The Rise of Agentic AI in 2026

AI that writes is no longer the headline. AI that acts is.

In the first episode of our Agentic AI Series, we sit down with Gilad Berenstein of Brook Bay Capital, Tahnee Perry, AI Expert Consultant, and Christian Watts of Magpie to break down agentic AI in plain terms. We unpack how it differs from generative AI, and why this next phase is less about single tools and more about systems of agents that can take real action across your workflows, data, and operations.

We also get candid about what this shift means for the travel industry right now. Gilad explains how agentic AI is reshaping startup economics, where technical moats are shrinking and advantages are shifting toward trust, distribution, customer lock in, and proprietary data. Christian challenges how travel tech teams think about capability, arguing that leaders need to understand what’s possible with today’s tools, not what was true even a year ago. Tahnee brings it down to reality with practical, field tested use cases, from data analysis and market research to automations that remove friction without losing the human touch.

From there, we map the traveler journey. Discovery and inspiration are rapidly moving toward chat and voice, while booking still rewards simplicity and clarity. We dig into AI visibility, including SEO, AEO, and GEO, the importance of context, and why seamless human handoffs still matter when agents fail. We also explore why travel advisors are not going away. Instead, they are becoming more valuable when they double down on taste, judgment, and human connection while automating everything else.

If you are leading a travel brand, tour operator, OTA, hotel team, or agency, this episode will give you a clearer lens on where to start, what to automate, where to stay human, and how to move forward without getting stuck in the hype.

Because the real shift is not just AI getting smarter. It is AI getting things done. And the companies that understand that early will not just adapt. They will lead.

Thanks to Maya for sponsoring this series!

👉 Listen to The Rise of Agentic AI Now

🔥 Special Thanks to our Season 7 Title Sponsors for their Support: Bokun, Civitatis, Intrepid, Kaptio, Propellic and Protect Group

The #1 B2B Travel Podcast Globally. Over 100 Episodes. Listeners in 125 countries. New Episodes Every Weds. 

https://www.traveltrendspodcast.com/

Transcript

Welcome And Series Kickoff

SPEAKER_03 0:05

All right, here we go. I am absolutely thrilled and honored to be surrounded by three of the people that I most respect and admire in the AI and travel space. Of course, we've got Gilad Berenstein, my great friend and longtime colleague, joining us. We have Tani Perry, which is someone that I have really, over the last couple of years, come to really respect and admire for the work that you're doing in AI and travel. And Christian Watts, who has been a big part of our podcast as well. He's been on several episodes, our AI Summit. So I'm thrilled to have these three individuals to kick off a three-part series on agentic AI and how it's transforming the travel industry. And first, I want to get everyone's take on agentic AI. I've been presenting at a number of conferences recently, and the way that I have been shaping it up is that generative AI allows you to create content and agentic AI allows you to take action. And I know, Gillad, I want to start with you because you had presented at the Virtuoso conferences. I had the privilege to do some presentations this year. I know you're a big part of their community. You're on the board, you go to a lot of their events, you have highlighted how Gen AI was a tool to accomplish your goals. And obviously things have evolved a lot since then with agentic AI. And so, how would you today in 2026, how would you define agentic AI? And how are you explaining it to your investors, companies you work with, and on stage?

SPEAKER_00 1:27

Yeah, absolutely. And thanks for having us. I mean, keeping this at a really high level, I think of the era we're entering right now as the agentic era of AI. I'm sure everyone in the audience knows this. Agentic is just a fancy word for agents. So the agentic era is just an era where we have many, many, many AI agents deployed across the world for us to use in our personal lives, in our business, in our trip planning, et cetera. And the thing that I think most about in the agentic era of AI is I think a lot about swarms. And swarms are effectively constellations, groups of AI agents, each on their own, kind of dumb, but together able to accomplish truly incredible things together. So the agentic era is about an era where we deploy these agents to help us do all sorts of things. And I think your summary was perfect, going from an AI that can communicate to an AI that can act.

SPEAKER_03 2:10

That's great. Thank you, Gilad. And I know you share a lot on this topic on LinkedIn with your podcast as well, of course. We share producers, most of our uh audience does know that. Um, and so I want to round out the conversation because Gillad and I have had a number of conversations on this topic. And I know a lot of people have benefited from those discussions. But Tawny and Christian being a part of this, and obviously there's certain things that you'll disagree on. And Christian, obviously, I always value your um sometimes obtuse insights as well. And so, but before we get to you, I want to bring Tawny in because Tawny is someone that I have actually referred her to a number of my advisory clients, and uh I have heard such amazing things back from them that we're like, we love working with Tani. She's taught us how to use Claude Cowork, and I can't believe what I'm now capable of doing because you've given people hands-on training, and that's something that I really wanted to make sure that we go from theory to practice in this conversation. So tell us, Tani, what is your view of agentic AI? Kind of level set with everybody, how this is different than the AI tools we've seen over the last couple of years.

SPEAKER_01 3:09

I think of Agentic AI as the AI having the ability to take autonomous actions. It can do things automatically without you having to sit there and press a button. And uh the way it works in in practice is if you've used Chat GPT or Claude and you're using the chat function, you put in a request, you tell it what to do, it does that, it spits out the output, and then you keep going. And it's this back and forth. Agentic AI is you build a system. So, as an example, I'm coaching a lot of people and teaching a lot of teams how to use Claude cowork at the moment because it is a very popular platform and it is amazingly powerful. But what you do is you build a system where you create all of the knowledge documentation, you connect it to all of your platforms, and then you design the system instructions to tell it what to do. And you can even schedule it. So it will go off and it will perform all of those things without you having to pay attention. Now, there's pros and cons to that because if you don't design the system really intentionally, it can go off and do some unintentional things that you might not want it to do. So you just you have to be, I think, planning when you're working with a gen tick AI, you have to do a lot more planning and system thinking than you do when you work with gen AI.

SPEAKER_03 4:20

That's great. Thank you, Tani. And to bring Christian into the conversation too, he continuously shares from all the conferences. I love uh hearing Christian speak, both on stage, but also when he does his short little videos in a park in New York or from a conference stall that belongs to someone else that didn't show up on day three, wherever you suddenly appear uh at an event. Um, and my wife actually knows your voice because I'll be in the kitchen, I'll be listening to you, and it's come up a number of times. She's like, Is that Christian again? I'm like, and I'm just giggling and smiling. And she knows she knows by my reaction, and she knows your voice. You've never met, but uh, I really enjoy your insights. And the certainly one thing you you were saying, and this is where I'd set the stage for your contribution here, is that you made a great point. Everyone's been talking about they've been doing AI for years, but really, guys, it started with Gen AI, which we know 2017 this was developed, 2022 is when this came out. We're having this conversation in 2026. Really, we're talking about what's happened so significantly in the last few years. So, Christian, tell us from your vantage point as an operator, as a founder, and entrepreneur, how do you look at and talk about and describe agentic AI?

SPEAKER_04 5:23

Yeah, I think we've all lost track of what agentic AI is, right? It's just all blended in. It is part of Gen AI. It's not different, it's just an extension of Gen AI. It's an extension of what ChatGPT kind of started. And in the same, in the same sort of line, everything's agentic now, right? So when you talk to your standard Chat GPT, it goes off and does stuff. You could argue if it goes and does a search, that's agentic. So I think both the definitions that we heard are are are fine. Everything's agentic now. It's gonna continue just getting more complicated. For me, it just it just does things, it's become a tool where it just does things like Claude Code, Claude Cowork. You give it a simple instruction now and it goes off. Whether that's spinning up a swarm or whether that's doing multiple steps, it's agentic. I think the piece that most people think is agentic is when it talks to other people and the bots start talking to each other. That's not really happening as much yet, mainly because of trust, but that's probably the next stage. But agenti's here today, I don't think most people know that or see that, and it probably doesn't matter. It just means stuff works now, and you've got access to multiples.

SPEAKER_03 6:28

Yeah, and that's where I absolutely want to, over the course of our conversation, get into some of the specific use cases and even actually just making sure that the terminology that we're using is accessible to everyone that is just wrapping their head around generative AI, and as you mentioned, it being an extension of and kind of the most people are not going to distinguish between gen AI and agentic AI, but the reality is part of this big shift and the change we're seeing is all of the capabilities that these LLMs are now introducing by virtue of the fact they can solve complex tasks and they can take action. And so I want to get into the realities of what that actually means for travel advisors, for tour operators, for everyone in our space, and also to leave them with some recommendations for how they should be utilizing these tools today. But let's stay on a more higher level and eventually get more granular. Uh, given each of your roles, investing is obviously a big focus for you, uh, Gilad. So I want to ask you this question first. We're seeing more and more companies in the travel industry receive funding, partly based on the fact that they are now uh an agentic or they're an AI first company. We saw this with mobile first. I'm sure you see a lot of pitches for companies that are telling you that they're um native AI startups. And so when you're looking at from an investment perspective, where do you see or what are you looking for when it comes to agentic AI? And what do you think is so significant about this shift and what it means for companies that are looking to raise, whether they be startups or established brands that are leveraging this technology this technology to be able to scale?

