#2 - Hospitality Is Now a Technology Business

Mike, thanks so much for making time today.

My pleasure.

I'm so happy to be in a CitizenM Amsterdam pretty, uh, pretty enjoyable lobbies these, so Yeah.

Yeah, it's really nice.

Uh, it's got some really nice lighting and I think the first time I was in a CitizenM lobby was in New York in 2017 or 16.

Oh yeah.

And Laura and I were there and we just like, oh, this is cool.

Nice.

This is cool.

This is different.

Yeah.

And uh, I think CitizenM has kind of maintained that flavour throughout the years.

Yeah, I think, uh, all our lobbies are distinctive but also very branded.

So you see one, you'll recognize them anywhere in the world.

That's true.

And the concept of affordable luxury is that we really invest in the lobby and, and you see us a lot on Instagram, so it must work pretty good.

Yeah.

And as, as much as I can spend 30 minutes discussing the color red.

That's okay.

I wanna just, because the first podcast I had with Paolo from Staycity Oh yeah.

He was just like, oh, I wanna hear from CitizenM.

Oh really?

And I'm so happy to have you here now and ask you straight.

So how is CitizenM approaching AI today?

Yeah because I read a few articles about you and you mentioned that.

AI, would have the number one impact on hotel operations in the next 10 to 15 years.

And that's quite a big statement, so I'd like to break that down and hear your thoughts on, what does that mean?

Yeah.

So obviously, um, we have to talk about AI because clearly it's just dominating the whole world at this point.

Um, and then so let's talk about it, you know, at the top level and then we'll bring it into the hospitality space.

So at the top level, it's for me, it is now what I'm calling Greenfield 2.0.

So, so it has changed, it's resetting everything similar to when, let's say the dot com arrived.

Okay.

And yeah, that was a bubble eventually, but it still reset and changed everything.

And AI, in my view, is bigger than the dot com.

Uh, and we are thinking about it as in like Greenfield 2.0., And what does that mean?

2.0., And what does that mean?

Well, that means you have to start thinking like an AI native.

And then you'll see at the moment there's a huge divergence continuing on where people are- you see a lot of reports that the use cases are not being met, the ROIs are not being met, and yet the 300 million, 400 million, uh, AI users continue to grow and, and use the product.

Yeah, 800 million.

Yeah.

So there's.

So that you can barely keep up with the number, right, because, and so clearly the adoption is happening and, and at this point, you know, the ROI is reportedly not arriving.

And I, I just think that's a, um, couple of things.

I think potentially trying to put AI as the holy grail to solve the biggest and uglies problems is probably still a difficult thing to do, where actually there's probably a whole lot of low lying fruit that, that you can basically go after.

And, and we certainly have done that so we'll get to cover that.

And then secondly, the space is so dynamic and so has so much volatility at the moment.

That's a very difficult thing to, invest in as a, let's say, making a business, making investment in AI.

Because some of the ones today will be bought by some of the big guys tomorrow.

Some of the ones that are point to point today, um, will be just, their functionality will just disappear into a platform.

Someone will just come out and release a platform addition and that whole product will disappear.

So how do you navigate your way through, the volatility that's in the AI landscape?

That's a, that's, that's quite a complex problem.

And then lastly, we're drowning in AI.

AI promises, and I just don't think at the moment anyone's really spent a lot of time thinking about the security and privacy elements that needs to occur to basically support it, right?

So for an enterprise.

We certainly at CitizenM have a history of being very innovative and, and my role is to continue to grow that innovation, but at the same time, I need a privacy and security wrap around that when it goes to production.

And at the moment, that's a huge impediment to the stuff that basically has to go right.

The GDPR, the PII.

What are you enabling the AI to do?

How much power has it got?

How do you control the changes that it's got?

How do you back it out?

So you know, there's a lot of things to consider about rolling it out.

And so that's complexity and that's, you know, that makes CTOs and CIOs have, a pretty big role these days.

And, and as you know, you can't do a business without tech these days.

So my final view is that the CIOs and CTOs will essentially, um, are the big callers over the next five to 10 years.

The CIOs and the CTOs will move into the CEO role.

Mm-hmm.

I believe that the business.

It's not as complex as the tech and therefore you can't, and you can't do anything without the tech.

So I think you're gonna see more and more that the tech is just going to take over the business.

Wow.

That's an interesting take.

Um, 'cause it's gotta go one way or another.

So historically it's been the business with some tech.

Well, I think in the future it's going to be some tech with some business, because I think that, uh, it's the only way forward.

