STR Global Unlocked with Simon Lehmann: Unfiltered knowledge for the short term rental industry

034: The AI Shift Most Property Managers Are Missing | Sab Mulligan, Arnon Cavaeiro, Chris Sorensen

• Powered by AJL Atelier • Episode 34

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 57:27

In this episode of STR Global Unlocked, we explore what an AI-native property management company really looks like and how AI is already changing operations, guest communication, distribution, PMS systems, scaling, and hospitality brands.


Together with Sab Mulligan, Arnon Cavaeiro and Chris Sorensen, we break down why AI is not just another tool or chatbot, but a deeper shift in how short-term rental and hospitality companies operate. The conversation covers where AI is already reducing costs, where it still fails, why human expertise remains essential, and how property managers can prepare for a future where AI, automation, brand, and guest experience become increasingly connected.


Questions answered in the episode:

  1. What does AI-native property management actually mean?
  2. How are leading property management companies already using AI in daily operations?
  3. Where can AI reduce costs without damaging the guest experience?
  4. Why is the PMS no longer the brain of the business?
  5. What happens when broken processes are automated with AI?
  6. Where does AI still fail in hospitality and short-term rentals?
  7. Should property managers build their own AI tools or work with tech partners?
  8. How could AI change travel discovery, direct bookings, and OTA dependency?
  9. Why will human connection still matter in an AI-driven hospitality business?
  10. What will property management companies look like in the next five years?

Resources

Stay connected:

SPEAKER_03

AI is changing property management faster than most people realize. But the real shift is not about adding another tool, another dashboard, or another chatbot. It's about building a company where AI becomes a part of the way business actually works. In this episode on STR Global Unlock, I'm sitting down with three people who are already working on this from very different angles. We have a brand strategist who looks at how AI can shape the guest experience and build strong hospitality brands. We have a head of product who is building AI into core systems and workflows of a property management company. And we have a CTO who's developing property technology to scale operations, distribution, and to full guest and host experience. So it's not a conversation about theory. These are people who are testing, building, and scaling AI inside real property management companies right now. Together we look what AI native property management company really means, what AI already creates real efficiency where it still fails, and why the future of hospitality will still depend on real humans. I am Simon Lehman, and for more than 20 years I've worked across short-term rental, property management, travel, and hospitality. I've seen this industry grow through major shifts from distribution and technology to operations and guest expectation. And if this conversation makes you think more deeply about your own STR business, your operating model, or the role AI could play in your growth, you can find a link below to book a consultation with me. Joining me today, Sab Mulligan, brand strategist from Sing in Malta, Arnon head of product from RBO in Berlin, Chris, CTO and co-founder from Ren Folk in Denmark. Sab and gentlemen, great to have you all here. Let's start with a quick round short answers. Sab, I want to start with you.

SPEAKER_04

Okay.

SPEAKER_03

AI and hospitality, hyperstructural shift.

SPEAKER_01

Shift for sure. Non-negotiable, it isn't a feature.

SPEAKER_03

PMS, still necessary or becoming obsolete?

SPEAKER_01

For me, non-negotiable is the engine. Not the brain, the engine.

SPEAKER_03

OTAs, partner or dependency?

SPEAKER_01

Partner, also tax on demand. Respect the rich though.

SPEAKER_03

And the head count in Zing in five years, up or down?

SPEAKER_01

Still up. Better careers, more hospitality-driven careers.

SPEAKER_03

Wonderful. Arnon, I'll go next with you. AI and hospitality hyperstructural shift.

SPEAKER_00

Structure shift. It's gonna be the basis of everything.

SPEAKER_03

PMS, still necessary or becoming obsolete?

SPEAKER_00

Become obsolete very fast because the foundational LLMs are getting so smart that they could replace in a few years.

SPEAKER_04

OTAs partner or dependency?

SPEAKER_00

Partner for now depending on how people guests are gonna looking for accommodation in the future, but partner for now.

SPEAKER_03

And head count in five years for RBO, up or down.

SPEAKER_00

We wanted to grow without hiring, so we staying the same.

SPEAKER_03

Okay, Chris, AI and hospitality, hyperstructural shifts.

SPEAKER_02

Structural shift for sure.

SPEAKER_03

PMS still necessary, becoming obsolete.

SPEAKER_02

It's just a bunch of tools that you build yourself, but the AI can use it.

SPEAKER_03

OTA is partner dependency.

SPEAKER_02

Uh just another distribution channel. AI will be another.

SPEAKER_03

And headcount in five years for land folk, up or down?

SPEAKER_02

I think it's gonna go up.

SPEAKER_03

Awesome. So let's kick this off and talk about AI native. AI native is sort of a new term within the short-term rental industry, and therefore we have invited you as guests to talk about what AI native is or AI-driven. And that's definitely something that came a long way when we originally heard about technology-enabled PMCs with companies like Vacasa and others who have propriety tech for years. Um PMS have taken over as commodity technology, and now it's going back again to being AI native. But can we can but that can mean that anything from a chatbot to a fully automated operation? So I want to get a very concrete answer. What does AI native actually mean in your business, Chris? And I want to start with you on that one.

