The AI use cases already changing hospitality, with Gilad Berenstein of Brook Bay Capital

[00:00:00] Gilad Berenstein: AI is coming for everything except for those things that we choose to keep in the human domain. Look around your peers, figure out what you think your peers will value being done by a human, and become good at that.

[00:00:23] Matt Welle: Hi, everyone. Welcome back to another Matt Talks Hospitality. One of the things that shifted in the last few years is obviously AI, where previously, every hotel was battling for the attention of Google and making sure that you're visible. Today, that's a very different picture, and guests are going into ChatGPT or any of these LLMs to ask questions about their trips. And it's interesting, but it's also nerve-wracking for hoteliers to be in that market today. So, I'm joined today by Gilad Berenstein. He's the founder of UTrip, which was the first AI-travel company that helped over 200 million travelers plan their trips before they sold that business. And then he moved into Brook Bay Capital, where he now runs this business where he invests in AI and hospitality, which is like the crossroads of everything that I love. So, I'm very excited to welcome him to Matt Talks Hospitality. Thanks for joining.

[00:01:07] Gilad Berenstein: Yeah. Thanks for having me.

[00:01:08] Matt Welle: So, did you grow up always wanting to be in hospitality, or did you always want to work with AI?

[00:01:13] Gilad Berenstein: Yeah. No. AI. So, that's the answer. I mean, I feel lucky to have found hospitality, but no, my dad was an AI architect in the 90s. I grew up in Israel, and I was a lucky kid in the back of the living room. Got to hear my dad and his friends talk about the future.

[00:01:24] Matt Welle: So good. So, what is it about AI that gets you excited?

[00:01:27] Gilad Berenstein: I mean, AI is changing our world in a way that I think is more profound than any technology we humans have ever created. And, you know, we can keep the AI philosophy for another time, whether it's all good or all bad or, realistically, something in the middle. But the amount of change we're seeing in this short period of time, I find to be extremely exciting, and it's an opportunity for us to create these win-win-win outcomes, what I'm always talking about in my speeches, in my essays, where it's no longer the seller versus the buyer. It's no longer the traveler versus the traveler. We can all win in this ecosystem. So, that's what I'm so excited about. And the fact that it can unleash so much creativity, so much business, it's just an exciting time.

[00:02:03] Matt Welle: So, a hotel that is not very technical, and I talked about, like, being visible on Google back in the day. And today, hotels are just lost because people are like, well, SEO is ruined and, like, getting visibility is very different. But I'm like, yep. Yep. But what do we do? Like, what do hoteliers do that are completely lost in this AI landscape today?

[00:02:20] Gilad Berenstein: Yeah. I mean, first and foremost, they begin to build internal expertise themselves, and the easy way to do that is just by committing to using AI regularly. So, as I say all the time, when you show up at your computer in the morning, you might open an email, a CRM, a Salesforce, whatever. You should also be opening your AI, and you should have it there with you all day because that's where you begin to identify small opportunities that become medium and eventually large opportunities. That is number one, personally. And until you build personal competence, there's no real reason to talk about business strategy, whatever. The second piece for those who already have some level of personal competence is to begin to create the new type of content that is required for you to be discovered in this AI era.

[00:03:00] Matt Welle: Love that. And specifically, the one thing that I've started doing is recording literally everything. In the real world and in the online world, I just press record on a note taker at all times because now my AI gets to know how I speak, what questions I ask, what are the things I'm interested in. And I connect my AI to consume all of this constantly, and it's really, really good. But also in physical meetings, I now, you know, I think Granola just launched an app for smartphones. So, you just use your Apple Watch to record physical meetings because that's really valuable context as well. Right?

[00:03:29] Gilad Berenstein: Yeah. Absolutely. So, as we said before we started here, I'm about to head to Vegas for Virtuoso Week. There'll be 100,000 individual appointments that week in Las Vegas, and many of those people have devices, plots or other kinds of devices they put on their table because you cannot remember the 50 meetings you're gonna have this week. But if you build that database, you can now build an external brain that has all this insight that you can now use for the rest of the year.

[00:03:51] Matt Welle: So, we met a couple of months ago at Mews Unfold. You were on stage with me, and I enjoyed that conversation. But AI moved so fast. Is there anything that's come on your path since then that you're like, this is the hardest thing right now that you need to know about?

[00:04:01] Gilad Berenstein: Yeah. Absolutely. Well, ultimately, AI swarms. So, I spoke about this on stage in Amsterdam together, but you're beginning to actually see people assembling these swarms. So, real quick, a swarm is just a group of agents that you construct that work together to accomplish overall tasks. Each of these is kind of dumb and simple on its own, but when you put them together, they can do really incredible things. So, we began to hypothesize a couple of years ago about these swarms that might include, you know, a justice support with housekeeping, with operations, with pricing, with guest service, with communication, with whatever, but they were all disparate. They didn't talk to each other, which is so many of the challenges people face with AI, like, why is it the wrong answer? Why is it so slow? All these kinds of things. So, we're beginning to see actual hotels beginning to put together these swarms, starting at the ground level. I'm not talking about this ultimate, most complicated swarm. But can you bring together a few agents that are maybe doing some stuff with commissions, some stuff with customer service, some stuff with operations, and together make them so much smarter because they can talk to each other? So, a few months ago, we were hypothesizing, and today we're beginning to see the very early examples of people doing this for real.

