AI Disruption in Travel: Session Recap: Key Takeaways from Victor Davidson at Digital Travel
At Digital Travel 2025, Victor Davidson, Senior Director of Data Strategy and Operations at TripAdvisor, delivered a case study presentation titled "AI Disruption in Travel: Four Trends and How to Take Advantage of Them." Drawing from his expertise in AI products and partnerships, Davidson contextualized AI's rise against historical innovations like the printing press and internet, emphasizing its unprecedented speed. This session was essential for travel leaders navigating AI's transformative impact on search, social engagement, and autonomous booking.
Key Takeaways
1. AI-Driven Search Challenges Traditional Models
Davidson traced search evolution from cluttered Yahoo pages to Google's streamlined results, now threatened by AI overviews that keep users on-platform, potentially erasing 50% of organic traffic per Gartner. Paid search dominates, sidelining content quality. TripAdvisor counters this via partnerships like with Perplexity AI, embedding interactive hotel maps and real-time pricing directly in AI engines—neither organic nor paid. This shift urges brands to prioritize deep AI understanding and strategic alliances amid declining SEO reliance.
2. Multimodal Social Expands Engagement Beyond Text
AI enables multimodality, combining voice, vision, and video, to transform social interactions. Over half of US users engage via voice like Alexa, where TripAdvisor deploys AI-summarized reviews for quick hotel insights. Vision tools in Meta glasses and Snapchat lenses offer real-time landmark ID, AR maps, and menu translations. TikTok drives full customer journeys, from planning to potential booking. Travel brands must adapt content for these senses, moving beyond websites to meet evolving discovery and execution needs.
3. Agentic Travel Automates Actions Autonomously
Agentic AI doesn't just advise, it executes, like Perplexity's Comet or ChatGPT agents booking TripAdvisor experiences for Orlando conferences. Users oversee as AI navigates sites, fills forms using stored data, and prompts for inputs like phone numbers. This raises questions: if agents visit sites, not humans, do ad models hold? B2B uses include CRM agents for personalized marketing. Governance, talent for GEO (Generative Engine Optimization), and redefined customer experiences are critical as agents scale.
4. AI Age vs. Internet Age: Shifting Customer Control
Internet empowered self-research over travel agents; AI reverses this, with users delegating planning to tools like ChatGPT for Bermuda itineraries. Social evolves from sharing to full experiences via lenses. Websites transition from empowerment tools to agent interfaces, challenging revenue from human-viewed ads. Davidson urges rethinking connections in an AI-led era.
Why It Matters
Victor Davidson's insights at Digital Travel 2025 spotlight AI's acceleration in travel, mirroring 2025 trends like 40% of travelers using AI for planning and agentic systems optimizing operations from room allocation to dynamic pricing. For industry leaders, this means tackling traffic loss from AI search (e.g., TripAdvisor's 400% ChatGPT referral surge) and preparing for autonomous agents disrupting ad models.
Opportunities lie in partnerships and multimodal content, fostering resilience amid Nvidia's $4T valuation signaling massive business stakes. Adapting now ensures competitiveness in a landscape where AI handles journeys end-to-end.
Actionable Insights
- Pursue AI partnerships: Collaborate with engines like Perplexity for embedded experiences driving direct traffic.
- Optimize for multimodality: Create voice summaries, AR lenses, and video content for voice, vision, and social platforms.
- Prepare for agentic AI: Redesign sites for autonomous navigation and test B2B agents for personalized marketing.
- Shift to GEO: Hire talent for Generative Engine Optimization over traditional SEO to maintain visibility.
Want more insights from Digital Travel? Click here to learn more about the program.
Click to View Full Session Transcript ▼
2025, Digital Travel. CASE STUDY PRESENTATION: AI Disruption in Travel: Four Trends and How to Take Advantage of Them
Victor Davidson, Senior Director, Data Strategy and Operations, TripAdvisor: I'm gonna talk to you about some AI trends in travel. There's a little bit of false marketing. It's not actually four trends. I'm gonna talk to you about three of them, but they're going to be extra compelling cases. So first let me, oh, all sorry about that. Going very fast here. Okay. So first, a little bit about me.
Again. My name is Victor Davidson. I'm the Senior Director of Data Strategy and Operations at TripAdvisor. I lead a lot of our AI product and AI partnership decisions. So I get to be a part of a lot of what we're building in the AI space, which is incredibly exciting. Prior to TripAdvisor, I worked at Wayfair, which many of you know, is a e-comm retail furniture supplier.
And then prior to that I worked at the Havas Network for about five years, had the privilege of working across a number of different travel brands, including Choice Hotels, Emirates Airlines, and Carnival Cruises. So getting up here and talking to you all today is really an intersection of two of my passion points, travel and ai.