SPEAKER_00 7:58

Yeah, absolutely. I think ultimately, if you get any real investor right now into a private room in a private conversation, they will tell you that investing in startups has gotten a lot harder over the last two years. At the end of the day, we spent much of the last 20 years looking for technical modes. As I often write, the quality of your technical founders was a differentiating enough that was often deterministic in investor decisions. And technology modes are effectively gone, especially in industries like ours, like travel, where we're not building deep tech solutions, we're building applications of AI, applications of technology. The tech mode's effectively gone, which makes it really difficult for early stage investors to differentiate when everyone's traction is at the same kind of early, low levels. And the thing is, Identific AI specifically is really changing the entire like financial model that VCs have kind of built over the last 40, 50 years. It's changing the amount of money startups have to raise and the timing with which they have to raise that money. Many predict that they'll have to raise less money, especially at the early stages when they're trying to build product, because a lot of that product is being built with no code and AI agents and other things like that to change the economics of it. The size and the per and the type of people you hire on your team changes. And of course, with the changing size of the team, you also change the economics for the VC, how much money goes in at which stage, how much equity they get to take out at that stage. And then the timescale changes, how long it takes to prototype and to build something. And I want to stress that from a VC perspective, the timescale is not necessarily a good thing. The fact that things can be built and prototyped faster for you is great, but also means that for your competition, the time to value, the time to build, the time to compete is great. And I'll go even further and say that there's an entire huge ecosystem of SaaS, software as a service companies out there. I personally am incredibly bearish on SaaS companies. This is probably more extreme than reality, but I sometimes say I never invest in another SaaS company again. But the point I always make is if Tawney goes and builds this awesome AI SaaS accounting thing for hotels, Christian are gonna say, that's pretty good. Let's go build our own competitor. And we can do it six months later for less time and less cost, which means our cost basis is lower, which means we can go offer the product for cheaper to clients. We can go undercut Tawny's product. But then the problem is as soon as Christian and I find success, Dan's gonna come after us with his own hotel accounting program. But basically, I think investing in AI startups has become very difficult. And I also want to share some sympathy for the founders. What good looks like is changing rapidly. For founders who started this journey one or two years ago, talking to VCs of what needs to be true for me to get a seed, for me to get an A, for me to get a B. Unfortunately, the advice they got one or two years ago is probably no longer relevant. And you find a lot of founders who are quite frustrated that they think what they have is good, but what investors are looking for looks a little bit different. Yeah.

SPEAKER_03 10:44

Well, it's really interesting you mentioned that, and I wanted just to uh underscore that for our listeners. When they're trying to wrap their head around agentic AI and looking at their stock portfolios, uh, we certainly saw at the start of 2026 some real corrections for SaaS companies uh by virtue of the fact the models that Claude was introducing in particular and the ramifications it meant for law firms and SaaS companies. And I know there's certain certainly people that believe that AI will eat all software. And then you hear other people that are actually, these are incredibly undervalued stocks now because they've taken a harder hit than they probably deserve, and some of these business models will be resilient.

SPEAKER_00 11:22

I mean, at the end of the day, as we talked about earlier, certain motes are going away, things like technical motes. The motes that are staying are things like customer lock-in and trust and proprietary data. Now, the thing is for these SaaS companies that are being undervalued, they have some of the only motes that matter. So while today you might look at some of these companies and say their AI is not as good as these upstarts AI, the fact that they have all the clients already, the fact that they have the data stream already, may actually allow them to kind of leapfrog these technologists. And just like startups can build technology quickly, so can the big guys.

SPEAKER_03 11:54

Sure. And speaking of building, this is where you know Tawny was part of our panel at the arrival conference in Washington last year, and she won. And this is part of the good segue to her weighing in on this, because the whole idea was, and Christian, you were there, you remember this. Two different teams had to build a solution for a tourism uh provider, and they each did presentations, and Tawny's team was head and shoulders above the other competition, and what they built within a couple of hours, then showcased back to the audience, and with the audience were the ones that decided, and I was got to be there to moderate. And so I saw Tawny in real time working with the team and building something out that would have before taken companies six months to a year to build out a marketing plan and to be able to figure out their personas and then be able to build a commercial with a character. So, Tawny, my question to you is now in this world of agentic, when you're working with clients, what are some of the most valuable use cases that you're seeing?

SPEAKER_01 12:45

Oh my gosh.

SPEAKER_03 12:46

And are you build- are are you building some of these things that Gillad is telling us can be done over a weekend to replace existing businesses?

SPEAKER_01 12:52

Some of them. And I don't think you can build them over a weekend, but you can build them. But you know, it's really interesting that your point on the competition at arrival, one of the things that I noticed about that was we did have some very similar suggestions and builds and recommendations because we were working with the AI models. And I think what people don't realize is they're getting all these outputs from these large language models, and they're they're also telling everyone else the same thing. And I think a lot of the reason why we won was maybe the presentation, the way we presented our solution. And one of the things that I talk to people about a lot is the the human element you add in to these solutions is about judgment and taste and your own experience, and that's what people will connect to. So Gillard is talking about this idea like how do you differentiate in the market? And I think that's a really big element. So on the use cases, I mean, a lot of my clients are in the marketing space, but I do work with a few that are travel focused. For me, it's all about what I talk to people about is what's your biggest pain point today? Where's all that friction? And that's the first thing you should target for building out an AI system. And it's really boring stuff. It's like doing research, understanding your ideal customer profile, uh, answering emails. The biggest thing that comes up a lot, especially for travel-focused companies, is they get tons of questions that don't really require a human to be online or on the phone for. So build a voice agent to answer them or build a chat bot on your website so people can go there and self-service. And I have this argument a lot, especially in corporate travel. I work in corporate travel a bit. They're like, the human service is like where that's so it's so important, it's so integral. And I think yes, to a certain degree, but that low level, when people are just fine, trying to find information, build an agent for that. Because I can guarantee you the traveler doesn't care if they're talking to a human or not. They just want an answer. So find a solution for that.

SPEAKER_03 14:49

That's great. Thank you, Tani. And Christian, I'm gonna bring you into this now, and then we'll let the conversation flow freely on all the other topics we want to discuss. But the reason I wanted to go in that order is because you have a very successful software company, a SaaS business model. You obviously, you know, you're based in San Francisco, you have been in the travel industry for 20 plus years. I'll be generous, because it's actually a bit longer than that. But I um but you also, by running a company like Magpie, you know travel tech. And so, and and then you're at all the conferences and you're speaking at all the conferences. So, in your view as a founder and an entrepreneur, what is your view about the opportunities presented by agentic AI uh for travel tech?

SPEAKER_04 15:31

I think it's I'll come back to what Gillad said because that was the that's a valid, it's a it's a valid cause of uh investment issues right now, but there's never been more opportunities, right? The stuff you can spin up, and you say you can't do it in a weekend. You can build a lot on a weekend, I think, today. Uh, but the stuff you can spin up in no time at all, it's in it's incredible. But it doesn't mean you're gonna get distribution, and it doesn't mean especially when you get into the big companies. I'm on a few WhatsApp groups and we discussed some of this geeky stuff, and some one of the people we know sent through a site this week and asked us for feedback. And I went on there, it was a trip planner, it was really good. I mean, it was if if you got if you'd done this four years ago, it would be everything you could ever imagine. And he did it in a couple of days, and everyone on the group's like, yeah, that's kind of cool, and then we move on. And just because you can build something amazing doesn't mean you're gonna be able to get distribution with it. So, like Gil had said, it's it's no longer the software. Anyone can build anything in in a weekend or a couple of weeks. We we've we've redone all of Magpie in the last few weeks. It's it's incredible. We're not gonna look at the code ever again.