SPEAKER_02

All right. I think for us, AI native means that we use AI in our daily operations. We use AI to automate and enhance and make a bunch of our internal operations more efficient. So that could be everything from like helping generate articles, creating listings, uh doing fraud detection, things that are very manually intensive work. We use AI to enhance that work or or like remove the human uh the hard part or the time-consuming part of it. We have an AI do that, and we just have a human in the loop to just confirm that things work as they should. But the hard work is an AI that does it.

SPEAKER_03

Arnon, what is there anything you want to add to that in how RBO is looking at that?

SPEAKER_00

There is some overlaps with Chris, but mainly RBO wants to be the wants to build a single source of truth that has enough context so agents can take decisions themselves. And of course, there is exceptions where you need humans in the loop since property management is very human-heavy activity as well.

SPEAKER_03

Which will remain human activity. Uh we'll talk about the human aspect of the business as well in the future. So where does where does you sit, where do you sit, Sab, in this when you're not a truly AI-native uh property management company? So, where is AI truly embedded for you at Singh, and how do you look at that?

SPEAKER_01

So for me, I'd start off by saying AI native would mean that the business is built around an AI brain that connects our three rails together. Property management, the three rails are distribution, communication, and operations. So it's not just the shiny layer you add to the top. So to test this at Zing, what we say is if you can switch it off and nothing breaks, then you're not AI native. You're just adding more AI features.

SPEAKER_03

So what would break if you would remove it from your point of view? What would break in your organization if you would remove it?

SPEAKER_01

In guest communication, for example. So our um AI agent, Zoe, she handles 85 to 90% of all communication across guests and owners. If it breaks, of course, this means that we're back to more of a call-center model where we're constantly our humans, our answering humans. That would that is what a break would mean.

SPEAKER_03

So, Chris, I want to follow up with you. Uh, you started to make some examples. So basically, within every workflow or within every piece of the business at Landfolk, you're basically utilizing AI and large language models as technology. So that could also be part of a recruiting process, your internal processes, anything the like.

SPEAKER_02

Yeah. I mean, sure. I mean, I think we started almost like three years ago or more, before ChatGPT even came out. Um and we started out by just, I mean, we're not that many people. We have a very small engineering team, so we've always been faced with figuring out uh interesting ways of handling these things. Uh and using AI was pretty obvious for us to use that as a leverage to be even more efficient. And as I mentioned before, I mean, we do it for listing descriptions, we do it for uh paid search, we do it for uh fraud detection. Um but we're also rolling out like more internal agents that that just live in where we work that can automate a lot of our jobs or at least automate a lot of the business insights because we build our tech stack in a way so that it's all um built on a single source of truth with our own tools on top of that. So when we want a business analysis to do something, we can just ask the AI and it helps us generate dashboards or reports like on demand. Um I mean it's just like everything that we do is just AI that is enhancing that workflow.

SPEAKER_03

Anything to add from any of you?

SPEAKER_00

I would have you as well. We say AI solving direct problems, but at the same time it's a tool that other departments can leverage as well. For example, you mentioned hiring. Hiring can leverage AI as well, or operations. Uh, they are having access to AI tools that they can build their own tools to improve the system. Example, we are using AI in Lovable to detect double bookings. Something very simple, but we don't need to have all this capacity on the engineering team. But also, operations team are using these AI tools themselves to solve their problems. And this increased the productivity so much.

SPEAKER_03

Yeah, I can imagine. So that's it's a great segue to talk about operation and cost structure uh as a next point. And because that's where the real impact should be at the end of the day of our business, it's still operational heavy. You already alluded to that in terms of you know, it's still human capital is important. That's where a lot of our cost is. Where has AI actually reduced costs in your business? And this is a tough question because being AI native, you say, well, we have not created that cost. But maybe I want to start with Sab on that to start with.

SPEAKER_01

First of all, based on what you were saying before as well, so I call myself an AI-first brand strategist, but for me, it's really important that the tech is silent and that the hospitality is what people feel. I don't want our guests to feel the tech. And I think that's really important to mention, because we're not really property managers, we're hospitality brands managing, in our case, both hotels and vacation rentals. Well, as regards to your questions about costs, the equation is broken. Before it was more keys, more head count. Now, that isn't like that anymore. In our case with Zoe, she has absorbed the equivalent of thousands of hours and conversations that, of course, used to sit on payroll before. So handling 85 to 90% of all conversations across the guest journey, from booking to post stay, plus keeping in mind the consistency, the fact that it is multilingual, the fact that she or it can answer voice notes as well. Of course, this is really important from my end with my brand hat on, um, because it's building your own digital twin, right? And in the past, I have found that very hard to build a team who speaks the language of brand when you're also very passionate about scaling as well. So sitting between, I want to host like I am boutique, but I really want to grow like that, there is no ceiling. And for us, that is where AI has come in. Of course, also opening service calls for our operations team. And then, leading on, probably we'll be talking about this as well, Agentic AI, which is helping us make business decisions. And that sits at the top of all our tech.