So that's the macro.

And then if we come down into AI currently, yeah, we have- so CitizenM has, two north stars for how we approach this so that we just have an idea and we don't drown in the tech.

So the first one is personalization at scale.

So for the guest, I wanna be able to do personalization at scale for you with AI.

And what does that mean?

Well, that means.

Adding value or removing friction.

So at the end of your journey, for example, and I noticed that you stay with us every two or three weeks, I'll get you that rebooking offer for the same room.

In CitizenM for the next two weeks with your discount and basically knowing that also, if you come to that booking, we know what temperature you like it at and we probably even can get the ambassadors to offer you that, uh, coffee 'cause it's already on your folio.

So we know it's your favorite drink.

It's a small example, but essentially.

Adding value.

And the other one is just removing friction.

So making sure that the invoice is perfect, that it's set up for your company, for example, and not you personally.

All that personalization.

So that's personalization at scale and we continue to use that as the way to, look at the guest and look at the guest for the total guest journey.

And then the second one is automation at scale.

So for the employees, for the ambassadors.

The more time I or the team get to create the space for the ambassadors to have more time with you and less time on a computer, the higher our ratings are, the higher our NPS is, the higher everything is.

Right?

So that part of magic sauce is pretty easy.

More time with the ambassador and the guest equals happy guest.

So our job is to then take that automation.

So we have a, currently we have a product called Mamba, which is the, which is the app for the ambassadors.

And if you wanna do a room move, you can walk up to the ambassador and say, you know, something's wrong with the room, we'll move you from 1, 2, 3 to 4, 5, 6 for the ambassador.

Now that's on the phone.

Click click click done.

Whereas previously they'd have to go to the desk and essentially do a whole lot of extra steps.

The advantage of that also is the ambassador doesn't need to know what system.

So even if that's using three systems, that's all just hidden now for, from the ambassador.

And if we decide to change out one of those systems or AI it, yeah, that's all agnostic.

So that's, so I like that from a tech point of view.

Yeah.

Yeah.

So those are the first two sort of, uh, or at least how we, how we approach it.

And then on the, let's talk about the employee automation 'cause I think at the moment that's the use case we've gone after in a big way.

'cause we are very careful with the personalization at scale by putting AI directly against the customer.

Does it have the tone who's, who's monitoring it?

There's a lot of things there that, that we are being, it goes wrong.

It could be, uh, if it goes wrong, it's gonna damage your brand directly.

And, and, and that, you know, that thing is talking for you.

And you saw the example where someone got a free car from the chat bot and that got upheld in the courts.

So it's speaking for the company right.

So we are going slowly on the AI directly to the customer, but fast on the automation at scale for the employee.

So we um, currently been running that for at least 12 or 18 months.

So we have two digital employees at the moment.

So that's how we positioned this.

So our first digital employee was called Robbie, and it's very clever to make it a digital employee, to put it, to personify it, because it's actually a complex piece of AI.

But when I talked to the CFO, I go, Hey, this is Robbie, and he never gets sick, he doesn't need a pension.

Basically he doesn't go on holiday and he's, 80% more accurate than his, uh, coworkers.

And he goes, oh, where can I sign?

So the packaging of your digital employee that way - we found that very successful.

And then secondly, our employee called Robbie.

It sounds boring, but actually Robbie can take a invoice from our inbox of which we get, 10,000 a month or something.

He will find that invoice.

Robbie goes away, finds the supplier, and our system finds the purchase order that's related to that invoice.

Checks that the purchase order and the invoice and the goods receipt are all basically correct.

So three way match and then we'll actually pay that invoice.

So that's a real world example.

And currently Robbie's doing about 70%, uh, of our invoices and where it's not, that's on us, it's usually data or some other process, some mismatch between, yeah, it's not, yeah, that's right.

It's, it's the quality of our data or the quality of the system or some other things.

Yeah.

So that's, yeah.

You got a question on that?

Yeah.

How did you build Robbie?

So where did that idea originate from?

Did you build it yourself?

Did you buy a software in the market?

Yep.

Can you talk a little bit about more details about the tech behind Robbie?

Yeah, so Robbie is, um, so we engaged a company.

So in the Netherlands there's a company called Free Day AI.

There's a few companies that are doing this.

We also use another one called api, APHY.

But we engaged free day, uh, and we were very clear about the automation at scale and the personalization at scale.

And we agreed to start on the automation at scale and the use case that had, let's say the Goldilocks of complex enough and rewarding enough but not crazy.

Yeah.