Now, before I actually dive into AI specifically, I thought it'd be helpful to anchor us a little bit in where we are in history because we're experiencing a time of radical change, and I think a lot of what AI brings to bear for us is actually a result of a number of foundations that were set hundreds of years ago.
So one of the innovations that happened about 500 years ago was the invention of the printing press, which we don't often think about today, but that was one of the first times that we were able to take information from one person and share it broadly. Now, granted the scale starts to change as we move more towards the present.
But that was something that revolutionized the way that we talk and interact with each other. The light bulb might be an interesting thing to also put on this chart, but the way we use energy and have actually used it to our advantage is something that very materially allowed us to use ai. And then of course, the advent of the internet and web browsers.
And then closer to today, we now have the dawn of ai. And so it's in this context that I wanna start to talk about some of the trends that we're seeing. But part of the reason that it's so important that we have this discussion today is just the rapid pace at which things are changing. Brendan talked earlier today and did a great job of teeing up a lot of the content here, but one of the charts that I love to look at is on the left hand side and what this is showing is the time that it took for any product or technology to reach a hundred million users in adoption.
And so as a frame of reference, Netflix, which I assume many of us use to consume content, took about a decade to move from start to a hundred million users an entire decade. Chat, GPT reached that same threshold in two months. So when we talk about things moving faster. It's not just marketing speak. This is very truly a rapid period of technology evolution and adoption.
Now, some of you might also say that's great that there's a hundred million users, but what does that actually mean for me and for business? And so there I have the Nvidia stock price, and that's the trend of Nvidia over the past 10 years, I believe five, 10 years. If you don't know, Nvidia is actually the manufacturer for a lot of GPUs, which underpin AI technology.
And as of last week, NVIDIA's market cap was $4 trillion with the T for most of 2025. Nvidia was the most valuable company in our world, and so there are very real outcomes and business that is being driven by ai. Which is part of the reason that it's so important for all of us to be talking about. And lastly, this is also important to TripAdvisor.
And so what you're seeing here is the traffic that TripAdvisor gets from chat, GPT. And so from the beginning of the year to about the middle of the year, we saw an astronomical rise. And this is continued, but we saw about a 200% increase in the traffic that we were getting from chat GPT. Now it's off of a small scale.
True. But this trend is continuing and so as of today, it's about a 400% increase. And so while it's small in scale today, what it will be a year from now is anyone's guess, but it will start to become more and more material. And so what I'm gonna talk about overall are three different trends that we at TripAdvisor focus on.
And in fact, these make up the underpinning of a lot of our AI strategy. So the first is how we think about AI search, which you heard a lot about in our first and second keynotes. The second is multimodal social, and then thirdly, agentic travel. And I'll talk about and define each of those as we get there.
So first, in terms of AI driven search, I wanna do another quick history lesson. Many of you will probably recognize this. This was the Yahoo homepage, circa 2000. And what is, at least for me, immediately noticeable is just how crowded it is. There's every link that you could possibly use and then a whole bunch of links that you will never use.
And this was the default for the way that we access the internet and gathered information that is until an upstart and Google came along and simplified everything, giving us just a single bar that we could ask a, not ask a question, but do a search from right. And Google's results were designed.
To show us answers to whatever information that we wanted. Now, most striking today is the amount of organic content that Google put forward. And as we all know, this is not true in travel today. Today you do a hotel search and above the fold, the first page of the sarp result that you get is all paid, right?
So we were already on a trajectory. Where it was getting harder and harder for us to connect directly with our customers. For many brands, whether they be in travel or not, organic search was the primary way that we got traffic to our site. It was the way that we thought about how to communicate with people, right?
Because Google trained us to think in terms of keywords, and if I have the right keywords on my website, Google will drive people there. And so his paid search became more and more prominent. It was less about do we have good content and more about do we have deep pockets? And sadly, that trend is only getting worse now that we have ai.
So this is Google's AI mode. Again, you saw screenshots of this earlier, but what you notice is that the core part of these results don't have organic or paid links, right? What you might get is a citation, but for the most part, Google is answering a question and intending for you to entirely stay on that page.
Or interact with other Google properties. So Gartner did a study a couple months ago and they said that 50% of organic traffic is going to disappear. Imagine that half of a large piece of your business is just gonna go away. And it's even worse when you look at other competitors like chat PT, because at least Google had some links that they're showing in AI mode.
If you go to chat GPT, generally there aren't any links. Now there are citations, and so you can navigate off of chat GPT, but it's becoming more and more important than an ai. Understand your business and understand it deeply. So this presents a lot of problems for marketers and business problems too.
How do we show up and how do we bring people to our website? And so you can talk about GEO, right? Generative AI optimization. But the reality is that channel. In terms of its scale is going away, it is declining. It's not gonna disappear, but it is not gonna be what it was before. And so at least a TripAdvisor part of what we think about is how do we find other ways of talking to people?