SPEAKER_03

Chris, um, while landfolk is taking an international approach, I mean, you need to build markets from scratch. And at the same time, you need to look at your cost of operation. So you're building that already off an existing model. But, you know, can you uh run us through that in terms of improved efficiencies and how you adopt that when you enter new markets as well to make sure you capitalize on that efficiency right from the beginning when you enter also new markets?

SPEAKER_02

Yeah, I think I think we just build our whole system about being efficient and automating things. And we automated things with I mean, we just used machine learning models before. It was cool to call it AI. Uh and now we call it AI, but it's still just automating different events or things that come into the system that can either be deterministic or it could be an AI that takes a decision here. Um But I think when scaling to a new market, we just build a platform that is easy to scale and use AI to do all the translations and then have real humans in the loop to qualify that everything looks correctly. Um but I mean when we scale to a new market, we still want to have like real humans there so that our uh hosts in new markets and guests in new markets have uh someone to talk to that actually understand uh their market and their problems and their use cases. So we actually hire people in those markets, people that understand that. But we can use AI to uh improve their responses to the hosts uh so that we still have real humans because we actually think that's uh uh a selling point for us, that you can actually talk to a real human being. Uh and that's actually what we see in new markets as we expand that people really like that they can call someone that knows them and knows how their market works. But then the person that they talk to is enhanced by an AI so that they can give even better answers and have more time to spend with them and talk to them and talk about their issues so that they can get even better listings. Because we have premium curated listings, and that's um we need to be more efficient so that our supply managers and guests and host care team have more time to spend with uh the guests and hosts to solve their problems. And then AI can do all the boring work behind the scenes.

SPEAKER_03

So there's a big differentiator between uh landfolk and RBO in terms of the growth strategy. RBO is growing more by acquisitions and has done several acquisitions, whereas landfolk has uh taken more the organic growth um approach as well. So, Arnon, from your perspective, you know, you're you're acquiring more legacy-run property management companies. You've just acquired one in England as well, and and and their integration onto a platform that you have built uh must be a quite a big challenge. And I want to readdress the the question of where the the efficiencies uh can be generated, and we're talking about operation and cost structure right now.

SPEAKER_00

First is the onboarding of new units. Uh, since we're acquiring very aggressively these last months, onboarding in units is something that AI can enhance the productive so much because there is pattern recognition that needs to be invented. But at the same time, uh good point that you mentioned here, uh Simon, is that for every acquisition we need to adapt to a new PMS, a new ticketing system, uh, and so on. So, what RBIO is doing, we are creating the abstraction layer in which we can connect any external tools, PMS, OTAs, ticketing systems, and then we translate this in the RBI system. Think as uh open banking. This is a known concept in finance, but why not building the open banking for property management? That's what RBI wants to go, because if you wanted to acquire aggressively new properties, we needed to have these abstractions later.

SPEAKER_03

Okay, let's talk about some challenges. Where has it not worked the way you expected it? Where where are the pain points in that Arnon?

SPEAKER_00

Before November, the um LLMs, the models were not so capable in order to process new information. But back in November, uh December 2025, the new Gemini 3 or Anthropic um Opus 4.5 model, things changed and things improved a lot. Uh but what we expected more from AI and what didn't happen is the um speed of resolutions and the quality. This balance is still tricky. Otherwise, you need to pay a lot for these models. And this gets really expensive if you trans, I think like transactions in hostaway, 19,000 bookings, uh attempts of the bookings per month. Uh, if you add LLM token costs, this gets really expensive. So to answer your questions, is the balance of cost and speed.

SPEAKER_03

Last November, it's light years away in the sense of AI already, right? So uh that's also shows nicely how uh things uh are are evolving, but there's still definitely still challenges in the cost aspect is definitely something you need to keep in mind as well. Um, I want to address that with with uh Sab as well. I mean, while you're more using a legacy tech stack, and we'll talk about that in a moment. Um, where the where hasn't it worked for you uh the way you wanted it to in what you have implemented in terms of your AI um processes, AI-driven processes?

SPEAKER_01

Well, speaking to an audience of property managers as well, for me it's important to remind everyone that AI isn't just a tool, but it's also your brand character. And so it's not just about AI taking over the tasks which are mediocre and mundane and repetitive, but it's also about building a brand character who can really deliver a strong brand message consistently and faster than humans. Because when you're scaling, then the amount of messages coming in and being answered simultaneously is incredible. So AI fits under brand character, um, is my message there. Where it breaks across guest experience, across operations, because we use AI there as well, is when our policies, our processes are broken. And that is the reality of it all. So if AI is answering incorrectly to a guest, that means the root cause is that we need to update the policy, we tweak it, we train, we test it, and then that won't happen again. Same with something that was um big learning for us when it comes to SOPs and scattered data and information everywhere, whether on jumping from tab to tab or on guest books, basically. Hard reality was that when we started to move forward faster in our journey of AI, when an operational process is broken, then AI is just gonna continue to make a mess of it. So root foundations, really focus on building your SOPs and um all your processes as strong as you can for the AI to work better.