[00:05:02] Gilad Berenstein: So, what does that mean? So, like, do I just deploy a bunch of AI apps, and then you've just plugged them into each other? Or…

[00:05:08] Gilad Berenstein: So, yeah. So, one way to do that is to have a bunch of AI apps. So, for people who are using Claude Cowork or other things like this, you have seen examples. This is not exactly the agentic example. I would think it's still really illustrative of how you can now connect your Google Calendar, your bank, your Slack, your whatever into this thing, and it can now all be in one central brain. That's an example. Now, of course, here, you're connecting services. It's not exactly the same thing, but I think the analogy still fits. The other place you see this is people beginning to train their own bots. So, one easy example for any hotel is a concierge bot. You take all the emails and questions you got from your guests about what to do in Rome or Venice or whatever city your hotel is in, and you take all the emails your concierge has manually written in response to your customers; that's your perfect training data, and now you create an agent that's trained on answering questions of what to do in Rome, for example. And then, as you train another agent that's really good at predicting no-shows, another agent that's really good at chasing commission, another agent that's really good at whatever, as you begin to connect these together, they'll be able to talk to each other, which will bring significantly more efficiency and significantly more knowledge to your ecosystem.

[00:06:07] Matt Welle: Like, I think the example of letting an agent consume your entire inbox is so smart because I think a lot of hoteliers don't think about it. You have all the data. It exists. They just need to feed it into this, like, private kind of network of AI that you have.

[00:06:20] Gilad Berenstein: Yeah. Absolutely. It's a gold mine, and it's sitting right there waiting to be utilized. And at the end of the day, I'm not suggesting that you replace your human concierge. I'm suggesting that you superpower the human concierge. So, now they spend more time on human connection, speaking to your guests, really engaging with the community as opposed to the computer in the back doing a bunch of research.

[00:06:39] Matt Welle: So, you very briefly touched on Virtuoso. You're a board member of this organization. It's a luxury travel agents' organization, I think. Is that a good description of it?

[00:06:47] Gilad Berenstein: Yeah. So, it's a consortium, and a consortium brings together a few different ecosystems. This is luxury travel agencies, with about a thousand virtuous agencies around the world with supply partners. These are air, hotel, cruise, tours, sort of, the best luxury providers in the world, with all sorts of other providers: technology providers, insurance providers, transportation - all the pieces together. And you bring them together both physically in one place this week, but throughout the year into a network so they can create truly magical experiences for the guests. And if you think about this, you know, one thing that Virtuoso did is it created this connectivity level before we had all these AI agents and APIs were able to connect it to ensure these experiences are really amazing. And I always say, like, when you go on a big trip, luxury or otherwise, there are so many providers. There are probably multiple transportation providers, lodging providers, experience providers, dining providers, and, again, lots of other people supporting you. How do you get them to play from the same songbook so the experience is amazing for the customer? That's where Virtuoso began, kind of sharing data and knowledge across the ecosystem. And it's, like, 40 years old, maybe a little older now, and it's really incredible what the organization has built.

[00:07:50] Matt Welle: So, I will now go into ChatGPT, and I have ChatGPT draw up my perfect luxury trip. How does a Virtuoso agent leverage ChatGPT differently than I would as a traveler?

[00:08:00] Gilad Berenstein: Yeah. Absolutely. So, first and foremost, if you begin planning on your own, which is a very common use case, you can actually invite your adviser, the concierge, the hotel you like, whatever, into that plan. You can actually now begin to plan it together. Now, a lot of AI agents have their own AI tools that they're bringing into it. So, there's a company called Trip Boutique, for example, I'm engaged with, which creates external brains, actually very similar to what we spoke about earlier, of recording all your emails or your calls or whatever. They take all the itineraries and all the recommendations you've ever made for guests within your agency and create internal wisdom, so when a new customer comes in, they want kids in Rome or whatever. You can say, last month, we sent these people, blah blah blah blah. So, effectively, you have your own AI. Today, it's stuck inside ChatGPT and GenAI. In the future, it'll be an agent. It'll be something you interact with through your phone, your computer, or something else. And your agent, in the SwarmMaker structure we just talked about, can now go pow-wow, talk to, communicate with the agent from your advisor. And once you've booked that reservation, let's say, the beautiful Four Seasons in Lanai, all of a sudden, your agents can now go talk to the concierge agent from the Four Seasons Lanai, And together, they can say, what's the best romantic dinner for Gilad and his wife to go to for their anniversary based on all the knowledge about the restaurants I book when I'm home, when I travel across my life, my adviser's brand and perspective with, of course, the knowledge of the concierge from the Four Seasons, but that's not it. If I were outdoors, they could bring in a weather prediction bot. If I care about sustainability, that could be a part of it. If there's transportation, it could be a transportation prediction bot. And together, they create an experience that's really magical for me, elevating all of their knowledge together.

[00:09:30] Matt Welle: So, I hear you, but does this actually exist, or is this just a dream that we would like to think this will go to?

[00:09:36] Gilad Berenstein: Yeah. I mean, I would put it closer than a dream, but no. What I'm describing right now is an end state that does not really exist yet. The time frame we're talking about here is at most a couple of years, and probably months. You know? Your first question was very, one of your first question was very wise, like, what's changed in the last two months? Well, a lot has changed in the AI world in the last two months, and we have to assume a lot will change over the next two months as well.

[00:09:57] Matt Welle: I'm really interested in UTrip. Like, this platform that you built, which is all around personalization, was a couple of years ago. So, I'm imagining that AI wasn't at the pace it is at today, but I'd love to just hear that story.