Sponsor Break

SPEAKER_02 16:34

We'll be right back. In the world of travel, you bring the wow, those extraordinary scroll stopping experiences that travelers never forget. And BOCUN, they bring the how. As a trip advisor company and the number one booking system across major OTAs, Bokoon is the complete booking and channel management solution for extraordinary experiences. From an intuitive website builder to synced availability, Bokoon handles the code so you can focus on the guest. Plus, scale your business through Bokoon's marketplace, over 2,600 OTAs and resellers. Are you ready to grow without the growing pains? Visit Bokoon.io today to start your free 14-day trial. Did you know that over 20% of people who purchase a ticket upgrade to a refundable option when given the choice? By offering Refund Protect, you make it easy for your customers to protect their ticket purchase against emergency circumstances including illness, injury, accidents, and more. Whether you're an airline, hotel group, tour operator, ticketing platform, OTA, or transport solution, when you partner with Protect Group, your customers get peace of mind and you generate extra revenue. With a trust pilot score of 4.8 stars and over 31,000 reviews, we treat your customers as our own, earning the highest customer ratings in the industry. To find out more, check out the Protect Group website at Protect.group slash travel trends. In a world of endless travel options, how do you guarantee quality? The answer is curation, and it's the core of Civitadas, the leading marketplace for global tours and activities. Instead of just more options, their promise is more quality. Their experts handpick the best from over 90,000 activities in 4,000 destinations worldwide, ensuring every tour meets the highest standards of quality, availability, and price. They focus this winning strategy on capturing the world's fastest growing customer market, travelers from Latin America. By connecting these travelers to their global portfolio of experiences in native Spanish or Portuguese, Civitadas creates the powerful customer loyalty your business can count on. Their model is a proven formula for generating massive demand and driving industry-leading conversion. For the travel trade, the message is clear. Partnering with Civitadas is your strategy for capturing the massive outbound Latin American market and turning their growth into your revenue. Visit Civitadas.com for more details.

Discovery Changes While Booking Lags

SPEAKER_03 19:01

And now, back to the show. When you see how AgenTech AI is transforming the way that travelers experience, like they discover and book trips specifically, I know you're very close to the team at FocusRight. You're you work with a number of uh startups in addition to having your own. You were at the Focus Right uh New York event just recently. I unfortunately couldn't attend at the last minute because of that um tragic plane accident. Uh so my flight was moved from Monday to Wednesday, and the conference was on Tuesday, so that didn't work. But I managed to actually watch all of the recordings. And for me, I love going to conferences mainly to learn, secondly for networking, because that's always been my joy of going to conferences, is that I've always been trying to figure out what I can learn and I can apply in uh real-world business situations. And I saw Michael Coletta do a fantastic opening session about how you know the usage of AI for trip planning has gone from like 30 to 40 to like 56% in like the last three quarters. Um, but he was also highlighting that Google is still performing well with because people are still using search engines as well. And so we haven't seen a drop-off yet in the OTAs, which is a world that you know really well. So I just wanted to get a sense, you know, as we're having this conversation in uh kind of the middle of 2026 about AAgenTix uh AI's impact on travel, where do you think it's having the impact today on a the user journey in terms of discovery and booking travel?

SPEAKER_04 20:20

I mean, uh it is having a massive impact on on user, yeah, on the discovery side, but I don't know if it shows for a lot of people. Mainly because Google, you know, Google a year ago said they were going all in on AI mode. It was last April actually they said that, and they haven't, right? It's still a tab there. They've not forced us down that path. It's it seems that we've all as humans accepted that the current version of Google is fine. It's got the summary at the top and then the results down below. That might be the end state. Maybe we'll never go to AI mode. Humans are very slow to adapt to new, and if it works, it works. And I think Google in the current state does work. But still, the more people that try direct chatbots, uh Chat GPT, or or everything else, they they don't go back. They're doing more voice, they're doing more chat. It's really sticky. So maybe it's I think the end position is the same. We will end up on chatbots or similar, but it takes the humans, it's taken the humans a long time to move. I think it's a good thing.

SPEAKER_01 21:16

Yeah, well, even on that point, yeah, because I I talk to a lot of people who track their their AI traffic, if you will, and even Mike's research showed this. The the traffic and the usage is doubling. Like every every week, every month. It's it's growing exponentially. Now it is on a small base. So when you look at the universe of traffic of usage, Google still has the lines share. But I I can't imagine that there has to be a future where ChatGPT is the the front runner at the moment. They're just they're growing so fast. So aren't we getting to the point soon where it flips and ChatGPT and other chatbots are the place people go to get their answers and to start booking travel?

SPEAKER_03 21:57

Well, so just on that, Tani, I'm gonna ask you a follow up on that. And then Gilad, please jump in on this because what I wanted to get a sense from your vantage point, and you obviously were at the conference as well. You're working with quite a few clients that are trying to implement agentic AI. Where do you see it having the biggest impact right now? And then Gilad, I want to bring you in on where you think it's actually going to cause the most disruption longer term. But what are you seeing right here, right now, Tani, with your clients and the implementation of this technology?

SPEAKER_01 22:24

Do you mean for search? Because you know, there's a whole there's a whole uh cottage industry, might be even bigger than that now, of trying to become a top source inside an AI chatbot. So AI visibility, they had a couple of sessions, SEO, AEO, GEO, there's all these acronyms now. But basically the thing is you you want to, your brand, you need your brand to show up inside these chatbots. And there, I mean, some of the traditional methods still apply, but there's all these different things you have to start thinking about. Like how does an LLM find your information? Why are they going to source you in an answer in a query? And then how are they going to position you? And I I don't know that there are there's nothing set in stone. This is such a new approach. And Gillard might have some thoughts on on this. But I do think people have to start paying attention and travel companies in particular, travel suppliers, because we know travelers are going to chat ChatGPT and asking, like, where should I go? What should I do when I'm there? What hotels are good? What what what places next to this thing I would like to visit? And you need to have a plan for that because if you don't, you won't show up. And yeah, you might still be getting some of the clicks from those blue links at Google, but that's not going to last forever.

SPEAKER_00 23:38

Yeah, absolutely. And they'll certainly echo that most of the hospitality brands that I spend time with are all seeing the number of qualified leads from Google go down each month. Sonny said it's a small base, so the discrete, the decrease is discrete, but it will continue to go. Now, the only thing I want to add to the conversation we had about the planning process, I thought all the points are really good, is that we just need to really break it down into more detail. The way people are dreaming and discovering at the very top of the planning process is different from the way they get verification answers near the bottom, and it's different from where they book. So today, I actually think that agentic AI has had relatively little impact on the way we book travel, even though it has had great impact on kind of the very top of the funnel, kind of how we dream. And then in those confirmatory steps. So does this hotel have a pool? Are there adjoining rooms? Like those kinds of things are really perfect for AI. And one of the things that we heard Expedia admit at Focus Right last year is that they test tested AI, full AI funnels, and then they tested traditional funnels with specific AI use cases within, and those outperform. So the idea and I was make the analogy of if I've already selected your hotel, you probably should not have a restaurant selection experience available to me because some of the time I'm gonna decide that the best restaurants are around someone else's hotel. So the whole point is when someone's already ready to make a booking, we actually want to get a lot of this AI stuff out of the way and we want to give them a place to make the booking and maybe reintroduce it later on. Now, to the second piece of your question in terms of like where I think there'll be great impact. So, first and foremost, I actually think that like from a macro perspective, the funnel will have less changes than we think. Like the players will change, right? The people who have the power will change. But I think at the end of the day, the funnel is really about consumer decision making. So I think the rough outline of the funnel will likely look the same, even though I completely agree that ChatGPT and Perplexity will replace people like TripAdvisor and others at the top of the funnel. But I think the human behavior is likely to stay the same. The two places that I'm most excited about agenda make an impact, one is back of house. At the end of the day, in our industry, we have taken a lot of employees who want to be hospitality professionals and we make them clerks. We put them with uh keyboards and with checklists, and we take away the hospitality. And it's a completely different job. And I think at the back of the house, we have a win-win experience and we can create experiences that are better for our guests, but actually much better for the staff who works in our industry to make it a more appealing place to work. And the second thing is all on product creation, right? I like to say that we've been selling dynamic packages for 20 years that all reality are not very dynamic. I think there's a lot we can do with AI and what's specifically with agentic AI to really begin to create experiences that are highly personalized and connect more elements in the city. I always use the drop ship example of getting Uber to drop off food from a restaurant that's not in the main tourist area and get them to engage if it's the right thing for the customer. I think it's really about back of house and about product creation.