SPEAKER_03

Chris, um, what comes first as SOP or AI? That's an interesting aspect to say we need to correct that if the if the LLM answers wrong.

SPEAKER_02

Yeah, I think it's a good question. I mean, it would be the same if uh if one of the employees did something wrong. You need to treat AI and especially the new more agentic AI as just another colleague with a bunch of tools. Uh and if if it uses the tools in the wrong way, you need to figure out how it won't happen again the next time. The same, exactly the same way as an employee would would be handled. Um so yeah, good operating procedures, good documentation, the right tool sets, access to the right tools and the right guardrails, that is the way forward.

SPEAKER_03

Okay, awesome. So we talked about standards, we we talked about uh we talked about what is possible, what is not possible, where the challenges lie as well. We talked about, and we're not using this uh podcast to talk about the right definition of AI versus LLMs and machine learning, but I could not agree with Chris more than you know, has it really got to do with artificial intelligence or actually it's machine learning models based on an LLM? Is is just you know it's a different uh it's a it's a different definition. I like that a lot better. But now AI has basically made it. But ultimately, um, we're still working off what's around in data, uh, especially in text as well. Um, so let's push a little bit further. We I want to talk about limits of LLMs and AIs. And and Arnon has already alluded a little bit back in November, but now we're in April. Uh, so that's already six months down the line in the compounding rate uh of any technology that means about 20 years of mobile phones and six months of AI. Uh, quite interesting aspect to think about that. So give it give me one function where you thought this will be fully automated and and it wasn't uh possible. Who wants to go first?

SPEAKER_04

Um I can.

SPEAKER_01

So I lead with that hospitality is the industry for humans. Okay. I want to frame it like that. For me, that's really um important. Um of course, our job from a property management point of view is to define how we want the business to run, right? The guest journeys, the operation, our brand promise. And then we're just looking for the tech that can make that happen for us. Like one prompt that I love from an agentic AI point of view is a prompt which has eliminated our human handover. So one prompt, tell me what happened in the last 12 hours, which could lead to a bad negative review. And the AI is just giving me that the AI is just giving me that information so I don't have to. Simon, on the positive side, then I would also mention a prompt that we're using now where it's um tell me who is celebrating their birthdays during their stay. Which means that instead of filtering through lots of check-in forms, the AI is giving me the answer within 30 seconds, and then the humans step up to make that call, deliver that cake, basically. So I think it's also about how AI and humans waltz together in this hospitality journey.

SPEAKER_03

Arnon, I want to go next with you. At the beginning, you said you know RBO can run a business literally without people. Um ultimately that's a goal to stay the same and continue to grow. So that's uh that's a a big uh challenge, I guess, as well. So where do you where do you see things haven't worked for your or AI had its limits that you'd want it to be fully automate, and then we'll talk about the human aspect of it.

SPEAKER_00

I think the um the issue with AI is not that only empowers the business, but also empower the guest side as well. And one thing that can create issues is on the support level. Why I mentioned that. Imagine a guest send a picture complaining that the fridge is broken, and then you as reading the message, okay, here is your refund, or how we can uh solve this issue of your staying. But the guest actually can alter the image using banana model from Google, for example. They can take a picture of your unit of your apartment and then say, Hey, in this apartment, make this fridge look broken so I can get some refund. That's tricky if you don't automate. That's very tricky. That's very tricky, yeah. Um, I saw people doing that, getting money back from ordering foods because I order a sandwich and then I send back, hey, and my sandwich came with mold. Actually, it was just an AI-generated image from the sandwich with mold and then got the refund back. So there is limits to automate so far. Like we didn't find yet the balance to automate 100% these issues.

SPEAKER_03

Yeah, that's an interesting one. You know, last last year I checked into a property and Brian Jeski was sitting on the sofa, and I didn't I said, I don't need anybody in my apartment. So uh anyway, choke aside, um, this is but this is definitely uh going to be a massive challenge, and and that's where the balance of human capital uh comes into the conversation as well. So, Chris, what are your thoughts on that as well, in terms of where do you think where you are at? Maybe not limitation of AI right now, but but maybe you want a little bit allude more about the balance of what Arnold just mentioned in terms of the human capital as well and how land folk looks at that. And you've already alluded to that a little bit earlier, to say, hey, we still want this human capital in the conversation with the homeowners who knows the market, but support them with AI tools. So maybe you have some thoughts on where limits are, and then where does the human aspect come into that?

SPEAKER_02

Yeah, I think had you asked me maybe a year ago, I would probably have come up with a bunch of limitations that I thought we're not gonna get there, but but at this point right now, I don't feel like there's that many limitations anymore, or if any, it's basically just the imagination that is that that is that is limiting here. Um But I feel like some of the issues that we are seeing, or at least is hard, is all the user-facing AI. So if we want to create AI for our guests and hosts to use directly, I feel like the return on investment there is not as big as doing internal AI and improving your own operations. Because creating all these user-facing AI tools is kind of like a wagon mole. There's just going to be someone who tries to trick that AI to do something. So you have to allocate a lot of resources to build that in the right way, and it's very hard, and you have to keep you have to keep improving it all the time. Of course, you can probably have an AI and an evil continually doing this, but it's something that you have to be over all the time. Whereas if you focus on the internal AI operations, uh you don't have the same limitations. Or at least the return on investment is is is bigger. And that's that's what we're seeing. So we've actually been deliberately holding off for building too many user-facing uh AI efficiencies, if that makes sense.