Was was essentially this three-way match auto paying the, the invoice and, and it also just had a huge transaction load and it was heavy on finance.

Yeah.

And it's, well everyone knows it's boring as bat shit, paying invoices all day.

So it's also gives them time to do more value added, processes.

Right?

So we went through partnering with, free day and, iterated our way through to Robbie, and that's been in running now for, I think about 18 months.

Amazing.

So if you were to approach AI, you just kind of take these use cases one at a time and say, okay, this is invoice processing.

Here's probably payroll questions maybe from employees.

And this is something else.

And kind of rank order it in terms of what's the most annoying and what's the most, yeah.

So you want to go for, um, essentially the way we do it is, we do a intake.

So we work on an intake with the business, and then that intake we structure around what is the frequency of the, of the transactions and what is the complexity of the transactions and what is the value of the transactions, right?

So triangulating those three should give you an answer.

So in this case, the invoice all the way through to payment, it's high frequency, it's high value, and it's reasonably complex.

And then you, you might have to rank that and start to, so you want a way to rank it so that you can, 'cause you've only got limited resources, so you rank it and you then have to use that to find you one, the business will lead the way normally and say, yeah, this is the one that we wanna, let's do this.

Is there anything you built in-house with your own team where you're just like, oh, this is so annoying, because you mentioned in the past you would build something in-house if it kind of helps position, CitizenM, as a innovator or kind of strategically ahead in terms of your competition.

Yeah, but you might buy generic stuff.

Is there something you've done in-house that, you would be able to talk about?

So we do a lot of innovation in house.

So for example, that proof of concept and what we call POCing stuff, we're doing that a lot.

So that and, for example, using Zapier and looking at the AI tooling and, but as an example, the automated room check that we wrote.

Yeah.

So every, which is unique, I think, to hotels.

So at the moment when you check out of CitizenM.

We spin up ARC, which is the automated room check, and it automatically goes in and it checks- is the iPad online?

Is it responsive?

Is the TV online?

Is it responsive?

Is the room online?

Is it responsive?

Essentially, is everything about that room for that next guest experience working?

Yes or no?

If it's not.

We put it out of service, we raise a ticket for the problem.

So if the HDMI is not connected, which we can tell, we'll do all of that pre-work to make sure the next guest does not have that broken promise.

So, at the moment, ARC's running for 40 hotels, 9,000 rooms, and, and it has really driven down, firstly the amount of tickets, but also it's driven up the customer satisfaction.

'cause we now aren't relying on the customers telling us something's broken.

We are already finding and fixing it.

That is a perfect candidate for us to start playing with AI.

So we would put AI over the top of that to start basically doing far more root cause analysis and start going, well wait a minute.

There's, you know, 15 TVs are reporting the same error here, so we think that that's related.

So those are the types of areas we would.

We would essentially, look at AI across those types of areas, and then the decision to do it ourselves or to get a third party to do it will simply depend on, cost, complexity and time.

We will choose to do a lot of stuff ourselves if we can.

'cause we have our own tech stack and some of the crown jewels are all ours and essentially built by us.

And so we want to have that flexibility, but by default, if it's big, and we've done the POC and then we do a pilot, which is one hotel, we will probably take one of our core partners and, and help get up because they also have to maintain and run it Yeah.

And do the support for it.

That's also painful.

Yeah.

Just, yeah.

And so that's not our core business.

Huh?

Our core business is essentially to ensure that we're just driving that guest journey and that guest satisfaction.

Yeah.

And you, you also have a lot of RPA we do robotic process automation, like 30 odd processes from what I hear.

So how does that work in the world of AI?

Are you kind of looking at all of them again and saying, I need to redo them with AI?

Is it just like they work, they're fine.

Leave them alone.

Yeah.

Probably six years ago now, when I first started RPA was before AI.

And so RPA, we got that going.

There's a company called Aphy that we, that we used and still use.

And indeed we had the, at one point, about 60 or 80 RPAs running.

'cause they were so effective for us.

When we changed PMS systems because of all the functionality that came in with Apaleo, which we knew we wanted, that took about half the RPAs out 'cause we got that as organic functionality.

So that's a good use of using the RPAs as a temporary measure, a new stack.

And then, subsequent to that, you know, Aphy's been retooling, to move more into AI.

And we also have basically moved some of that functionality into ai, either ourselves or using a different partner, but it's already shrinking down.

So probably at the moment we've still probably only got 15 or 20.