[00:10:07] Gilad Berenstein: Yeah. Absolutely. So, I already shared I knew a bit about AI from my dad being a kid in the back of the living room listening to him. And fast-forwarding a whole bunch, when I graduated from grad school, I'd done a couple of startups in undergrad before that, so UTrip was not my first startup. But then after finishing grad school, my buddy, Matt Rob, and I took a trip to Europe, and we were those basic Americans who went to London, Amsterdam, Paris, and Barcelona, and we had a great time, for the record. But I constantly felt like I had, you know, misinformation, holes in my itinerary, just all sorts of little things that every traveler has faced and experienced in their lives. And I came back, and I began to think about, you know, how would I plan my trip if I were to do this again? And I realized that there wasn't a tool out there that would take my preferences and the huge amount of information that's just in the world and mash those together for my benefit. So, UTrip was born in 2011 to become TripAdvisor 2.0. We thought then, as I still think now, it's broken, lacks personalization, and all those kinds of things. And our goal was to effectively become an AI-powered, highly personalized TripAdvisor. And over a couple of years, we learned that we are pretty good at data and pretty good at product, and we had great usage, we had great NPS, and great time on page, and we had no bookings. And because of that, we effectively had no revenue. And I always remind people: 2012, 2013 was a very different era on the Internet. Like, we weren't putting our credit cards into a whole bunch of random websites. Like, there was still a lot of anxiety about that. So, when we really studied user behavior, we noticed that lots of people came and planned on UTrip and then went to Booking.com or Expedia.com or Marriott.com or whatever it was to go book that actual hotel. So, we did end up pivoting from a B2C business to a B2B2C business, where effectively we kept the same core product, the same core value proposition, etc, but instead of delivering it directly through UTrip.com, we delivered it through all sorts of partners we had around the world. And this was an incredible era because, you know, we started this licensing thing in 2014, right as, like, SaaS was being born and right as licensing became kind of a more traditional business model around the world. And we were very lucky that we began to work with DMOs and tourism ministries and quickly added hotels and airlines and cruise lines, etc. And we're really proud to work with over 150 brands.

[00:12:08] Matt Welle: If you were to rebuild that business today, as in from an AI native, like, in the time of LLMs, would you build something different?

[00:12:15] Gilad Berenstein: I would build a different, but not from the core AI. I mean, obviously, the core AI would be different because we had to build our own models in-house, which took years. And I was saying my investors were much more patient with me than I am with my companies as an investor now. So, obviously, there's a huge advantage now where you can get things off the shelf that are quite incredible, and you can fine-tune them, whatever. Really, what I would think about now is the whole thing I spoke about on stage is on dynamic experiences. And for a long time, we've had dynamic packages, and I like to think that UTrip had the best version of dynamic packages, but it was still a mad lid. It was still a pre-existing story with a number of holes in the story, and then the algorithm filled in the holes with a small number of pre-existing options, whatever. Today, you have the ability to start everything from scratch. Every trip, every user begins as a story, as a narrative, and then you have the ability to use AI to go identify the right hard product, to drop things along the way, etc., etc. So, I would begin by thinking about: how do you build for an end of one? How is every trip really built just for me? And by doing so, we achieve the win-win outcome that I'm always looking for - better trips for travelers, increased margin and share of wallet for the travel sellers, which is a double whammy financially, with the third benefit of reducing the negative impacts of tourism and increasing the positive impacts of tourism. I often like to say, for every city that has too many tourists, for every community that has too many people coming in, there are 10 or 50 that are begging for more tourists to come and spend their dollars in those cities. So, by telling new stories, we can always say, here are other places you haven't been to that are gonna offer that same romance, the same sunset, the same thing that you're seeking in a place that is gonna have a really positive impact on both the environment and the people who live there.

[00:13:50] Gilad Berenstein: Would you commercialize it differently today? Is there a commercial model where people pay for that end-to-end kind of booking experience? Or can you actually even do end-to-end bookings on, like, AI tools today?

[00:14:01] Gilad Berenstein: Yeah. It's a really hard question. Right? So, on one hand, you wanna say, like, the best business model would be some sort of subscription business model to the traveler or to the consumer. Right? I wrote an essay a while ago, but the problem we have is that as long as your business model is commissioned, you will never be perfectly aligned with the traveler. At the end of the day, Expedia believes that the best hotels in the world are on Expedia.com, and Marriott believes the best hotels in the world are Marriott Hotels. And we all, and Chase believes that their partners are the best partners in the world. We all have that bias. So, the only type of tool that would not have that bias is a tool that you actually pay for. Now, can you build a big business with that? I'm not sure. I mean, you know, we have trained consumers that all this travel content and stuff should be free. So, I'm not sure if there's a big enough market. There's no doubt that there's many small markets. We're seeing a lot of travel clubs and members' clubs who travel, like, charging sometimes a lot of money. But these are, in the grand scheme of things, very small compared to the size of our industry. So, I don't know, but that's something that I would love as a consumer, that I would love as a person in the economy. Now, my favorite business model these days is really upside business models, effectively where you're creating further economic activity and taking a piece of that. I've talked for a long time about how SaaS, the business model, is under threat. Because at the end of the day, if you build me a widget and the widget's pretty good and I have to pay you every month or record for that widget, there is Lovable and Codex and whatever where I can go build a competitor. And what a lot of people in tech forget is that for the average operating business in the world, having a 70% as good solution for 30% the cost or for free, is often a lot better than a 100% solution for 100% the cost. So, it's one of those things where, if you're purely creating a solution, I think it's under threat. And if you're saving me money, that's also under threat. I might be willing to pay you a lot to save me money this year, but now that's the new baseline. So, I'm not gonna be willing to pay you a lot next year to save money again. So, the only place I'm really gonna be willing to pay you a lot into the future is if you create commercial outcomes for me that I want to happen in my business. So, as we think about this dynamic experiences thing I've been talking about, well, I can imagine a new capability, whether it's a B2C OTA or travel agency or whether it's a B2B capability provider enabling other people to do that. But all of a sudden, if you increase the size and the margin on the package, you deserve to get paid some percentage of the economic lift you brought there. That's a business model that I believe in.

