Filip Filipov on AI, Travel Data, and the Connected Trip

0:00ZS is a global consulting and technology firm that partners with companies to improve life and how we live it. Working with the world's top travel brands, ZS brings together data, science, technology, and human ingenuity to transform ideas into consumer loyalty. They're ready to help you elevate the journey, improve operations, and uncover preferences and motivations through their AI-powered solutions. In the instances that call for more, they help travelers choose the finer path and never look back. With more than 1 ,700 clients in 170 countries, ZS has built the expertise, fortitude, and vision to help you deliver impact right where it matters.

0:38All right, everybody. Welcome back to Skift Take Sessions. I'm Will Slickers here in London as we are hosting the Data and AI Summit. And I am here with my special guest, Philip Filipov. You were just on stage with Seth Borko. A great session. Quick, you know, random Q &A, I would say, of like data numbers. What does it belong to? But Philip, you're the CEO of OAG And just for the listeners that don't know Can you tell us a little bit about OAG Your role and what you guys do in the industry Thank you Will and thank you for the invitation Glad to be part of this event and also this interview OAG is Almost a 100 year old company Which collates and aggregates Schedules, flight status and pricing Information and provides it back To airlines, airports, financial institutions And trial technology companies So we aggregate a lot of fragmented data points and we give it back to the industry so they can make better decisions.

1:30I love it. And we're going to talk about fragmentation because you actually said something around that in your session with Seth. But before we do that, I want to talk about the blame in AI because Seth walked through that. He gave instinct as AI agent, all these rules and things to follow around booking a flight, but he forgot to do his pre-check, his TSA pre-check. And so that's one little mistake that doesn't ruin the trip, but he does have to go through a longer line. But let's say an AI agent doesn't book the trip properly, books the wrong flight, but says it confirmed the flight that you wanted, etc.

2:07Who's to blame? Is it the data that the AI agent got or is it the AI agent itself? I mean, in Seth's case, I think it is Seth because probably Seth didn't tell the agent that he has the pre-check and that should be used in further bookings. In this case, it's interesting because sometimes the data is at halt because it's not connected. I mean, that's what we try to solve as a company so I can talk knowledgeably about it. The better, the freshest, the most complete data you have, the better decisions you can make. The agent right now at this stage is acting more of an automation tool. So the input that you give it is equal to the output you're going to get out of it.

2:42In the future, when that mistake happens, because AI will be a lot more capable, if the agent made a premeditated decision saying, in this case, I will put Philip on an earlier flight rather than a later flight. I think it will be the agent and that needs to be fine too. But ultimately, we're users and we have to accept that sometimes it's our fault. 100%. And it was Seth's fault. So Seth, when you watch this, we all know it was you. It was him. Yes. And I want to ask you too, because I think, not saying everybody in this room, but a lot of travel companies, as I'm scrolling through LinkedIn, are all saying now that they're data-driven.

3:19You are an actual data company in the space, so I 100 % believe that when it comes to you guys. But what makes a difference between a company that is, or maybe thinks that they're prepared to be a data-driven company rather than it actually is prepared to be data-driven with the implications of AI using models, building on that data? I think part of what we're missing is, even the companies that say they are data-driven, they might have some data sets or live data in different silos within the company so the data that they have needs to be connected it needs to be accurate and it needs to be fresh you cannot be working on information that happened like months and months ago i think it's a nice change that we've seen before because if previously you would think of product people as some genius who shows up and says i believe the world wants x and i'm gonna build it and we've had that in history and they're incredible individuals.

4:13Let's assume the rest of us are not Steve Jobs and Elon Musk. I think everyone else needs to back their statements with data. And I like the fact that we're moving to this data driven world, but I think there's quite a bit of distance we need to travel to actually say all the decisions we're making are data driven. Yeah. Well, Seth also made the comment too, you guys are kind of in the shovel business right now. There's a goldmine going on and you guys are selling the shovels, which I think is true because data, I never looked at data the way I do or did until AI really became as big as it is.

4:46And so one of those things that you've kind of been talking about is fragmentation. And this is a huge conversation. I've talked about this on multiple of our other podcasts. Fragmentation in our industry, when I was a hotel front desk agent, I had to log into five different systems just to log someone in and give or check someone in and give them a key. And you said that that's not going to go away. Fragmentation is not going to go away, but it's going to make travel feel connected. So do you think fragmentation will change on the B2B side as AI and data become more relevant? Or is it still going to be fragmented, but we're just going to use AI differently to help connect the pieces?

5:21I think we'll use AI in its most empowering state. So let me break that down a little bit. Today, if my flight is delayed, maybe my wife knows about it and that's it. But if I'm going into a conference like this one, I live in London now. But let's assume I'm staying at a hotel. I shouldn't be stressing out my agent or somebody needs to tell the hotel that I'll be checking in late. And they shouldn't let go of my room. And maybe they should actually offer me to use a local delivery service for food. That doesn't happen today. And I think with AI, we can actually unlock it because there are many different pieces that needs to be stitched together for that experience to really shine.

6:01I think AI really unlocks it. Because previously, you could do it, but it was a lot of work. Nobody invested in it. You don't have a transaction at the end of it. So people invested money somewhere else. Now with AI, given that this experience, putting it together is a lot cheaper. I hope to see a fantastic new way for us as travelers to experience travel. Now, is data sets going to stay fragmented? Absolutely, yes. Like every hotel in a small village in, let's say, Malaysia is going to have its own little data silo. And that will be there forever. and maybe we'll access it or maybe we won't be able to do it.