SPEAKER_03 26:34

Yeah, it's really interesting. I certainly agree with you about the you know operational efficiency is CRM systems and making sure that um you can utilize this technology to power the back office. Now, a lot of travel advisors are concerned about this technology, and understandably so to a certain extent, but I just wanted to uh turn a uh a presentation that you and I each gave into a question here for Tani and Christian, because you had talked about to the Virtuoso community about generative AI being a tool to accomplish your goals. And it was a great presentation. And I know the number of people that loved your keynote and it made them feel much more comfortable with AI and put them at ease. As I was talking to them about agentic AI, I was pointing out examples and I was using the example of chatbots, and there was this graph that David and the team from Virtuoso had shared where 29% of the preferred partners at Virtuoso were using chatbots or virtual assistants. And I my my concern was actually this is great technology for hospitality, to your point, about um uh and certainly for hotels. Companies like GuestOS is a good example, like and I mentioned them because they uh I've used them as a demo on stage because we used uh their platform for our AI Summit. And when it when you look at hotels or you look at airlines, it's a great use of that technology, but you do not want to train your high net worth travel clients to use a chatbot instead of speaking to a human. And so uh the crowd definitely enjoyed that part because it was just like this is where your human value, the human touch, everything you know about what's so special about Virtuoso stays relevant now more so than ever. And so when I think but here's here's the question I wanted to relate to Tani and Christian. You hear about chatbots, and what I'm talking about we're talking about today is really more co-pilot and how you can leverage agentic AI to accomplish tasks and goals on your behalf. So putting the chatbot aside for a moment, given that it would have use cases in some industries more so than others within travel, when we're thinking about how you can harness the power of agencai and what it can be used to uh accomplish, that's what I'm I'm really curious to get both of your perspectives on from really like so Tani, it's like what defines a true agent and what capabilities should companies be thinking about about how they can leverage this technology? And Christian, I'm so keen to get your point on that as well. It's like, what does an AI-powered travel agent look like in practice today? But Tani, maybe let's start with you.

Where Humans Still Win

SPEAKER_01 28:48

Oh man, I was gonna pass that one to Christian. I'm gonna take Christian's I'm taking Christian's question because I actually work with travel advisors more. And when you you asked what makes a an AI-powered travel advisor, it's someone who applies AI to the back office. Like Gillard was saying, he's really interested in applications for that. And a travel advisor, what makes them special is them and their experience and their personality, their taste, their judgment. And that's something AI, I don't think, yet can replace. So they need to own that. And that's the thing that that's going to appeal to their customers, the people they're advising. They they want to work with the travel advisor for them. And so they should focus on AI for the other stuff that the client might not necessarily see. So running their back office, doing research, maybe setting up the itineraries. And, you know, there's all this, there's all this hype around voice agents and avatars, and you know, you you can like basically clone yourself, but who loves to watch a video of an AI avatar? I mean, at some point it might get really good and convincing, but right now you want to you want to meet someone and look them in the eye. That's why conferences and meetings and events are so popular now. Is I think people are really craving that human connection and being able to see people face to face. And so travel advisors should lean into that. That's what makes them special. Sorry, Christian, I stole your question.

SPEAKER_04 30:11

No, I can I can I can I can adapt. I think I think humans need to be, I think we need to be more honest. We're we're we're gonna fight the next whatever it is, 10, 20 years because the agents are here, the the the robots are here, they're better than us at most. I think we need to be honest at the things that the robot and the AI is better than us at. But it's not everything. Um but but we're gonna fight this all the way. The the developers are still fighting, it's oh the AI can't quite do the code yet, and it's it can. People are still bringing up the Canadian Airlines thing where it gave someone a$200 refund four years ago as some kind of evidence that we can't do voice chatbots or whatever. It's it's it's ridiculous. The the AI agents are better than humans in 99% of cases. I'm making numbers up, whatever, but but we need to be honest. But when when humans weren't humans, which is I think is a lot of the time, we're we're better just because we want to talk to humans, we want to see humans. We've just got to be honest at what we're better at and when we should hand off to the chatbot. So use the tools where they can make us more efficient, where it saves time, where the client's fine with it, and then use ourselves for the times where we can add value, speaking to a client, making them feel reassured that everything's gonna be okay and we'll make the changes for them. But I think we need to stop fighting and pretending that we're better at things that we're just clearly not.

SPEAKER_01 31:29

Can I tell a story here, Dan?

SPEAKER_04 31:31

Go, please.

SPEAKER_01 31:32

So I I just moved and so I updated my my house and car insurance policy, and it's not travel, but I think this would apply just as much for travel insurance, if you will. And I'm using a platform not to be named. The in the beginning, that the user experience was amazing, and I was talking to their AI agent, and it's I think it's like 99% agent-driven. And you go through this chat uh experience where you give all your information, it works on the policy, it delivers it, and then it kind of walks you through how you're all set up and they have a lovely app, and it's like, wow, this was so amazing. And I compare that to working with something like State Farm, where you have to find a local agent and they might call you back in three days and maybe not call you back, and then you've got to talk to them for an hour on the phone, and it's just a hassle, right? This the initial experience was awesome. And then I had an issue where they wanted to send me this device I had to put in the car and it doesn't fit in my car. And so now I'm stuck in this loop with this agent who keeps telling me to do things that don't exist, and I'm so frustrated, I'm ready to quit. And I'm like, to me, this is an example of the agents are amazing until they're not. And as a travel company, if you're building these things, you need to think about what that journey looks like for the end user to make sure that when the agent starts to fail, there's a seamless handoff. Because what's happening is those hallucin hallucinations that we know, Gen AI is really, it does a lot. It's coming up in this experience. And now it's starting to damage their brand because I'm I it's not good anymore. So I do think, like to Christian's point, the agent can take over a lot, but it's not everything. And that was a great example of a fail.

SPEAKER_00 33:06

Yeah, if I could just add two quick thoughts. First, I loved what Christian said a minute ago about knowing what we're good at and what the AI is good at. There's only one thing I want to add to that, which my view is that AI is coming for everything, except for those things that we choose to keep in the human domain. And I think that's a really important question for all of us to be asking as an industry, but really as humans, as society, what are those things that we want to keep in the human realm, even if AI can do them as well or better than us, that we really want to keep in that realm? And this, I think, is a perfect connection to what you talked about with travel advisors. You know, when people ask me, do I think the profession of a travel advisor is safe? To me, that is an easy, obvious yes. That profession will be around forever. And the thing is, it's because the thing that a travel advisor provides, a good one, is human connection, it's good taste, and it's vibes and experiences that you want to have with another human in the experience. Now, the analogy I would use though is that while the travel advisory profession is safe, individual travel advisors are not. I use the analogy of if you walk into your accountant's office and there's not a calculator on the accountant's desk, you should probably turn around and walk out. Why would you ever use an accountant who's not using the world's best tools to do the accounting job? I think the exact same analogy applies to travel advisors. I love my human travel advisors, but I expect them to be using the world's best tools to do the best possible job for me and my kids.

SPEAKER_03 34:26

Well, so I got a couple things to add in there, and then I think this is gonna uh uh uh be a fascinating uh topic to discuss. So the one of the things I just want to uh bring us all together on, and one uh that with this Agentic AI series and this being our very first episode, one of the things I want us together is to cut through the noise that we're hearing, and you guys have each brought up a couple of valid points, uh, which is the fear-mongering on both sides. One, you highlighted uh Christian just about the the Air Canada example, and Tani, obviously, you've heard number there's a lot of people that are concerned about their data and privacy and or their companies not allowing them to use it. And so there's certain people that are continuing to be Luddites on this topic because they're, you know, they're they're not embracing it because they're they're listening to those signals, and then you're also hearing people on the other side saying, Oh, you're gonna get so far ahead, you're never gonna be able to catch up. And and so there seems to be two different extremes of vantage points here. And the reality is neither um are the true reality. And the one of the things I finished my presentations on uh most recently is I uh I emphasize that in 2026, it's not a question of whether you use AI, but it's how you use it. And I think this is the difficult thing that we at the moment we find ourselves in, is because in 2026, this is very different than 2001, uh, with the dot-com era, where you knew how who you had to go to for a website or for digital marketing or for social media, and the consumer journey was changing, yes, but you could find partners to work with. We're now at a point, and the the reason I wanted to have the three of you here, we have a top 50 list of the most uh knowledgeable people in traveling AI, and you guys are in the top 10, if not the top five. And so people can find your profiles. And right now, what I encourage people to do is to follow each of you. Go and see what Tawney's doing, listen to Christian, see Gilad's presentations, and they're trying to gather that information from hearing what you're saying. But this is where I find the challenges that we start talking about chatbots, and then you find someone using a chatbot not for the right purpose because it's like, wait a second, you're losing the human touch. You're passing your customer off to a chat bot when you should be the first line of you as soon as they hear your human voice and make that connection. But it's very difficult for companies to wrap their heads around how they should be using this technology. And so that's why I've, you know, I'm I'm I'm hitting that point home because people listening to this are looking for that direction and advice from us for how they should be embracing these tools. Tawny, you mentioned right at the beginning, one of the most important things here in 2026 is visibility and making sure that just like in the early days of SEO, that your business is relevant. But when it comes to actually taking a platform like Claude Cowork, so many of these travel advisors ask me, what are the use cases for it? And the number of travel advisors that came up after my presentation and said, Oh, I've been using it for the last six weeks. I've stopped using ChatGPT, I do everything on Claude, it's totally changed my world. And I say, send me some examples of that so I can showcase them to others of how you are leveraging this technology. You know, you hear people like Travis Pittman from Tour Radar that you guys all know. He now uses Cloud Cowork exclusively to build all of his presentations for his board meetings. It used to take him four or five days. Now, by virtue of having access to Confluence, Slack, his emails, building up the skills in the platform, he's like in 45 minutes, I can have 75% done, and I just finesse it. And so you keep hearing all these game-changing, but you're you're hearing it from people that are trying to just apply it in their companies. And so I think we're just at this interesting stage right now where it's happening so fast, and there's obviously a risk of people being left behind. Tani and Christian, love to get your views here and like as far as some of the real world applications you're seeing and what some of the recommendations you would have for people listening to this, trying to wrap their head around where should I start?