How do we make the best of the situation that we're in? And so one of the ways we think about that is through partnerships. Because with all of this rapid change, it actually creates a lot of opportunity for smart thinking and new ways of P speaking to customers and bringing content to them. So earlier this year, we inked a partnership with Perplexity.
If you're not familiar, perplexity is an AI search engine. Smaller in scale than chat GPT. But by reaching out to them and establishing a partnership whereby we share some of our review data privately with them. They pre-built an experience with us. This is neither organic nor paid. This is something that's built into the perplexity experience so that if someone is looking for a hotel, they can go to a map, they can interact with it, and they get the benefit of all of Trip TripAdvisors, real time availability and pricing.
And so I would challenge all of you to think about partnerships as an added way of driving new traffic. Now also, as we think even further outside search, obviously they're social, and I would expect that many of you already have robust social strategies. But what's interesting about AI and what AI is bringing that didn't exist before is multimodality.
And so when I use that term, I'm referring to things other than just image and text, right? And so one of the big ways that multimodal is taking off is invoice. More than half of the US population engages with voice in some capacity or another. Now, Amazon Alexa isn't a social platform per se, but more and more the way that people expect to interact with the world around them is via voice.
If anyone has kids, they will likely resonate with you that if a kid picks up a cell phone today, at least my kids. They don't think about typing something out. They think about talking to it. And so in terms of an implication in a world where more and more of how we engage is via voice, we need to make sure that our brand can show up there.
One of the approaches that we've taken is to be really thoughtful about our review content and what it might look like on an Alexa surface or a Google Home surface. And so we're actually deploying a lot of AI in helping to create review summaries so that if someone calls out to Alexa and asks to learn about a hotel, we've already got a pre-canned snippet that takes all of the review content that we have on a hotel and summarizes it down to two or three lines.
And so that can easily carry over to any brand. You just, we just need to be thoughtful about how we're consolidating that content and the way that customers want to engage in voice. Another modality is vision, right? For many of us, the way that we see is our number one sense, that's how we engage with the world.
And so both Meta and Snap have rolled out lenses or glasses that we can use to engage not only with the social platform itself. But with the world around us. And so like an obvious example in the meta case is someone's walking around in a destination and they might not know what a landmark is, and just by looking at it with the glasses, they can learn about it.
They can essentially start to do discovery while they're in destination. This has many other implications, right? Because you could think about a more mixed reality use case, which is something that Snapchat's going heavily against, whereby a user could have a snap lens on and actually create a map in the palm of their hand, and it will move as they move around.
Another great use case for people that are traveling in a foreign country is menus and direction. So the Snapchat lens in real time can do translation so that if you're out at a restaurant and you're looking at a menu and you need it translated, it happens instantaneously. And so this is just another way that we can think about engaging with our customers, but obviously requires a new way of thinking and a new way of building.
It's less about our website content. And digital as that used to be, but more about how we engage with them with different sites, sounds and motions. And last but not least is video. Video also has been around for a very long time, but anyone that's been on TikTok knows that TikTok has not just about being social, it's not just about exchanging information and ideas, it's also about execution.
TikTok is very much trying to take the entire customer journey and make it possible to do on the platform. So while there aren't many travel use cases on TikTok where you can do the full lifecycle purchasing things, retail on TikTok is exploding. And particularly for younger millennials, more and more of the customer journey in travel is a function of social.
They're doing the planning, the thinking, the iterating. The selection of destinations, of hotels, all in the social platform. And within the near future, they will actually do the booking itself on the social platform. So one might ask what is left? Is AI going to start doing everything for us? And the answer might be, yes, agent travel.Victor Davidson, Senior Director, Data Strategy and Operations, TripAdvisor: So if you're not familiar, an agent refers to a type of ai. That doesn't just share information with you, but can actually execute things on your behalf. And so when we think about Agentic Travel, it's a customer being able to talk to an AI and have the AI actually book things for them. As a function of our partnership with perplexity, we got access to Comet, which is Perplexity, agentic Browser Chat.
GPT also has one GPT Agent. And as you'll see. It looks just like any other web browser when you start, but in addition to asking it questions, you can tell it to do things. It can automate tasks for you. And so in this case, what I'm doing is I'm telling perplexity that I'm going to a conference in Orlando and that I'm gonna be there for a set of dates and I ask it to find things for me to do all on its own.
It opens its own web browser, navigates to trip, TripAdvisor, and starts searching for listing. I'm in the loop so I can see everything, and you might be able to see, depending upon how close to the screen you are, you can see that it actually shares, its thinking out loud. It tells you, I'm going to TripAdvisor.