SPEAKER_03

Can you allude to that a little bit? More user-facing AI efficiencies in relation to cost?

SPEAKER_02

I think I think it's not related that much to cost, it's more uh resources. That you need to spend a lot of resources on continually improving it and uh make sure that there are no prompt injections and uh that you don't leak any data if you build uh user-facing AI tools. Of course, we always be building AI tools. We built uh DAIC and our dream search, but that was more like how can we do something that is not a chat interface? How can we do how can we use AI in a way that um gives the user the user another experience which isn't just chatting with an AI, but more like how can you dream about where you want to go and then we return uh inspiring images to you. So that's not just a chat where we have to worry about uh prompt injections and stuff like that.

SPEAKER_00

Yeah, just adding here to Chris as well. We don't need to grow our own vegetables if you can buy our vegetables, right? Uh and that's the problem with AI now. Like it's so easy to build new tools, uh, but to maintain these tools as well, it increases all operational costs to maintain these tools and all these uh evolves that it needs to do in order to work properly. So these builds that's buy uh it's still a valid question in AI.

SPEAKER_01

And from my end, to build on what Arnon is saying, this is why property managers should focus on what they do best: hospitality. And the tech should not be something that they're focused or thinking on. I am not in favor of people experimenting and building their own AI agents. I think we should really elevate hospitality by doing the hospitality side and working with a tech partner who can make it silent, act like our growth engine. Um, also a note: are our guests, and they are, more AI first than we are in this space, right? So there's the expectation that AI is playing in our space. So I really think we need to step up and move a bit forward faster as a collective, of course.

SPEAKER_03

Yeah, thanks for the shout-out, Sab. I could not agree with you more, but you know, at the end of the day, we have 150,000 PMCs around the world with different challenges. Yes, we started talking about hospitality only about three years ago before it was not considered hospitality. Um, more glorified cleaners or you know, all the challenges that we have. But now tech adoption is also uh quite material in making that process work because there are so many inefficiencies in our industry still today. And obviously, with the three uh participants of our podcast today, with you, that you're setting a new benchmark to what is out there. But let's not forget the vast majority uh is still dealing with problems with the bedlin and and and and and the cleaning processes and and everything uh else around. So let's talk about that uh in terms of tech stack, uh, and and I want to go a little bit deeper, build versus build versus buy. So we came a long way. When I was running a property management company in 2005, started it, we were running off AS400, uh, and then we build our own PMS based on SAP uh in 2006. Uh and and we came a very, very long way. At that time, no PMSs were out there, there was no other technology. Moving fast forward to 2025, we have over 420 tech providers within the short-term rental industry, talking about SaaS, PMSs, RMSs, channel managers, operation platforms, uh, you name it. And now there is there is a shift. So we're seeing a shift back towards proprietary tech, which I've already started to talk about at the beginning. So I want to start with SAP first, because I will challenge you uh in the in the quick questions you said the PMS is your backbone, uh, which I think Arnold and and and uh Chris would probably highly disagree with because that's the cornerpiece for them to run the business. So some of you um you are building your own layer, but you're still using a traditional tech stack. Is that correct?

SPEAKER_01

Okay, okay, so I what I meant was the PMS is the engine. So for a long time the industry behaved, I agree, that the PMS was the end all. That's not like that anymore. We also need to have a hospitality brain, basically. So engine and brain. PMS is that engine. In our case, we use guesty as our PMS, and then boom, sits on top of it all and is our brain, and that is where the agentic AI comes in, helping us make business decisions as well, and helping us run the three rails, whether it's communications, um, distribution, and operations. Because let's face it, those are the three rails, and we have six guest moments basically. And the tech needs to support us quietly throughout all of those six moments. I think sometimes things it become so noisy, right? We're all talking about AI and tech stack, and we we start to to create um more of a panic than we should. And so maybe this is why I talk a lot about us focusing on the hospitality when you mentioned, I agree, issues and problems when it comes to um cleaning. You also said processes though. So I would turn to tech. How can AI or automation? Because for me that's different, help us become better there. So when automated tickets, service calls, which are happening instantly. That is part of the brain, not the engine, which is the PMS.

SPEAKER_03

Are you using any other uh software like operation software, revenue management software?

SPEAKER_01

No, no. Um, again, our mindset here was we don't need a ton of tech partners, we need solid brands who can grow as fast as we want to scale, basically, right? So our portfolio is diverse. Hotels and vacation rentals. For us, it was really important that our guest experience remains the same because you can have a guest who is looking for a holiday with his family and friends, and so a vacation rental fits. That same guest might come on a business trip, and now a hotel suits him better as well. So this is why we talk about hospitality rather than just a property manager just handling vacation rentals.