SPEAKER_01 38:00

I should let Christian go first here because I stole his thunder last time. So Christian, you go.

SPEAKER_04 38:04

I don't know if you I don't know if you did, but I because I I think of different weird stuff sometimes. But yeah, I think I think everyone knows now. There's been a revolution in the last what, four months with clawed code and then clawed co-work. It it has taken things a step up. The the the future, the immediate future for the next year or two is clawed cowork or similar, right? OpenAI is talking about a super app, a super operating system. I don't know what they're doing, they need to get a move on, but we're all gonna have these systems where we've got our docs on the side, we've got all these tools, and it just does things. So my my my I guess my main thing is people just need to dive in. If you're still using a AI and you're just on Chat GPT asking it a question, getting a response, asking another question, you're at you're at level one. We we did that three years ago. That's fine, it's better than nothing, but you need to be using things like Claude Cowork. Cloud cowork's so accessible, it's a chatbot that talks to your files. You just do simple things like go and change this thing in all of my files, and it'll go and change the year from this to this in 250 files, or go take these 20 Excel sheets of customer lists and combine them into one, and take these folders of invoices and match them all up with names and save them again. It just does things that you that you want. And the the best way to use them is to ask it how to use it. Ask the chat bot, just tell it about your company and ask it what kind of tools you could use and how you could be more efficient today. You've got you've got someone by your side all day long that's not gonna judge you. Maybe it's not gonna judge you, maybe it does, but you can ask it questions and it can tell you what to use, how to use it, how to better your life. So just jump in and ask it questions and and kind of get there.

SPEAKER_00 39:42

Our friend Jeanette Roush always says that having a Chat GPT or a cloud subscription is like having an 80% McKinsey sitting next to you all day long helping you. And I think what Christian said is it's exactly right.

SPEAKER_01 39:54

And I I was going to acknowledge that it's for someone who's just used to that back and forth, using something like Claude Cowork, it it can be tricky. It is difficult. There's some terminology in the platform that's hard to understand. And Christian's point about asking Claude for help is amazing. It's it's like a tutor right there with you. I will say, if you can, there's some guides out there that will help you. I actually have one and I'm happy to send it over, Dan, and you can include it in the show notes. It's a it helps you with the setup. So all the steps that you need to get started and understanding what the platform looks like and how how to operate it is a once you have that foundation, it does get much easier to use after that.

SPEAKER_03 40:34

Well, I'd love for you to share a few of those. I know your time uh together is a little bit more limited. Everyone, this is amazing we managed to get everyone together for this recording. I could have an in a conversation with each of you individually, but I thought it would be that much more dynamic and interesting for our listeners to have you all together. So Tani's gonna have to have to head to the airport first, then Gilad's heading to the airport shortly afterwards for an around-the-world trip, um, which is why this is so special and valuable to get this time together. But knowing that our time is gonna be a little bit more limited, Tani, I want to stay with you for another minute to actually share some of those examples. So for practical advice, when it comes to someone, I've kind of been framing it recently as kind of beginner, intermediate, advanced, and it's always hard to find a level. What I've typically found is that people are somewhere between like high beginner and low intermediate. Like it's like it's very few are intermediate, medium, and advanced. So everyone's used Chat GPT or they're familiar with these, and they're but they don't necessarily aren't very sophisticated even at prompting, or you know, they're not familiar with how to use different platforms. And so you kind of realize very quickly that you know you are dealing with, you know, still very much dealing with beginners. And so, what would be some of the advice when it comes to agentic AI? What are some of the recommendations, first off, you're giving to clients to get started? I think that would be super helpful for people listening to this to know. And then we'll make sure we follow up with your recommendations so people can find that as well.

SPEAKER_01 41:53

Yeah. I have a foundation or a framework I teach, and it's the three F's it's foundation, function, and flow. And the reason I teach it in those three steps is you need to have the foundation, and that's really it's simply being organized. You as a person need to have all of your folders and files in an easy, simple layout so the AI agent can access them and find them and it knows what it's looking at. The function piece is making sure that you have the platform set up. Is it connected to all the other platforms that you use? To give you an example of how I use this, is I have my Cloud Cowork run a daily briefing for me that comes to my Slack every morning. And to make that valuable for me, it needs access to my email, my calendar, my Notion database with all of my to-dos. It also needs to be connected to Slack. And I give it a really detailed prompt and then it goes off and it looks at all the things that are happening to me on that day and then delivers me this briefing in Slack every morning. Now you could do the same thing for your travel. You, I mean, you there's all this information that exists in your ecosystem. And if you give the agent, co-work, called co-work, access to this, and this is what I call flow, is when it just goes off and it does these things for you. But you can't have that flow piece unless you do the two prior steps, setting everything up and making sure everything's connected.

SPEAKER_03 43:11

And just as a follow-up to that, because I think that's going to be valuable, especially for individuals that are looking for how they can use it, the example of Travis or many of these travel advisors. When you're dealing with companies, I know you now uh work with one of the companies I collaborate with, Blaxford and you know, premium RVs, and that they're one of the ones I mentioned that loves working with you. When you're coming into a company like that, certainly there's a benefit to um David learning how to use it for his own benefit to manage his business. But then as it when it comes to the practical applications for their businesses, you mentioned optimization for um to be discoverable being one of them. But what are some of the best use cases? Because I think that would be helpful for us for yeah, for for travel businesses to be able to apply that technology. Is it like the back end systems, as Gilad mentioned? Um, what what are some of the things you initially say, oh, actually, here's where you can use it right away. And it will make a real difference.

SPEAKER_01 44:02

I think for for travel companies in particular, there's two areas that are really they're a great use case. And the first one is around data analysis because travel companies have so much data. You have all of the information you're collecting from bookings where people are traveling, what they selected. For BlackFeds as an example, they have, I mean, their spreadsheets are like they are azizy because they have so many different elements. The price someone paid, the discount they received, where they went, how many nights, et cetera. And they get these spreadsheets and it's just a bunch of numbers on a page that takes someone a really long time to figure out and pass. And you know, you can build formulas in Excel. But what you can do now is you can connect this data directly to something like Claude or Claude Code or Cowork, and you can start to build these analyses that happen on a scheduled basis or whenever there's new information. That's one of the things that I've been working with them on is like, how do you automate some of that? The second thing, the second use case they're finding really helpful is market research. They have a really good presence in a couple of markets, but they obviously want to expand and move into different areas. But what are those areas and what what are the implications? How do they need to be thinking about people's desire to book RVs in those places? And using things like deep research and the some of the information they already have from what I call their ideal customer profile, Claude can go out there and start to do these research reports for them that they would probably pay a company$50,000 for. And you just let the agent run, it does the work for you, delivers back a presentation or it delivers a report. And then we were working on with David, like how do you then convert this into something that's consumable for your team, for your board of directors, for investors? And so you can take that report and you can have these agents go off and then convert it into a PowerPoint or whatever format you want. So I think those are two use cases anybody can start with today and should be using today.

SPEAKER_03 46:02

That's great. Thank you, Tani. And I'm going to come to you next, Christian, when we talk about spreadsheets. But Tani, I know you have to run. So I want to make sure that everyone knows where to find you. Obviously, taniperry.com. You've got an amazing weekly newsletter. Uh, we'll make sure those show notes are added as well. Um, but just I want to give you the last word before you sign off. Any last final recommendations or ways to connect with you after hearing this podcast?

