers out there. Happy cooking.

41:42Sarah Kopit: Hope your recipes come out great as you're listening to us or wherever you are listening. Your run, your morning run or your evening run or your cooking, whatever you're doing as your podcast and fold and close. I hope it goes well.

41:54Seth Borko: Excellent. All right. See everybody next week. Thank you.

41:57Sarah Kopit: Bye, everyone.