SPEAKER_03

Um, yeah, that is our so that's an interesting aspect that you mentioned here, Saab, talking about, you know, they want you want a partner that can grow as fast as you want to grow. And I think that's one of the great segues for Chris. It's probably one of the core reasons, but probably many others for land folk to think about hey, we want to be AI native entirely and build the entire tech stack uh in-house without using any third-party tech. Uh, what is the main reason behind that?

SPEAKER_04

You're mute.

SPEAKER_02

All right, let me try. I think I think the main reason is that we me and my technical co-founder previously sold a company to Airbnb where we built everything from scratch, and we got to see from Airbnb and Inside Out how it works if you have everything you uh defined yourself. Um so we quickly decided that we need to build our own internal operations tooling, our own marketplace, so that we could provide the best experience to the host, but also to the guests, but also to ourselves, uh, for our own employees. Uh and the only way that we can be in control is if we build it ourselves. Then we're not beholden by someone else's APIs or user interfaces. So we just started out building out anything that we needed. Uh I think there's a saying that you should do whatever makes your beer taste better. Uh, anything that doesn't make your beer taste better, you should just buy. Uh, so of course, we don't build everything ourselves because that wouldn't make sense, but we build everything that makes uh the core user journey better. Uh and that's distribution, that's internal operations, that's like everything about that whole journey for vacation rentals. That is something that we build ourselves. And in a in a world where AI is able to do more and more, it turned out to be a very beneficial thing to have done because then it has all the tools uh at its fingertips suddenly. Because anything that we can do in our internal systems, an AI can do as well. Uh, so we're basically just layering the AI on top of all the tools that we have already built for ourselves.

SPEAKER_03

Yeah, I like that. And and especially the core strategy of the foundation is driven through pre for through past experiences without having to go down into the channel of, hey, what kind of technology are we going to use? We just do that. And you know, a lot of people sort of start using stuff and then they start building things because they think they can do it better. But you you actually had your core strategy already set from previous experience that we want to have everything in-house, and that's basically how we're gonna go about this.

SPEAKER_02

Well, yeah, we we paid a lot of startup tax, uh figuring out what not to do and what to do.

SPEAKER_03

Yeah, I'm sure, like everybody else, and we sure like the Danish beer as well. So that's interesting saying uh too. So, Arnon, how does that apply to RBO? I think where did you come from a purely AI native uh product strategy? Or um have you got everything in-house or using any other type of tools?

SPEAKER_00

I think if you zoom out, we see that the cost of coding is going towards zero. Right now, Simon, if you open cloud codes and you start to prompt uh what do you want to build step by step, let's say in one, two days you can have like a basics prototyping working. Of course, not gonna be a world-class software, but it's gonna work. So the question is shift from can we build it to should we build it? Because this is gonna add true value to the vision where the company wants to go. And then bringing back to RBO, since our growth strategy is based more in acquisition, then we need to build the abstraction layer so other agents and tools can connect to RBO and operate with the full context. Because that's the mode as well in the AI. It's not the software anymore. Everyone can build software right now with cloud code. But the context itself on why and how the decisions are made, this is essential for agents, operates, and that's where we're focusing now.

SPEAKER_03

Yeah, that's super interesting. And at my uh sweet age of 55, I become a programmer, and I never thought I would ever in my entire life, but you are so true that I can do things that even uh make my children go, hey, my dad is now an engineer, uh, as well, which which is just mind-boggling and how fast it goes as well as well.

SPEAKER_00

An analogy with Excel. Uh, I believe that uh building software is gonna be something similar to Excel. Everyone's gonna know the basics of Excel, but if you want to go pro, then you are in VBA macro level of Excel. That's gonna be the programmers. But everyone's gonna know how to do the basics of programming. It's the Excel.

SPEAKER_01

Yeah, my daughter is a fan of Lovable. So she has vibe coded a couple of really cool projects from like her obsession with Wicked, the movie, all the way up to homework plans. So I think these are the skills that kids, not just adults, should be learning.

SPEAKER_03

So interesting to see how we get when are we gonna be tired of that and when the human aspect is coming back into the conversation. But I think we're not too far away from that either, and realizing how important the human interaction is going to be. And we can, you know, we've seen a lot of predictions, and one of which was Sakas, uh former OpenAI guy who um was talking at keynote at Darm Conference in Florida, and he reckons that we're all gonna spend so much more time together and and be interacting again while the machine works for us. So let's see how this bowls out. Um, one thing I want to definitely touch on is distribution um and control. And one of my opening questions was, you know, is the OTA are we going to be dependent uh or not? And I think a lot of the PMCs out there uh are talking about what is and Arnold uh answered uh in his view that it's gonna be defined by how is the future traveler going to search and how is that behavior going to change. At the end of the day, I don't think that uh a ChatGPT or a claude will become the merchant of record when you start searching for a property. Um and I have also tested uh landfolk's uh search function, which is absolutely insane. I I love that one. Uh, and it just gives me incredible results, just you know, sitting down at a family table, and everybody can scream in what they want and uh and then how it's uh actually selecting property. So the future search is still uh, you know, we still don't know. Everybody says you need to be you need to be able to show up in these LLMs somewhere so you can be found while the people search, but then ultimately search and booking and transact is still out of there to be seen. Most operators are still heavily dependent on OTAs. And I obviously had CEOs of OTAs on the podcast as well, and clearly they're not gonna give up that fight without a battle, and they all have their relationship with Claude and O and ChatGBT as well, so they're making sure they sit on top of these results uh themselves. So I would like to talk about your approach, and and I I want to start um with Chris first. Um and because I know through the interview with with Christian as well, is that Landfolk has definitely strategically uh an ambition to build a very strong uh brand themselves and not just being solely dependent on third-party distribution. While my argument is having been in this industry for over 20 years, there hasn't been any meaningful consumer-facing brand in this industry ever, ever, ever before because it's so fragmented with the exception of some specific geographical operators. And I guess Denmark is a perfect example with Down Centers, Tolgestrand, you know, the classical operators, uh, Novasol, who've been in this industry forever in Denmark, and they sort of cut the market themselves. But in in general, nobody has cracked the nut of building a direct brand and reducing dependencies on OTAs. How does land folk think about that? And what is your strategy? Whatever you can share.