SPEAKER_01 46:25

Come find me on LinkedIn. That's where I'm most active. I would love to meet you all. Just do a quick search for Tani Perry. There I don't think there are many Tani Perrys out there. So yeah, come say hi. I always love to hear about your journey with AI, what you're working on. Give me your hardest problem. So come connect with me on LinkedIn. Tell me the biggest thing that you're having a challenge with, because I always like a good problem to solve. So that would be great. Thank you. It's been wonderful. Great conversation.

SPEAKER_03 46:47

Thanks so much for joining us, Tani. Safe travels. We'll speak soon.

SPEAKER_01 46:50

Hi.

SPEAKER_03 46:50

We'll be right back.

Context And Knowledge Bases Matter

SPEAKER_02 46:54

Propelic is the leading full service travel marketing agency built for how people discover travel today. We help travel brands grow through SEO, paid media, analytics, and AI-driven visibility across search and generative AI platforms like ChatGPT. Recently named to the Inc. 5000 list of fastest growing companies, Propelic works with leading tour operators, hospitality brands, and destinations worldwide. If your brand cares about demand capture, performance, and winning in the next era of search, Propelic is your partner. Learn more at Propelic.com. Looking for travel that's purposeful and hassle-free? Intrepid Travel, the world's largest adventure travel company, offers over 900 locally led tours across more than 100 countries on every continent. Every journey is designed to connect travelers deeply with people and places, while they take care of the logistics so your clients can focus on the experience, not the details. With small groups and off-the-beaten path adventures, Intrepid delivers meaningful, immersive, and responsible travel. Explore the world with purpose and ease at IntrepidTravel.com. Travel has changed. But many multi-day tour operators are still stuck with spreadsheets and disconnected systems. Today's travelers expect more. More choice, more flexibility, more unforgettable moments without the friction. That's where CapTO comes in. Built on Salesforce, Captio is a modern platform designed specifically for multi-day tour operators. From tailor-made trips to group tours, from rail to river to cruise, we bring your itineraries, bookings, payments, and operations together in one seamless system. With automation, real-time inventory, and supplier integrations, we make complex travel simple. So you can focus on what you do best. Serve travelers, not spreadsheets. Visit Captio.com to learn more. And now, back to the show.

SPEAKER_03 48:50

Christian, let's talk spreadsheets, man. That's your business in some extent. Content management. So yeah, tell us what you're seeing. And Gilad, this is where and please jump in as well. Some of the use cases you guys are seeing that I think we can point people towards. And you also don't suffer fools gladly. You call it nonsense when you see it, and that's what I appreciate about you. Give us the real deal.

SPEAKER_04 49:09

Speaking of LinkedIn, all the AI slop nonsense on LinkedIn winds winds me of all these long posts of it's not this, it's that, and I'm like, stop. But no, that my word of the year last year was data compression. My word of the year this year, from here on out, is context. I kind of like those words because they sound really boring, but they're really relevant now. So context is what Tani's talking about. Having all your data. People don't know what data means. They don't know what having your data in one place. But when I talk about it, context is having your company information, what you do, what people's job roles are, how your company works, having that in one place enables you to build all of this stuff. We just re-released Magpie 36 hours ago. And it mostly written, well, it's all written on on AI. And for me, the best thing about it is it documents. Everyone knows engineers are supposed to document their work. I don't think any actually do. And the idea of that is for the next engineer can come back and read it, which they're never gonna do. But now that AI can, and when it builds the code, it documents it, it documents it for the for itself, but it also for the humans. So now I've got pages and pages of descriptions on every page, every feature, every little tool within Magpie that I can use so we remember what it does, and just to create user docs. So now I've got this knowledge base of everything in Magpie, and I can do one prompt and say, write user docs for all of my customers to do the top 10 things they do. Four seconds later, I've got everything in one place. So that's the most important thing. There was a there was a great essay. I don't know if you guys read it, the Jack Dorsey essay this week. And he talked about you should read it. He talks about the way organizations are structured, right? And he goes back to the Roman Empire because that's what you're supposed to do, and we have all these hierarchies because it's all about information, right? Passing information from the top to the bottom. So the emperor is trying to invade England, and it takes 10 layers to get the information down to the bottom before people start marching. And then when they start marching, oh we've hit a problem, there's a there's a there's a channel, and information has to go back up to the top. And we've built all these structures around that information transfer, and now that information transfer is gone, everyone has access to all the information all the time if we allow them. So company structures can completely change today. If you're actually AI first, which nobody is, you could revisit how you organize your company. And now is a great time to do that. If you're not too big to move, now and I'd start with reading this essay and just makes you think about how you don't need to have the same structure you've always had. You can actually do very different things now because the information is no longer contained by layers of people passing it on to the next person.

SPEAKER_00 51:45

That's awesome. I mean, the first thing I'll just underline is what Christian said is one great use for AI is to do things you didn't have time to do before. We see this all the time in the travel industry. To use another example from travel agencies, the best travel agencies send recipes and songs before you go to a country. They send narratives and other kind of content to really help you remember and prolong the experience after. But we all know we're just too darn busy to do a lot of these things. So, one great use for AI is to do things you've always wanted to do that you've never had time to. And it's also psychologically an easy place to start because you don't feel like you're replacing your own job or things you're doing, because by definition, these are things that you never had time to actually do. The other similar context that I help people think about where to adopt AI is I encourage people to create a two by two, you know, kind of a little consulting chart where you plot at the top things you love, at the bottom things you don't love, on the right, things you're good at, on the left, things you're bad at. And you actually begin to plot the key elements of your job. And what you notice is there's some things in each of the quadrant. And I always tell people the things in the top right, the things you both love and are good at, leave it alone. The things on the bottom left, the things you both dislike and you're not great at, that is the perfect place to start deploying AI. You have the dual motivation. You don't like it, so you want AI to automate it for you, and you're not great at it. So you have a reason to believe that AI can lead to better results. So that's specifically how I think about it. And then just lastly, a couple things for the travel, travel businesses specifically. Um, Christian, use the word context. I completely agree. I really think the brands need to focus on deeper engagement with your guests. Page counts probably never mattered as much as we thought they did. They certainly don't matter at all anymore. So it's not about how many pages or how many minutes. It's about the depth of the engagement, it's about the context you can collect from that engagement. And similar, the other side of that same coin is really advising business to focus on share of wallets, not just a number of transactions, but how big of a share of wallet can you grab and can use AI and AI agents to begin to create these dynamic offerings that give you a better share of wallet for you and a better, more personalized experience for the guest, which is the win-win outcome. I think AI is here to help us achieve.

Real Risks And How To Mitigate

SPEAKER_03 53:51

That's great. Thanks, Gilad. One of the things that you, as we look at the opportunities that are very real and meaningful, and that's where I wanted to get a sense from both of you, and also uh make sure we're leaving our listeners with practical, actionable advice, and they can actually, and and we will do some more blog posts on this. But one of the things I want to make sure we do cover together is some of the risks that are real and valid. And so uh to the point about pointing to examples of where it can hallucinate or where it can uh where it's not perfect, um, is an easy way to cop out of doing anything. Because when, you know, there there's this is another major moment of uh change management in most organizations. And so I use TUE as the example because they were one of the most impressive presentations at last year's ITB on how they had embraced AI within their organization. Because you hear many companies talk about the fact that they won't allow their company, their employees to use these platforms because they're worried about their proprietary data or personally identifiable information or risks that could be associated legal and otherwise, especially for larger companies. But what they need to do, and what some of the most successful companies have done, uh, is build their own LLMs for their own organization that and they brand them and all their team members use them. And so one of the things that's always stood out to me, and I learned this from them, but I, you know, find a champion, get everyone involved, and then also find the right partners to work with. That's sort of my recommendations. But but people are still bringing up the risks involved. And what I want to ask you both is what's real in terms of risk that you do need to manage and be aware of? Um, and what are some of the things that are just noise and just getting in the way? So maybe Gilatsa, stay with you for a minute because what do you think in terms of the risks around utilizing AI that you need to just you know mitigate against?