[00:16:12] Matt Welle: And imagine everyone can build anything, but not everything is worth rebuilding, because of the cost of it. How does a hotelier that's listening to this think about what do I build? Like, a chatbot? Anyone can build a chatbot today. Like, but why would you if you can just buy it off the shelf? So, how should they think about what they buy and what they build themselves?

[00:16:31] Gilad Berenstein: Yeah. So I'll say, step one, the low-hanging fruit, the easy stuff, is whatever your team is doing over and over and over again right now clearly is not accounted for by your system because some human is doing it over and over and over again. I would start there. This is the crawl, walk around. This is the perfect place to crawl, real value, etc.

[00:16:46] [AD]: Thank you so much for listening to Matt Talks Hospitality. I've really enjoyed these conversations. I did wanna take a moment to talk about the Mews Revenue Management System. Two years ago, we acquired Atomize, and we then, from that moment, started rebuilding the whole platform from the ground up at the core. So, the idea was that the Mews Operating System becomes the core platform that everyone in the hotel lives in, but we had to rebuild the whole user experience inside that same platform whilst we optimize the algorithms. That product is now live. Hoteliers and revenue managers love it because once they log in to Mews, they get their business intelligence insights. They can manage the rates and the connectivity to the channel managers. And now, all the price recommendations are sitting in that same platform. And this makes the life of revenue managers much more interesting because it turns them into revenue strategists. You can see this demo straight away, directly inside the system. We can get you live by tomorrow, and we can get recommendations flowing to instantly make your business a better business. I hope you enjoy it.

[00:17:40] Gilad Berenstein: As we think a little bit further and a little bit higher, I think really the question is, where is your business truly different from the other businesses in the ecosystem, and these are things you wanna build, and where is it the same? So, the easy example is accounting. At the end of the day, very few hotels are gonna have to do accounting differently than other hotels. That is a place where you definitely want to license, to borrow, to take from someone else. This is again in a further future where you might be able to get, like, literally an AI agent that is coming from whether it's an accounting firm or QuickBooks or Mews or whoever that allows you to do these kinds of things. So, where there's truly no differentiation for your brand and the customer experience, I would not be building my own software, my own experience.

[00:18:20] Matt Welle: The outcome you wanna get is always the same in a finance system because it needs to always balance at the bottom line. Right?

[00:18:25] Gilad Berenstein: That's exactly right. So, that's the easiest category. And I would even say in category two, you wanna go outside, which is things that don't make a difference to your customer. Right? Like, for example, like, you don't need to build your own Wi Fi gateway for people to access Wi Fi on the property. Like, whoever you're using is just fine, and no customer's ever gonna choose your hotel because you have a better Wi Fi login experience than someone else. But if there are things that are truly differentiated, it's easy to think of examples in the experience space. If you are a wellness hotel, if you're a sustainability hotel, if you're doing something like that, that's where I would really consider building your own tech in something that really accentuates your differentiation, your brand, your what you're bringing to the market. That's where I would consider building myself.

[00:19:03] Matt Welle: Have you stayed in a hotel and actually, at the end of your trip, thought, wow, they actually used AI really smartly. And, like, I've had differentiated experiences because I stay in a hotel every week, and I have yet to experience that.

[00:19:15] Gilad Berenstein: Yep. So, the answer is probably no on the travel experience. Like, I have not yet been to a hotel that did something with technology. Obviously, they have staff that's done amazing things for me with technology. Where I have is in kind of operations and back office. So, I like to use the example of the Shashi Hotel. It's a small hotel brand in San Francisco. I believe they've built 3 hotels in the Bay Area, and they've done some really incredible things. So, firstly, it was a new build, so they could do things that you can do in a new build. They put a bunch of sensors in the walls - sensors for noise, for heat, for dampness and water, for all sorts of different things like that that can effectively control the experience and give a lot more information in real time. So, one immediate benefit they got is a dramatic cost in their property insurance. Because now, before the leak goes across the entire tower and every room in the tower has to be renovated, the leak could be caught in the specific bathroom, or it's happening in the wall or something like that, to protect that. Another experience is around things like noise and heat that can begin to give automatic alerts to the room mix or to whatever without me having to complain, without a stab at how I'm gonna go up about things like the noise, the different things they're doing like that. So, there's a lot of hardware that's beginning to be put into hotels, especially new builds, that creates efficiency and operational capabilities inside the hotel today. And then the main thing is, once the hardware is in the walls with the right kind of software, you can actually begin to think about more and more complicated use cases within there that would do even more. So, that's one small example, but we are just getting started. And, of course, in the hotel space and, of course, in the transportation space as well, things take slower because you live in a physical world with physical property than we expect them to come from an OTA or someone who's living purely in the digital world. So, just because maybe some of them are moving faster, I think it's a short-term difference, not a long-term difference.

[00:20:57] Matt Welle: Speaking of, like, real-world applications, I just bought this thing. It's called Eight Sleep, which is like a bed cover that you put on your bed. And it comes with, like, a machine which cools down your bed throughout the night. It uses AI to measure, and it's like I'm on day five now, and I've had the best sleep. And, like, you can feel maybe the energy is quite high on this podcast because I had a phenomenal sleep because it's, like, getting so much better every night at figuring out how do I get into my deep sleep and the best rest.

[00:21:25] Gilad Berenstein: This is like a really small tangent, not really about AI, but I recently had a conversation with a number of wellness experts, a woman named Mia, who opened a bunch of the wellness offerings for, like, a bunch of the Star Wars branch, like the St. Regis and the West or whatever, has really been thinking about these for a really long time. And they talked about all these hotels doing incredible things to offer wellness or whatever and forgetting the bare minimum, which is quality sleep.

[00:21:46] Matt Welle: The thing you do the most is sleep.