SPEAKER_02

I think, yeah, I mean, as you said, I think. Brand is very important to us, and we started out building the marketplace. But we also saw that to be able to fulfill the needs of our hosts and provide even better services for them, we needed OTAs to give more bookings to them. So, I mean, we see OTAs as an acquisition channel for us. And then there's something going through the OTAs, something is going directly through our platform, but we have all the possibilities there. And I see I see AI as another potential distribution channel. Whether that's going to be through OpenAI or Cloud or Gemini or who knows, maybe someday Siri will be able to do more than just set an alarm. But it might be that that's going to be a distribution channel as well, you know. I just think that I think the important thing for us is to have the best relation to the homeowner. So that when our homes get booked, it doesn't matter if it's directly or through an OCA or through Siri or whatever. We just need to have the best relations and own that that connection to our homeowners. So we need to best build the best tooling and have a very, very good relationship.

SPEAKER_03

Excellent. Love that. That's the only asset we have in the industry's relationships. Our homeowners, let's not forget that. And uh, you know, that's the only asset this industry has is trust. So, uh Arnon, I think you know, with your brand called RBO, it could be very misleading that you're heavily dependent on a on a channel out of San Francisco and going to be like that in the future. What is the approach of RBO over time in relation to the dependencies on OTAs and building your own distribution as well?

SPEAKER_00

We wanted to build tools that agents can access in the cloud. What it means, for example, for a website, something very simple that companies are doing nowadays is adding a TXT file explaining what your company does, um how it operates, so ChatGPT or cloud can read these TXT files and can rank you higher in when you're prompting your questions there. So we believe that yes, OTAs they are still the gatekeeper, let's say, but that advantage is not gonna last long because depending on user behavior, and we are saying that yes, users still go into OTAs directly because that's how we learned. But what about the new generation that's gonna be born and raised with ChatGPT and Cloud? They are still going to these OTAs or gonna ask directly. So it's a matter of time. It's a matter of time to go directly to those tools. Um, and then we don't know how it's gonna be the power of OTAs as gatekeepers anymore.

SPEAKER_03

Sab, you're a local brand in a smaller market, obviously. So you could look at that, totally differentiate, and say, look, as long as our houses are booked and we have good occupancies and our cost of acquisition and customer acquisition costs are under control, we we don't care too much. Where is your thought on the distribution side, dependencies and OTAs and building your own brand?

SPEAKER_01

So there's the discovery side, right, which is happening and it's something that we are learning, how AI can help us get discovered. But then for me, again, from a brand perspective, it's also about the the recognize, the refer, the return. So guests are spending like 30 minutes, an hour on the OTA channel, but then they're spending four days in our properties. What an opportunity to build memory, loyalty, and belonging there so that they can come again. And I agree with Chris, great for a first guest to come through an OTA channel, fantastic visibility with OTA, but then it's our responsibility to transform that guest into a Zing guest, basically. And there we're doing a lot of work with AI. We have a playbook, which is really cool, where we can identify if the guest is here for travel or maybe it's an anniversary or something special. And then the agentic AI can send a reminder oh, anniversary is coming up in eight weeks, for example, or is it time for another break? And the AI is doing that work for us again, for the humans to show up. So when it comes to AI, two things: discovery, which is something really still new for us. But when it comes to returning guests and how to build that loyalty, we are more confident now with AI than we were three months ago.

SPEAKER_03

Interesting. And you just gave uh a great next point uh that I wanted to touch on before we wrap this up. And interesting enough, AI agents would give me uh a reminder that the anniversary is coming up, which uh should save me a lot of grief as well. Um, anyway, so so let's talk about the discovery, and that's a you know, coming from the marketing side, you already touched on that now that AI uh you add AI into the equation of distribution and and it discover it the discovery shifts from research to conversation to AI-driven recommendations. What does that mean for your business if that is exactly happening? And Chris, you have built this function on your on your own web page already, but where does that how would that impact uh uh your business?