6:36But fragmentation is part of travel. That's the beauty of travel. So I think it will always remain fragmented as a source, but I hope it will be a lot more connected as an experience. So the customer won't see the fragmentation though. Their agent's probably going to see the fragmentation. Ultimately, yes. I mean, do you care? I mean, I would love it if every time you check your flight status, it comes from us because that's what we do. But ultimately, do you personally care? You just want it to be accurate and available. Yeah. So that's where we're headed. I love it. You know, I have a bet. And this is my...

7:09I'm a big Uber fan, big Dara fan. I know Uber's a customer of OAG. But who do you think is going to win at making the connected trip for the customer from door to door? Right? Like from the Uber ride to the airport, from the airport to the other Uber to your hotel. Who's going to win? An OTA? Uber? Independent brands? I don't... I think it is the false race a little bit. I think Dara is incredible. And what's happening with Uber and the evolution into hotels right now is just fantastic. Also, Booking has been trying to do it for some time. Expedia has been trying to do it for some time. And I don't think there needs to be a winner because ultimately the closer we get to that goal, the better it is for the traveler.

7:53I think with the authentic experience we have right now, you don't have to. Dara and Tim doesn't have to own flights and trains and hotels and first mile and last mile and airport check-in. They don't have to own all that. They can be really good at what they do. And then an agent maybe will be able to connect all of that across. Now, each one of them will try to capture more of that trip. Fair enough. And if they can do a great job, I hope they will win it. but I don't think it's going to be any of the players because I don't think there is a finish line at the end. There's a lot more that can happen.

8:31Is your passport expiring? Are you ready with your luggage? Then how far into that end-to-end trip you go? Did you make sure you checked with the kid's schedule? It's cool. How far do you stretch it? And it's endless. I love it. It's a hot take. I like it. I want to go into my final question for you. It's around betting on AI. So what burns money the most when it comes to testing and implementing AI at a travel company? I think right now we're token maxing, if you will, because people simply don't know how powerful the models are. And sometimes we're throwing easy problems that can be literally solved with a VLOOKUP in Excel to a model that's multi-billion and costs a lot to actually run.

9:14I think over time, we're going to become a lot more precise in where we use AI and how. I think coding will be one area which will be fantastic for optimization. I think data analytics, provided that you have the data, will be a fantastic area for growth. I also think that some of the operations within a company will be optimized. But right now, we're still on the adoption curve. And there's some time before we all can say, that's it. We've adopted the best and the latest. And now it's time to optimize. So it will burn money for some time. We're not going to see it immediately affecting us. Well, what would you say to follow up on that?

9:49Just like, what do you measure and what metrics really matter? For us? Yeah, for you guys or even customers that you're talking to? Let's start with us because that's a little bit easier. I mean, we're 350 people. And for us, the important piece was to make sure that adoption is high, that people are using it every day. And right now, I believe 94 % of OEG, which is a 100-year-old company, our employees use at least once a day some form of AI enabled tool. So I love this. But also we're nimble and we're small and we can move fast. Yeah. If you're sitting at an airline with 4 ,000 people and revenue management and technology, and you actually have people's lives in your hands, are you going to tolerate release into the workforce to go and do whatever you want to do?

10:35Yeah. Probably not. So I think we're still on that adoption curve, especially for bigger corporations. I mean, banks are not able to use anything but co-pilot and work. True. So, yeah, a few years before that, that hits really the P &Ls in an appropriate level. So what metric would you say is important for you guys? For us, it was adoption initially. Now we're looking at our state, which are self-reported. So how many hours have we saved in doing what we do now that we use AI? It's in the thousands and it's great. Then the next question will become like, what is the outcome? Like, are we actually working on the right things?

11:13we're still on that stage and that journey. But yeah, we started with adoption. Now we're looking at the savings. Yeah, I think a lot of people are on that same spot where it's like, okay, we've saved money. We've saved time, maybe a little bit of cost. But what do we do with this now moving forward? What's going to be the important goal to have focus on it? So the job still remains. The task is a little bit different. We still need to get the code done and deployed. Yeah. But would I have an engineer sit down and write line by line for the next six hours and then have somebody go and check the code and for a PR and all that.

11:46Does this make sense anymore? Can you spend 30 minutes checking the code of an agent or even ask an agent to check the code of your agent and then move on and build something else? I think the productivity gain here is I hope it is for growth rather than let's just cut costs because now two people can do what 25 people used to do with it. Cutting costs is like the low hanging fruit. I don't think it's fair as well. I think it's very limited. We should run businesses logically and we should be careful about the cost. Yeah. But for us to really grow, which is the purpose of a business, to grow and expand and serve more customers, we should really invest that safe time into building better products fast.

12:24For sure. Yeah. All right. Well, I really appreciate it, Philip. It's been a great conversation. I know there's going to be a lot more happening here at the Skip Data and AI Summit. But for anyone who wants to check you guys out, where do they go? OAG.com. We're right there. Easy. Easy. Well, make sure everything's linked to the show notes. Make sure you check out the next session, and we'll see you guys on the next one. Thank you.

13:10They're ready to help you elevate the journey, improve operations, and uncover preferences and motivations through their AI-powered solutions. In the instances that call for more, they help travelers choose the finer path and never look back. With more than 1 ,700 clients in 170 countries, ZS has built the expertise, fortitude, and vision to help you deliver impact right where it matters.