[00:21:48] Gilad Berenstein: Yeah. That's exactly right. They talk about how, like, you might invest in these beautiful blackout shades that really block all things out, and then there are all these lights from all the little pieces of hardware that you put in the room that make it really difficult for some people to fall asleep. And the same thing you were talking about with the temperature control where, like, it's like they might have these amazing systems, but they're not very precise. So, I actually think that some of these small things make a big difference. And, again, this is where AI agents can also help. So today, I, effectively, can just set my thermostat to one temperature around there. But in reality, I actually prefer to fall asleep in a pretty cold, maybe 66-degree room, but actually would prefer to wake up in, like, a 68 or 69-degree room. And, like, again, that's a small thing, but as soon as you have AI agents and this smart hardware in the room, all of these new use cases become possible once you determine that matters to your customers.

[00:22:33] Matt Welle: And I think people with money that do luxury travel, the pillow menu was something from the 90s; that way you could pick your pillow. That is not the thing that gets me the best sleep. It is my bed at home that's infinitely better than the best hotels. And I'm going this weekend for a personal trip, and I'm already worried that I'm not having my bed with me. And I would pay for a hotel company that would actually invest in a bed where I can just plug in my app so that it instantly knows the sleep that I need to get. So, if you find out a hotel brand that does this, I will go there.

[00:23:02] Gilad Berenstein: You see, I love that. The small version of that is when I do car trips, I bring my own pillow. I'm just gonna sleep better on my pillow than I will on some random other pillow. And that's the thing. So, I think wellness is about those things. It's about sleep. It's about nutrition. And as much as I love a wonderful spa, and, don't worry, I really do love a wonderful spa, there are things that I think are much more basic that can create a lot of value.

[00:23:23] Matt Welle: So, you invest in AI and hospitality. What are the things that get you excited as an investor where you put money in? Because there's so much; we can all build anything with AI today. But when do you think, like, yeah, this is one where I should put money in because this could be big?

[00:23:37] Gilad Berenstein: Yeah. Absolutely. So, a few areas I'm excited about. So, one is an area called Back-of-House Ozempic. At the end of the day, we have a lot of staff in the back of the house doing things that are not only not enjoyable for them to be doing, but are not adding direct value to our guests. So, as I say all the time, I don't want any fewer humans working in your hotel. I want fewer humans locked in the back hallway, in the back office, in the laundry room, etc. So, one is I wanna invest in operational efficiency. I wanna invest in things that allow the staff to go be with the guest, to do hospitality, which I always say is good for the guest and good for the staff. That's one area that I'm really interested in investing in. Second is everything around AI harnesses, AI rails, AI infrastructure around the agentic future. I say all the time, it's really hard to predict. Actually that's not possible to predict how these channels will kind of even out over the next 5 or 10 years, but there are probably things that we know will be true no matter what, like exposing content in different ways and things like that. So, very much looking at investing in AI infrastructure and rails and that entire kind of ecosystem. And then the third one I already talked about a bunch, which is dynamic future. I really believe that the win-win-win is on offer, and I'm looking to invest in all those kinds of things. One more just to add in. I'm a huge believer in the financialization of travel, something I've been writing about for almost a decade now. Hopper is, of course, the poster child of these kinds of solutions, you know, the leave for any reason, the price lock, those kinds of things that are effectively our financial instruments. These are puts and options and insurance contracts that are being built for the travel industry. And I love them because they actually solve real problems. Right? So, if you think about the Hopper Solutions, it's like, we've all been concerned about booking a hotel when the price goes down. We've all been concerned about booking a hotel and the photos are not exactly what we imagine. We've all had these little concerns. So, what they did was they said, I will solve your consumer problem. And, of course, what we do in the economy is we pay for people who solve our problems and do so in a way that's really high margin. So, I'm really excited about that. And then as I talk about this dynamic future, there's a lot of financial capabilities that come together to make that possible. So, that's another area that I'm really excited about.

[00:25:32] Matt Welle: What's one of your portfolio companies who are killing it right now?

[00:25:35] Gilad Berenstein: Well, today, since we're recording it, I can share that Faye, a travel insurance company from Israel that I've been a part of for a long time and I'm a huge believer in, raised a very large round, a 50-million-dollar round with a great local VC here in town called Madrona, who led the round. And, really, the insurance space has been a space that's been largely overlooked and forgotten in our industry, partially because there's been a long time when even people like me, who are trying to be creative or having a hard time being creative, or what can you really do in this space. I think Elad and the team there have come up with some real solutions, and their real attempt is to bring value to the traveler even when something doesn't go terribly wrong on the trip. And then, of course, when something does go terribly wrong, making sure it's really easy, like, making sure it doesn't become another awful experience for you to have after that. So, that's one I'm really excited about.

[00:26:18] Matt Welle: But where do they sell the insurance? At which point do they get in touch with the customers? Is it at the OTA level or at the PMS level? Or…

[00:26:24] Gilad Berenstein: So, right now, it's the OTA and agency level. There's also, of course, a direct-to-consumer element as well that can be booked after you make the booking, but I think they have a lot of ideas of how to integrate it with many other partners as well, over the coming years. And I imagine this new funding…

[00:26:35] Matt Welle: If we do online check-in, right, I can imagine, like, pre-arrival, you'd wanna slot in, like, an insurance program.