SPEAKER_02

Um yeah, I think that's a very good question. Uh I mean we tried to build it ourselves because we thought that that would be a better experience for our guests. Uh, but I don't see, I mean, I I think I think OpenAI and Anthropic could could do the same and build it into their systems. I mean, they could just use our tool if they wanted. Um But I think if we just own the uh the relationship and have uh very, very good content and very good images, um, that discoverability is just going to be better on on other platforms. Uh they're gonna benefit from that for having that direct relation and good content.

SPEAKER_03

Okay, so I want to quickly move on about scaling, uh, because that's an important thing that we need to think about, especially in today's age as well. Uh, you know, we we sort of took an approach to say, hey, no, we could just maximize the business that we have, but I think in terms of raising costs and everything else around us, I think we're doomed to scale uh regardless uh to make a meaningful impact in the business and be economically viable as well. What do you think breaks first in terms of scaling? Is it technology? Is it operation? Is it supply quality? Or where are the real constraints? Thinking about scale. And Arnon, you're a you're part of a company that acquires a lot of business as an acquisition strategy. So, what do you think uh breaks first in your scale strategy?

SPEAKER_00

Human capital, right? Um humans, um training them, making them to operate at the level that we expect. I think that's that's still the bottleneck. Especially, for example, coming to cleaning, like what is expected, as well as customer support, of course. Um, let's say 80-85% is automated, but this 15% left, um, it's where you can create the wow moment in the customer support. So humans are still the bottleneck when it comes to scaling.

SPEAKER_02

Chris, I think I agree on that. I think I mean we have a very high bar for the quality, and we interview all hosts who apply to keep you part of our platform. So keep scaling that quality is is difficult and it involves humans. And we need humans in that process. But we can we can scale uh or automate certain parts of the process.

SPEAKER_03

Sab, what breaks first when you scale?

SPEAKER_01

I think what we have learned is quality of foundations is the most important thing. Our processes. Because uh that is what starts revealing the cracks. I think scaling can actually feel calmer and you can add more keys and you can create more time for hospitality if first you go back to the drawing board and see what your processes are like. For me, all about processes, playbooks.

SPEAKER_03

Interesting, all about playbook processes and potentially standards as well. None of you said uh regulation, so that's very encouraging. Um, you know, when you scale, maybe regulation can also be a factor of break, but we could considering the time of what our discussion, I want to talk uh and wrap this up about the future of property management. So, you know, being AI native on two fronts here, and and one uh being very adaptive to optimize the business. On the other, let's look ahead a little bit and and how does the business look like in in five years? And and what does a property manager company in view of you when we talk about November to April in six months, and we're talking five years, so that's light years away uh in our uh way of thinking uh when it comes to AI. But let's try to do that. What does a property management company look like in five years from today? Who wants to go first?

SPEAKER_00

You can go. Um, as I mentioned in the beginning, uh, when you ask about the PMS, I see as uh the invisible layer is not gonna be um all the front end with dashboards that you have to do 10 clicks in order to get what you want. So invisible PMS, uh you can connect directly to a terminal. It's beautiful for agents to execute the job. And in the end, the real advantage of uh property management company is gonna be the context as compounding assets, um ontology to build structures for your data. In the end, if in five years AI is gonna be the biggest lever and AI is built on the top of data, you need to have decisions tracers on how this data is built. That's the call ontology. So in the end, in having context is gonna be the differentiator for the companies that are gonna survive and the companies that will not in the AI world. SAP.

SPEAKER_01

In my crystal ball, in 60 months time, we're not gonna be property managers, we'll be hospitality architects, we'll be curating experiences, we won't be doing busy work, so no distribution, no tickets, no WhatsApp. Everything is like fully automized, systemized, we're focused on how guests feel. We're crafting memories, tech, well, the big players are keeping everything running smoothly in the background, and then we're really calling ourselves hospitality brands.

SPEAKER_03

Interesting. Uh Chris.

SPEAKER_02

I think it's an interesting question. I mean, five years or so, so far out into the future. I mean, but yeah, I agree. I think it's uh the company knowledge that you build up, the the human connections, the the relationships to your to your hosts that you build up. That is that is your core, that is what you have. Uh, and all the tools that you just uh make available to all AI agents. So you might have a marketplace where you can like distribute your houses through, but it might also just be an AI to generate a UI for whatever guest wants to book something for you. Uh but who knows?

SPEAKER_01

It's like can I just add will be learning about guest psychology, sleep, why bedding is important. We'll see how we can design hospitality based on senses, just because we won't be firefighting. Positive.

SPEAKER_03

Interesting. So to wrap this up, that's uh uh you know, that's how you differentiated answers you get from a CTO, from a head of product, and from somebody who comes uh from a different angle. So that's been uh it's more with the marketing and the guests in in focus. So that's been uh amazing. So what's clear from this discussion is that AI native is not one model. It's a spectrum from where I see it, with very different approaches to technology, operations, and then potentially scale. And the companies that figure this out first will likely define the next generation of hospitality. So I want to thank Seb, Arnon, and Chris for participating on this podcast. Thank you all for joining me today on STR Global Unlocked and see you at the next session. Thank you very much.