I see the search bar, I'm gonna enter experiences. I'm gonna look for the type of things that you like. And so at any point I can interrupt it, I can stop it, I can refocus it, but it's acting almost entirely autonomously. And so as we move forward it has created a list of experiences that I could potentially book on TripAdvisor, and let's say for example, that I wanted to try the Kennedy Space Center.
So I'll tell it. Please go and book the Kennedy Space Center, and off it goes. It'll again create its own browser window, go back to TripAdvisor, pick up where it left off, and start to actually book that experience. Now, I'm gonna pause just for a second so that you can all let that sort of sink in. That this agent is what's visiting TripAdvisor.
It's not me. And so this has a number of different implications for all of us, for all of travel, that in a world where people aren't visiting your website anymore, what's the purpose of your website? And many of us. TripAdvisor included, have major revenue streams that are derived based on the advertising that we put on our website.
And in a world where an agent is seeing that advertising and not a human, does that business model make sense? Will advertisers continue to pay for that? These are outstanding questions. The last step is when I ask it, it's identified. The experience that I want, it's identified the dates that I want and I tell it to go and book it.Victor Davidson, Senior Director, Data Strategy and Operations, TripAdvisor: And so the great thing, maybe scary thing, is that Perplexity already knows my name, it knows my email address, and so it can go into the form on TripAdvisor and start to fill it out. As it starts doing that, it has a question. So one of the items on the field is phone number, which it doesn't have access to, and so this is where the interaction comes in.
So I give it my phone number and it picks up right where I left, where it left off, and starts to fill in the phone number and it continues to move down the booking flow. The last piece you'll see is that it brings up my credit card because I've got that stored in my internet browser and it asks me if I want to book it.
And so the world in which we can actually have an agent go out, navigate the web and conduct activities for us is very close. So today the AG agentic browsers. GPT agent and comet are relative, are very small in scale. Both the perplexity and OpenAI have only really rolled these out to some of their premium customers, but this is something that is coming.Victor Davidson, Senior Director, Data Strategy and Operations, TripAdvisor: And as I mentioned, this presents a very thoughtful foundational question about what our websites are and how we maintain a connection with customers in a world where an AI is doing the vast majority of the work for us. Very quickly, I also want to touch on the fact that agents aren't just a B2C use case.
So more and more companies are starting to deploy AI agents to actually help with marketing activities themselves. And so a good example is amp, which is a very B stage startup company, but they're basically deploying agents in a CRM context where they pair one agent to one customer. And so that agent is focused, razor focused on just one person at a time, and can optimize the email send based on time of day, the copy, the number of communications, and actually making real a promise that we've been talking about for decades in terms of one to one marketing at scale.
There are also a whole host of other trends and topics that I'd love to talk about, but I have about 50 seconds left. So I'm not gonna be able to do that. But all of this presents very real governance problems, right? Because how do we make sure that our customer's information is still secure in a world where our customer isn't actually the one entering the information and talent, obviously.
So what are the types of roles that we need to be hiring for in a world where agents are running the show or where we are? We are optimizing for a EO or GEO and not SEO. And last, but certainly not least, all of this is changing and evolving. What it means to be a customer and what that experience looks like.
So I'm gonna leave you with three final implications. And I'm framing them based on what it was in the internet age, which we are no longer in versus the AI age that we are. And so if you think about first what search and what the internet meant to travel. Initially, if you go back even before the internet, if you were planning a family vacation, you had to use a travel agent generally, right?
Travel agents knew the destinations, they knew all the right hotels. They could actually manage all the booking for you. The internet actually revolutionized that. We have OTAs now, but it actually empowered customers. To be able to do all that research on their own and have a high degree of confidence that, yeah, I can book a trip just as good as the travel agent.
What we're seeing is actually the inverse in the world of ai because people are now relinquishing a lot of that control to the AI they're asking. You can go, any of you can go to chat GPT, although I'd prefer you do it on TripAdvisor, but you can ask, how do I plan a seven day trip to Bermuda and the AI can plan it all for you.
So the way that we think about bringing our content forward is really changing. Second, in the internet age, social was how we started to interact with other people. We learned about what someone's doing or why they're doing it, and we got to humble brag about all the great trips we went on in an AI age.
It is much more all encompassing. It's not just sharing information with one person, it's sharing it, one with the whole world. But then actually doing the entire experience, right? So there's a world in which you're wearing your lenses the whole time that you're doing a trip and you're booking the entire trip through those lenses.
So social is becoming a larger space for customers to interact, and brands need to be present in that, in the way that consumers expect them to be. And then last but not least, in the internet age. It was empowering for customers. Look at all the things that the internet can help me do. But today we're starting to enter a space where we are not doing anything.
The AI is doing it for us, right? And so the purpose of a website and the way that we think about interacting with our customers is going to change. And so with that, thank you very much. I'll be around for questions.
Thank you so much, Victor.