SPEAKER_00 55:29

Yeah, absolutely. So the number one thing I hear about is wrongness, hallucinations, misinformation, etc. I will say the number of hallucinations, the proportion of hallucinations coming out of the flagship models is down significantly over the last two years. It's not perfect. I'm a huge believer in human in the loop. I'm not saying that it's ready for prime time on its own, but I do think that some people have these ideas in their mind that might be one or two years old and the models really have gotten better. And this is also where there's different kinds of AI, different kinds of AI tools that really force the model to stick much more closely to the core content. So things like notebook LM. So I hear a lot about wrongness. I think it's something we should still be aware of, we should still pay attention to, especially in premium and luxury part of the market. But I don't think it's to the same extent that maybe it was a couple of years ago. Similarly, I hear a lot of people on the business side of the house saying that AI is not 100% there, so I'm not gonna use it. And I always feel like that is such a greedy mentality that someone has. Like if I said I'll do 70% of your job, you're gonna say no, thank you, or not doing 100%, I don't want your help. That seems kind of ridiculous. So my whole thing is too, is like just because it's not 100% of the way there does not mean it could not add immense value to what you're working on. And as I said already, I'm a believer in the human of the loop. So we actually don't need it to be at 100% because you're still engaged in the thought process. And then the last one I'll say is loss of the human knowledge of our businesses. At the end of the day, there's a lot of knowledge in our businesses that is not critical. You know, little things about the way information flows or about specific product works, et cetera. But what's really important is that as we adopt these AI agents and these AI solutions, we don't lose a picture of the macro view of what we're actually trying to do here and how each of the components adds together into that macro view. So my whole thing is I'm all about introducing AI everywhere, as long as we keep the human nature of our world, the human nature of our businesses. And don't forget that our job as humans is to be critical thinkers using this new technology.

SPEAKER_03 57:24

Thanks for that, Gillad. And Christian, you're obviously of the era of crossing the chasm. I think it was probably like the um one of the probably one of the first tech investors, especially given where you're based. And Jeffrey Moore, I might have said the landmark book going all the way back to 91, but it's still relevant today, especially now when we're looking at embracing new technology. The people who are the innovators and the early adopters, the ones that are currently using OpenClaw, for example, and uh and those are the people that are also willing to accept some of the risks associated with the unknown elements of giving it access to all of your systems. And sometimes things can go haywire, and other people say, no, I'm not quite ready for that. So those are the innovators, those are the people that are early adopters. Um, but Christian, I'm keen to get your take on the risks and the challenges that those people are willing to just you know uh push forward with. But how do we make sure that we do cross the chasm and bring the majority of uh travelers on this journey and people in the travel industry that actually to address the risks that are real? So that late majority, the early majority where the businesses really shift and change more quickly now with AI than I think when we saw with uh the dot-com revolution, it's gonna move much more. But what are some of those risks that are real that we should be worried about but still move forward?

SPEAKER_04 58:38

Yeah, there are real risks. There's there's bigger societal risks, which we haven't got time to get into. I I love Glad's second point when because I sell software, right? And when you have a new solution, a lot of people's first thing is, oh, that great, this is gonna solve everything, every problem I've ever had. And when you tell them it's only 90%, they're like, oh, that doesn't really work then. Uh there's there's no single tool that solves everything. Chat GPT, just as a chatbot, is gonna help you. It's not gonna do it. And then people they they find a hallucination or they talk about the Canadian Airlines thing or the one person that got lost in Tasmania one time, and out of billions of searches, these and they they do hallucinate. So if you're sending information out to clients, you do need to verify it. You should double check everything. If you're just gonna send them a list of waterfalls that are critical to their trip, make sure those things exist. So it is still something you have to verify, but humans make mistakes as well, and you'd verify human communications if it's important. Second part, I think people get hung up on it's gonna steal my data and it OpenAI is a trillion dollar company today. They don't care about your list of that first of all, they're never gonna train on things like emails or names or anything like that. They don't care about your lists of clients and your list of very special itinerary information that you've been building over time. They they don't. They've got they know pretty much everything in the world. They have the world's data. Having said that, there are ri there are risks. And if you're especially a big company and there's some liability there, ask your lawyers if you must. But ask Chat GPT what not to do and what not to upload. I think you misspoke when you said we're gonna build an LLM. No one's building an LLM, right? That's that's a job for the trillion dollar company. But yeah, a fine-tuned version of it of an LLM, which is like a GPT or something that you've kind of got your information off on the side, that's what every company should have. So you're not uploading all your information every time. But there are there are risks, but yeah, it gets overblown. People get overexcited about this thing wanting to steal my data or leaking email information. It's it's never gonna, I shouldn't say never, it's not likely to do that.

SPEAKER_00 1:00:42

Yeah, and if I can add a thought here, this is something that I've worked a lot at as board as a board member in many companies and as an advisor to some boards. And I think it's really important for leaders to think about net new risk, not overall AI risks. So Christian used a perfect example. Humans make mistakes too, not just AI. So if you think about a company that has hundreds of humans on the call as agents, as salespeople, as whatever, and then you think about introducing an AI, well, leadership says, what if the AI says the wrong thing? Well, my response is what if your hundreds of employees working from home say the wrong thing? So I think it's really important for businesses not to think about AI risk, to think about net new risk that you're introducing into the system when you adopt AI. Sometimes you're adding risk, and you should be aware of that. Sometimes you're actually reducing the net level of risk by introducing AI, even though the AI, of course, has some own, some of its own risks it's introducing into it as well. That's it.

What Leaders Should Do Next]

SPEAKER_04 1:01:33

I think the big the biggest risk is is is our fault. This thing generates so much content so quickly, right? We all know that. If you have a session where you're asking questions for half an hour, you could end up with 50 pages of stuff. And there's a tendency to think, oh, that was 99% good, so let me send that out. And a lot of it's just us not verifying it, stuff that we know we should we should verify because it creates so much content so quickly. You turn a chat button and all of a sudden it's having thousands of conversations, so you just let it go and you forget that it was only 99% accurate in the first place, but you got lazy and you didn't fix that last 1%.

SPEAKER_03 1:02:09

I want to make sure that for our listeners, when they're trying to prioritize the next 12 months in their business and how they should be embracing AI, agentic AI specifically. I'm keen to get each of your take on this. Obviously, we've talked about some of their use cases, we've talked about some of the opportunities that you see. And I'm asking you specifically to speak to leaders now because as we know, change happens from the top and from the bottom, and and you have people that we're encouraging inside their organizations. I spoke to someone just recently who works for a large insurance company. I won't call out the specific company, but most people would know who they are, and highlighted just how frustrated they are that the entire marketing team doesn't use AI. And so there's organizations that are struggling to figure out uh how they should be utilizing it. Um and when you speak to owners or you speak to um executives, most of them are keen to figure out how they break through. And a lot of them are now, it's the expectation to keep their CEO role or to keep their executive position. They need to be figuring out how their business can become more efficient, uh, generate more revenue. They're all trying to upskill themselves on AI. So for each of you, given your senior experience in the industry and your uh deep knowledge in AI and of AI and where we are, what do you think leaders should be prioritizing? Christian, why don't we start with you? What do you think industry leaders and travel should be prioritizing as it relates to AI in the next year?

SPEAKER_04 1:03:29

Yeah, I'm gonna go back and count on the amount of times you've called you've called us like senior or like something that refers to I'll edit those out.

unknown 1:03:36

Yeah.

SPEAKER_03 1:03:37

I'll edit those out with AI, no problem.

SPEAKER_04 1:03:39

Yeah, that's uh I think as I think as leaders, I think you need you need to be in touch with what's possible. You need to know what's possible because I think there's a huge gap between what people think as they've heard about what's possible and what's happening today. I think of code because I think code's the biggest gap. I know most people are not writing code, but that's just an example of millions of engineers sitting out there still writing code today where it's completely unnecessary. They shouldn't be doing that. But it's gonna be the same with cowork next. So I think leaders need to understand what's possible. They need to be on the cutting edge. And then then, yeah, if you've got a big team, it this this stuff's really disruptive. I and I asked Tony this a couple of weeks ago because she's speaking to a say a marketing director or VP of marketing. I said, How many of those people are actually just trying to get the tools because they want to be the last person standing? And that's that's that's scary, but that's real. There's teams of 20 marketing people looking at this thing thinking, we only need one person in a couple of years. That needs to be me, because I've got a mortgage and I've got a this and we all have bills and mortgages and rent and all that stuff. So I but I think as a leader, you need to be honest, you need to because you need to be able to see the future. None of us can see the future. We can't see three months ahead, but you've got to see as far as you can. You need to adopt the tools. Four tools, I I I don't care if people use Gemini, Chat GPT, whatever, but the the concept of Claude co-work, the concept of having this set of files over here that you're talking to, you need to learn that. Because whether that's whether Claude's winning now, whether OpenAI is the next thing is the best thing since Slides Bread, it doesn't matter, Gemini will have the same. That concept of working is gonna be here for the future. I'm convinced of that for at least a few years. So I would get your employees on a system like a Claude co-work, where you're working with the files and with others, and that becomes your sort of background system of running the company.