[00:26:40] Gilad Berenstein: That's exactly right. And combining these things, we were talking about this financialization. Right? So, I'm in Washington, and I'm sure people will know that this summer, we've had terrible smoke and terrible fires here in Washington. Like, we have a family vacation plan for two weeks from now that is very likely - I shouldn't say, let me be optimistic. I'm very hopeful the smoke will clear. We'll be able to go and enjoy our family vacation, whatever. But the thing is, there actually isn't a smoke insurance policy for a trip like that, and that's something I probably would have bought. And using financial instruments and AI, you can probably, on the fly, create this kind of policy for me and add it at the hotel booking checkout or on the pre-arrival or whatever it is. So, yeah, I think there's a lot of opportunity there as well. One other one that I'm a big fan of is Resort Pass, another company I've been involved with for a long time. You know, today, hotels mostly monetize an out-of-town audience. Resort Pass: 70% of their guests are locals. They create opportunities for locals to engage with hotels, to bring more business to those hotels, and often to utilize underutilized assets that the hotel already has - another company that's seen unbelievable growth over the last year.

[00:27:44] Matt Welle: So good. Like, the amount of times that I'm in a city and my hotel doesn't have a facility, like a gym or something, that I'm like, where do I go? I just wanna work out. That's all I want. I'll buy a pass into a hotel any day.

[00:27:55] Gilad Berenstein: A 100%. And it was, for example, like, I love staying at, like, boutique hotels and historic hotels, and these kinds of really beautiful small hotels. There's never gonna be a beautiful pool and hot tub; never is a big word, but there's very rarely gonna be a beautiful pool or hot tub at that thing. And for the day, I'm like, I'm happy to go pay $50 to go spend a few hours by a beautiful pool.

[00:28:11] Matt Welle: My husband knows that I only stay in hotels with pools for private. Like, for business, I stay in any bed.

[00:28:15] Gilad Berenstein: Anyway, yeah.

[00:28:16] Matt Welle: But for private, I'm like, yeah. I want a pool. And I might not even swim in it. I just want the knowledge that it's there when I want it.

[00:28:22] Gilad Berenstein: You see, my equivalent crazy thing is a hot tub. I would feel like if I can spend 15 minutes in the hot tub at the end of the day, it's like, whatever, good or bad, the day is gonna end perfectly.

[00:28:32] Matt Welle: Chatbots are everywhere. If you were a hotelier, would you put a chatbot in or not?

[00:28:37] Gilad Berenstein: I mean, the answer is probably yes. Like, I'm sure there are times where people are calling the phone, and everyone else has already picked it up. I'm sure there are times where they just want the towel or that little thing in there. So, I think the answer is yes, but I would not ever replace opportunity for human connection with a chatbot. And I use the example of check-in all the time because check-in is one of those experiences where I was I don't wanna give the name - but I was at a very high-end hotel in Asia recently. And we had about a 10-minute check-in experience, and the entire thing was transactional. And I get it. There's lots of regulation about they need photos of my passport. Lots of regulations

[00:29:11] Matt Welle: Oh, you know what I hate? The leather pads that they come with, with an envelope on. I was like, I just wanna get to my room. I'm tired.

[00:29:16] Gilad Berenstein: That's exactly right. And I always say, like, if it was late at night, could they have automated that whole thing and just taken me straight to my room? Or since it wasn't late at night, could they have done a hospitality experience? Just to continue the analogy, this hotel, an unbelievable spa, an entire spa floor on, like, the 30th floor of the hotel. Imagine if, instead of a 10-minute experience, he could have said, "We actually know that you're really passionate about this. Let me show you right away the product hotel you most wanna wanna see." Imagine if I were telling you that story instead. That's something I'm really excited about, where AI can take care of the legal, of the operation, of the required things so that humans can do hospitality.

[00:29:51] Matt Welle: Can I talk about something else that irks me? Because I'm hoping that the LLM will pick it up so that my next hotel will not do this. When the bellboy goes into the room and he wants to show you every facility, like how the AC works, I was like, no, I got it. I know how a hotel room works. And I'm like, don't waste my time on that or waiting for the luggage for 10 minutes. Like, then I'll just drag it myself. I don't wanna wait.

[00:30:11] Gilad Berenstein: Let me give you another AI example that I love. So, this is for Atlantis The Royal in Dubai, which some people would argue is the most high-end large hotel in the world. So, they have, for security reasons, all the staff at hotels in Dubai have these, like, location beacons for different security reasons in Dubai. But what does that also do is, and we're gonna, like, we talked about the hardware on the walls earlier. When you have that knowledge, you can come up with more use cases. So, if you ask for towels or toothpaste or whatever you're asking for at the hotel, they won't tell you exactly how many minutes they estimate it's gonna take to get the shaving kit there. And my point is, when I'm asking for shaving cream, I'm usually about to get in the shower. And the real question I care about is, should I get in the shower before or after you deliver me the shaving cream? And the fact is, I actually don't even care if it's 2 minutes or 12 minutes. I just wanna know if it's 2 minutes or 12 minutes. And that information, to me, felt like a piece of hospitality, a piece of luxury, where I wasn't kind of sitting by the door twiddling my thumbs waiting, but I knew that I should jump in the shower or whatever. A small, tiny thing that to me was a really nice luxury experience.

[00:31:09] Matt Welle: That is really good. When we met last in Amsterdam, you told this story about your daughter and fairy tales and personalization. Can you maybe share that? Because we've got a different audience here than we probably have on stage there.

[00:31:19] Gilad Berenstein: I would love it. And this is, like, the ultimate culmination of my dynamic products concept, dynamic offers concept. So, I have a daughter named Rose. You might be able to hear her. She's playing outside my office here. And she really, really loves princesses. So, they would say, assuming that one of you sold me a dynamic product, what would that look like? So, when the van came to pick us up here to take us to the airport, the itinerary would be turned by AI into a video that would tell us about all the cool things we're about to do in Paris and all the princesses we're gonna run across on our trip. When I got to the airport, I would have four seats selected. Let's say this is a luxury trip. Not just together, but some of you might know Lufthansa has this new business class where the four seats face each other. So, me and my kids can watch a movie together and have lunch together on the plane. When we got to Paris, the AI would choose the right hotel from the huge proliferation of hotels in the world. I always remind people that when I started 15 years ago, Marriott had 8 brands. I think today, they have 37 brands. Each of them came to just, like, a different customer. So, huge proliferation of hard products. So, selected the right hotel. We got to the hotel. We got upstairs. Hopefully, they were a quick check-in experience. And there were princess toys waiting by the bathtub that Amazon dropped off this morning to make sure that my kids' transition from jet lag into the trip was really fun and enjoyable. And the next day, when a tour guide came to pick us up for the history walking tour, he had AI-generated stories about princess fairy tales sent to him so you can include those stories in the walking tour so my kids are not bored as we walk through Paris. And then my wife happens to be gluten-free. And most of most people know that most Parisian pastries do not offer gluten-free options, so Uber Eats stops at the last moment with a box of gluten-free cookies so my wife can also enjoy a Parisian dessert. I can go on and on and on. But the point I always make is, in this whole example I just shared, there are only two pieces of hard cost I introduced into the trip - these were princess toys for the bath and gluten-free cookies for my wife. Let's call each of those €15. So, €30, but that's, you know, it's a lot of money for gluten-free cookies, but whatever. Let's call it €30. So, I just increased your cost of delivering the trip to me by €30. How much more am I willing to pay for this trip? I think it's almost certainly €300. In a luxury setting, it might be €3,000. So, we're now talking about an order-of-magnitude difference purely by telling a story and stringing together these pieces of experience for me in a way that's really all about me and my experience. As we talked about before, you get the triple whammy. You get a better trip for me, increased margin and share of wallet for you, and then many more business. In this case, it's whoever sells princess toys, whoever sells gluten-free cookies, whatever, getting to participate in the upside of tourism. That's a future I'm incredibly excited about.

[00:33:46] Matt Welle: I think it's a really exciting future. It's like you said, if we learn anything from COVID, it was that travel matters, and you'd wanna double down on the experience, the memories that you have when you're locked up in your home for, what, was it, 2 years that we were locked up?

[00:33:59] Gilad Berenstein: Absolutely. And I always add, like, as we think about the future of our world in this AI era, what I call the great commoditization, assuming we don't get one of the dystopian, awful outcomes, almost all of the flavors of the positive outcomes include more time for leisure, more time for self-discovery, and more time for human connection. So, I am just a full believer that we're already in an amazing industry today, and it will be one of the fastest-growing industries over the coming century.

[00:34:22] Matt Welle: I agree. My last question is always, like, what's something that irks you in a hotel today that actually, with technology, we can solve in a year's time that you would like to see a hotel brand solve for you?

[00:34:33] Gilad Berenstein: Yeah. Absolutely. So, this is tiny and nitpicky and hashtag first world problems and all this stuff. So, when my phone charger is on one side of the bed, and they don't realize which side of the bed I'm sleeping on, it just feels like this huge human oversight where they put the water or the slippers or whatever they're doing on the other side of the bed. And it's just really, really small, and it's super easy to fix. I haven't think in less than a year's time. So, there's already companies creating computer vision capabilities where, effectively, the maid or whoever can take a picture of the room when they're done setting it up and done cleaning, and it can give them real-time alerts. Some of this can be brand standards of, you know, we put two water glasses, not one or whatever the brand standard for the hotel is. But a lot of it can also be about guest experience, about the fact that it looks like there's actually three pairs of shoes in the room, and they do remember to put three towels in the bathroom, not two towels in the bathroom. It looks like the chargers are on the left side of the bed should you put the water on the left side of the bed. So, I think these types of things supercharge our ability, and any human's ability to perform their job at a high level. And I'm not trying to make this sound easy. Like, right, like, if your job is to clean hotel rooms, you have many more rooms in a reasonable period of time. It's a hard job. I'm not saying that they're, like, lazy or whatever. This is where AI can effectively supercharge them and help them think about the things that they are not able to think about themselves. I know that's really exciting. So, are robots coming to a hotel near you soon? For people who've been to Asia, you've probably already seen some of these robots deployed in hotels and restaurants. But they're…

[00:36:05] Matt Welle: I've seen it for room service or, like, one driving around that you put plates on. Right? That's the baseline kind of robot.

[00:36:10] Gilad Berenstein: Yeah. It's the baseline. And then, actually, the most utilized robot in hotels is a vacuum robot. So, effectively, an industrial Roomba.

[00:36:14] Matt Welle: Roomba.

[00:36:15] Gilad Berenstein: You're putting largely into conference space and into kind of large meeting space. You're not really using it in hotel rooms as much yet, but I think that's all gonna change. So, I'll give you one example. There's a company called Bellboy Robotics out of Israel. Their robot's called Billy. They've deployed about 30 or 40 kind of mass-market hotels in Germany and Italy right now. And Billy can do a few things. It can do a pre-cleaning. So, it can come into the room before the house cleaner can come in, and it can remove dirty towels. It can remove garbage bags. It can send photos to maintenance if things need to be, like, you know, fixed or whatever it is. It could do a small number of tasks before. Additionally, it can scrub a toilet, which I have to imagine is the least enjoyable part of anyone's job who's cleaning a room. And I say all the time, we live in an industry full of labor shortages. If you simply said, come be a maid in our hotel, but no one has to scrub a toilet, you would have the most applicants of any hotel in town. So, I think robots are coming…

[00:37:07] Matt Welle: Does this robot exist and is it operational today, or still in the ideation phase?