SPEAKER_00 1:05:28

Yeah, absolutely. That's great advice. I would just add two things from a slightly different perspective. Um, really thinking of my role as a CEO coach and as a board member. So, first is leaders need to create cultures of innovation within your company. At the end of the day, the people are the hard parts. People don't like to change the way they work, they don't want to take unnecessary risk. Most people are not the entrepreneur crazy people, right? So, first and foremost, you need to create a culture of innovation within your company because innovation, as we said, comes from everywhere, it comes from the top, the bottom, and the sides. So, you need to have a culture where everyone knows that they're allowed and even encouraged. innovate. And one really important point to building a culture of innovation is you have to ensure that in your company, people are rewarded more for taking a risk than they are punished for taking a risk that doesn't work out. At the end of the day, if people think they're going to get punished for trying something new, more they get rewarded for trying something new, they will never be innovation in your company. So first and foremost is you have to build a culture of innovation. And right now, because everything is moving so quickly, this matters more than it ever mattered before. The second piece, which is related, is it's the job of executives and leaders to create mental models. Mental models are sh are fancy terms for shortcuts to thinking, creating mental models for when and how to deploy AI within your company. And someone mentioned this early in the conversation this is not going to be the same for every company. For every brand and for every company between your brand, your customers, your service, your your risk profile, there'll be a different way you want to deploy AI. And the way to do this safely and the way to enable innovation across the organization is for leaders to create these mental models and to spread them across the organization. So let me give you a one concrete example. In a hotel you may say the the question of when do we deploy AI, you may say that our mental model is we deploy AI when the speed of a response outweighs the human touch of a response. So what does that look like? If someone is asking your concierge about the the uh tipping culture in your country or the dress code at that temple up the street, they're probably going to the bank or packing their bag and you want the answer right now. If someone's asking a concierge where do I propose to my girlfriend or where do I do the family photo shoot, they're probably very happy to have the concierge wake up tomorrow morning and give a really thoughtful human touch, human creative answer. So for one brand they could say our mental model is when the speed of a response outweighs human touch, we deploy AI. Like I said, for each company this will be different but it's your job as leaders and as founders to create these mental models, to spread them across the organization and by doing so to enable everyone in your team to innovate with AI.

SPEAKER_03 1:07:48

That's awesome Gilad I really appreciate your perspective there.

SPEAKER_04 1:07:50

I think Christian do you want to add something to that you're smiling I was just because I used to employ lots of concierges and I'm just picturing the concierge and someone says where shall I propose to my girlfriend and oh just do it over there. Last thing you want is oh just do it oh just do it around the corner. You want someone to think and stop and pay attention it's a good point.

SPEAKER_00 1:08:09

And tell your human story that makes you feel a certain way. You know I used to do the analogy we have these local experts at UTIP. This is before creators and influencers are really a thing and we have this James Beard award winning chef named Tom Douglas here in Seattle. And I always used to say Tom Douglas can give you a recommendation of where to have a salmon dinner tonight that's going to mean more to you than the best AI reading all the reviews and all the research in the world because Tom's a human and he's communicating to you in this human level. So absolutely I think at the end of the day there are just these things that we want the human not just to give us the right answer, but to make us feel a certain yeah that's great.

Personalization And The SMB Advantage

SPEAKER_03 1:08:43

I've so enjoyed this I want to leave everyone with one big takeaway from each of you and that is what excites you most about age and take AI in 2026 and as we look forward what do you think is the biggest opportunity or the most exciting development that you're going to be paying close attention to being either investing in or involved in and Christian I'd love to start with you and then finish off with you Gilad but yeah Christian what is what what excites you most since you are always talking about this and you're such a senior leader within the industry. So I've but tell us tell us what excites you most about where we are with uh with AI.

SPEAKER_04 1:09:20

For me today it's it's what you can build. I I love building stuff and now I'm directly building I'm in clawed code all day building tools and building processes and and it's it's what you can build. I think it's and it's a chance to learn. So my sort of last advice I think push yourself. Now's a chance everyone can push themselves one level higher than they were comfortable about. I said earlier you've got this tool now to ask questions. I don't code I spent half my day taking screenshots and saying I don't know how to configure this thing just tell me where the button is tell me which which which tab to go to and it does it all day so I can push myself to another level of something I wouldn't have been comfortable with six months ago and everyone can do that because you've got and now a teacher to help you get there. So I think push yourself deeper than you would have done before. New software's intimidating and scary and people get lost and you do get stuck but you've got someone to help you now so I think pushing yourself one more level deeper the stuff you can do now is it's incredible the last few months and I don't think most people have jumped in yet.

SPEAKER_00 1:10:19

Yeah absolutely I completely agree with that. The two small things I would say is one, I'm really excited for the next era of personalization. You know I was someone who's been talking about personalization for nearly 20 years in our industry now. And I think the things we did in the 20 teens were really cool and I'm very proud of those things. But we are just in a whole new world right now. And I'm just really excited for the opportunity to create more win-win-win outcomes, better trips for the traveler, bigger share of wallet, more profit for the travel seller and the ability to reduce the negative impacts of tourism like over tourism, environmental impact by spreading the value of tourism across many more places. I think that's something I'm really excited about for the industry. And then the second piece I want to say is I refer to this era as the golden era of small and medium business. And this is not really good for me as a venture investor but it's good for the rest of us. And we are in an industry that is largely made up of small and medium businesses. Most of the best travel companies are small and medium businesses. And it used to be that small and medium businesses were at a real systemic disadvantage compared to large businesses and their ability to deploy technology, to use data, all those kinds of things. That disadvantage is effectively over. Your company no longer is disadvantaged by the lack of engineers or the lack of data scientists on your team. I think this is an endlessly exciting time to be a small and medium business and I hope we see many more incredible businesses pop up.

SPEAKER_03 1:11:34

Yeah I'm excited as well for bullish for the companies that can now utilize this technology to be much more competitive than so I I I'm I lean towards the startups rather than the incumbents in terms of where the opportunity is but there's opportunity for both and that's probably a conversation for next time. But Gilad let's make sure if people want to connect with you what are the best ways to do so I know you're incredibly active on LinkedIn you've got your podcast as well but yeah what would be the best ways for any listeners for this to be able to connect with you afterwards.

SPEAKER_00 1:12:05

Yeah absolutely LinkedIn is absolutely the best place and I want to hear from you and I often put things out about the innovations that I'm excited about as a person as an investor if you're working on those things please do reach out.

SPEAKER_03 1:12:15

Awesome thank you and Christian same for you please obviously everyone knows you as magpie but obviously you're a you're a thought leader in this space so where should people follow your journey?

SPEAKER_04 1:12:24

Yeah same LinkedIn I'm in that a lot if you do write AI slop I will comment and call you out in front of in front and the public.

Next Guests And Closing Links

SPEAKER_03 1:12:32

But yeah LinkedIn happy to message me or put a comment yeah that's where I spend a lot of my time thank you so much for joining us on our latest episode of Travel Trends and our first in the deep dive into the world of agentic AI. I hope you enjoyed our opening episode with our good friends Gillad Berenstein, the founder of Brook Bay Capital. Great to have you back on the podcast, Gillad. Tani Perry joining us again who was part of our AI conference last year. She is truly an expert in this space as you heard today. And Christian Watts our old friend as well founder and CEO of Magpie. Now next week we're going to continue the conversation with Michelle from Mindtrip. Looking forward to bringing her back to the podcast. And then we're also going to have Brianna McNeil from Travel AI. You've very familiar most of our listeners with Travel AI so great to have them back as well. And then we're going to be joined by Benjamin Manzie, who is the chief customer officer at Maya. And Maya of course is the sponsor of this series so I was keen to have him part of the conversation but specifically because of what they have built. And just so you have some context, Maya is the AI agent for travel companies, helping brands automate conversations, support internal teams, personalize engagement at scale and convert more travelers across channels. You can learn more at MayaTravel.ai that's M-A-Y-A Travel.ai thanks again to Maya, Ben, and the team for being a part of our agentic series and helping bring this to life and we're going to have a third episode as well and we have three more amazing guests so I will announce those details next week but please make sure you are subscribed on the streaming platform of your choice to be notified when new episodes go live. And then don't forget we do post clips and highlights on our social channels which you can find on LinkedIn, Instagram and YouTube at Travel Trends Podcast. And as everyone should know we do send out a monthly newsletter to recap all of our episodes and tell you what's coming up next, including the events that I will be at over the course of the year. So make sure that you are registered at Traveltrendspodcast.com. And finally and most importantly overall given this is our Igentic AI series, we have officially announced the dates for our summit October 28th and 29th 2026 and right now thanks to our sponsor Tor Radar you can register for free. The first thousand registrations are free and we'll be announcing more details next month on all of our speakers and sponsors thanks again for joining us and until next week safe travels