[00:37:12] Gilad Berenstein: No. No. So, it's operational in about 30 hotels in Europe right now, and they're preparing to scale and deploy many more in many more places. Today, Billy has one arm coming off the back. Eventually, it's gonna have two arms so it can carry more things. The other use case is as the company's name implies, Bellboy Robotics. And the company is also assistant Bellboy. So, the human Bellboy is taking the bag out of the car. It's greeting you; it's saying, "Welcome." It's opening the door. The robot can effectively take the bag. I think they can pick up a bag of up to 60 kilos or 70 kilos, a relatively heavy bag, and they can kind of stick it on its back. And they can now go in the elevator and go drop it off either outside of a hotel room door or in some sort of common space upstairs, something like that. So, I think robotics is really exciting. And it goes back to everything we talked about earlier: how do we allow our humans to do more hospitality, more human connection, while still ensuring that all the operations are being done at this highly efficient way.

[00:38:01] Matt Welle: I worry often that the education system isn't ready for what AI is automating, so that the humans we deploy in hospitality are actually focused on that. And how do you worry about the hospitality education system today, and what should change there?

[00:38:13] Gilad Berenstein: Yeah. I mean, broadly, I have a six- and a three-year-old, so I'm worried a lot about the education system because of AI. And I ask the question a lot of myself. You know? A lot of my success has come because my dad introduced me to the stuff in the back of the living room so long ago, and the question to me is, like, how do I give my kids the advantage of really understanding this technology intuitively, not being, you know, kind of tourist into this world of AI without robbing them of the creativity and the freedom of the amazingness of being a child? I think that's a really difficult question. Bringing it into our sector, right, like, it's really about the human connection and human impact, right? And when I speak at universities, that's the number one question I always get is, like, well, what the heck do I do in this new era? And my answer is always the same. I don't know what, but I know how to think about it, and it's the same with hospitality. I always say AI is coming for everything, except for those things that we choose to keep in the human domain. Look around your peers, figure out what you think your peers will value being done by a human, and become good at that. So, now, of course, in travel, there's some things that are pretty obvious. We're all seeking human connection. This is how we were bred. This is who we are. We are seeking human connection. We're seeking novel experiences, and we're seeking sense of belonging. Right? I mean, Airbnb's whole genius market has always been about a sense of belonging. That's, like, why they've been so successful for so long. So, those are the things I would focus my staff training on, is how do you make someone feel really welcome? How do you make them feel like they belong? And how do you make them feel like they can be authentic to themselves and the place at the same time? You know, when we started UTrip, we spent a long time with psychology professors at the University of Washington studying consumer decision-making. And when you ask people what leads to the best travel experiences across genders, across ages, across everything, the answer was almost always authenticity. The authenticity has two elements to it, and our staff needs to enable us to have both elements within a trip. First, authenticity to the place where I am. If I'm in Rome, if I'm in Amsterdam, if I'm wherever, I wanna feel like I'm living that way. I'm eating that food. I'm engaging with the people and the culture and the art, whatever, but I am still me. And I do also wanna be authentic to myself and my views and my beliefs and my needs, whatever. Good hospitality professionals help you balance those things. So, that's how I would think about training our future staff. And guess what? Their job would be more enjoyable than being a clerk in a back office somewhere.

[00:40:26] Matt Welle: Like, I think the skill of asking open questions and deeply listening and then capturing what you've learned into your systems and let the LLM do with it what it wants. But, like, that deep conversation with the customer is actually really exciting once you figure out why they're in town, what the special occasion is.

[00:40:42] Gilad Berenstein: And this is where AI can help as well. Right? So, another company I'll share is Abra. Abra is a guest intelligence platform that can, today, they don't listen, but maybe in the future, can just straight up listen to your conversation with the guest. But the example I would use is, let's say you're talking to that bartender in that hotel in Hawaii, and you say, "I just love mangoes so much. Can I have a mango margarita or whatever it is?" The bartender could really easily, whether it's typing or through voice, say, "The guest in Room 342 really loves mangoes." And then tomorrow morning, a housekeeper will get a suggestion that instead of the bowl of strawberries, you put a mango in my room for the morning amenity. Right? Small example, but this is where the AI is directly furthering that human connection.

[00:41:19] Matt Welle: Yeah. I stayed yesterday in a hotel, and I've been there 4 times before. And I had made it clear that, you know, I didn't want a bucket of frozen well, actually unfrozen, like, water in there with a bucket of warm wine. And they actually listen, and the AI is, like, feeding it now because that's in my profile, probably, I'm imagining. And now they know I'm lactose intolerant, so they don't put, like, a cheese platter there for me, but they find they found lactose free treats every time I come, and it's now consistently done twice. And they're like, okay, there's something here that they're using the LLM to field my profiles in some way, and I was like, uh.

[00:41:52] Gilad Berenstein: And, again, you highlighted a win-win outcome. You know? So, my example is I don't drink much, but I really, really love chocolate. So, all of a sudden, instead of a bottle of wine that even at wholesale is gonna cost you $30 or $40 to put in my room, you can now put that beautiful $8 bar of local chocolate. And not only did you spend less than 20% as much, I feel seen and cared for and engaged with. And, again, I think that's a future we should all be excited about.

[00:42:15] Matt Welle: I could literally talk to you forever, but thank you for sharing. If you're listening to this, we talk very fast because I love your energy, but you can slow us down as well. I hope you figured it out if you've made it this far into the episode, but thank you for joining me today.

[00:42:27] Gilad Berenstein: Yeah. Absolutely. It was a blast. And thank you for what you guys do for our industry.

[00:42:30] Matt Welle: I hope that you're enjoying this episode. Because if you do, make sure that you press that subscribe button. It really helps grow the audience and keep you informed whenever a new episode drops. We have some of the best hoteliers on this platform, and I love asking them the hard questions so that, hopefully, you can improve your hospitality business as well.