But with the Marriot integration, 'cause some of their stuff is not the most modern, we'll probably still have some RPA that'll be, that'll be fit for purpose for that.

And I know there's a, there's a big Marriott deal that's just happened big.

Yep.

You probably have a lot of work to do right now, so I appreciate you taking some time to talk with us.

It's, would you mind telling us what that looks like behind the scenes from a data point of view and from a technology point of view.

Yeah.

So obviously it was a huge deal and it was announced, you know, whatever, six months ago.

And we are gonna get access.

So, so from a business point of view, getting access to the 230 million Bonvoy members and the whole Marriott infrastructure and economy is huge.

And it will be huge for us.

So what does it mean high level?

So if you think about the guest journey being pre-book, book, stay- essentially the pre-book and book is now going to be leaning into Marriott.

So that all those capabilities will be run by them.

The website, the availability, the OTAs, it's all essentially the pre-book and book.

It's gonna be on Marriott and they have massive economies, and massive volumes there.

And, we expect to see some, you know, big improvements from that.

What we have to do is essentially turn on a tap, which is, you are right, it's the end of this month slash November.

So we're under, um, a lot of pressure.

The teams are working hard and, and it's looking good, but it's a big migration coming up end of October, November, and that's essentially then giving the reservation availability to the Marriott system and making sure that all our systems are turned on, that the kiosk, when you arrive, you can arrive as a Bonvoy member and abra-ca-dabra knows you are Bonvoy and

it knows that you get a special room and you get a free coffee, whatever.

Yeah.

Uh, and that we can do all the point system.

So obviously Marriott is all about points and everyone's just super excited about points and we had to make, so a lot of the work is around the integration of that point system and all of that.

So the arrival right from the kiosk and then right through to checkout and also the use of the Marriott systems. So they've got at least half a dozen of their systems that we need to use.

To, I don't know, check a guest out.

They have a system called GXP.

So their GXP system is essentially the CRM system.

So we have a lot of adoption to do, to integrate essentially into our journey, parts of their journey that are mandatory.

So it's a lot of work.

It's quite a bit of change management.

And I think the most important thing is it's very important to us not to wreck the secret sauce.

So, so suddenly putting a whole lot of friction into the guest arrival when the kiosk can be really snappy.

Yeah.

There's a balance there.

And we have to make sure we guard the guest experience, when you're dealing with such a big corporate who have a lot of requirements.

How do you guard the innovation DNA as well?

Yeah, because if you don't take care , then you could be just worrying about integrations all day long.

And then, you've got all these AI initiatives ongoing, so how are you thinking about that?

Yeah, so we, um, we were just talking about it last week.

So we went to see Marriott last week and we had a good session with them, and one of them is the innovation approach and strategy.

We believe we still have a lot of firepower in what we call the stay part of the guest journey.

So from the arrival to the, to the checkout.

Essentially our 40 hotels are a perfect little pilot.

We're a very small pilot for Marriott.

So they see the use of CitizenM in that space and also because we are still so, tightly integrated as a unit, our speed is very high for these types of innovations.

So the working process will probably be that we will take the lead, still do POC, pilot and and roll out, and then those types of things will then basically be, yeah, showcased up into Marriott.

So for example, they're super excited about the kiosk.

That's kind of their number one thing.

They don't have a kiosk and they're very interested in our kiosk.

And at the moment that's kind of got the, so in, in a way, a lot of stuff that they've purchased for them is innovation.

So right now, without us doing anything, we could probably sit back for a couple of years because we've got a lot of stuff that for them is, is new.

Kiosk is number one.

When I was doing research for this interview, I was just like, how many things are you guys doing?

Like the loyalty app, the kiosk.

Yeah, a lot.

The room tablets and all these things are huge projects.

Can you talk a little bit about the kiosk and tell us what's so special about it?

Yeah, so the kiosk, well, two things.

So firstly, the, the original kiosk was here, you know, 15 years ago or whatever at Schiphol.

So at that point, the industry literally said, you guys are crazy.

No guest is gonna check themselves in.

It's insane.

So, you know, full credit that they went on with that.

Recently, we've essentially, refreshed all of the kiosks.

So it's now basically running off a iPad.

It's running on Flutter, which is, just super agnostic for the hardware and the screen size and all that stuff so you can run Android and so super happy to have that.

And, it's being, produced and run for us by Rock Company called Rock Soft, which is a Netherlands company and they've been with us for a long time and they do the in room iPads as well, and Mamba.

So it's very useful because there's an example where we consolidate those products into one core partner and we can get a lot of mileage out of that relationship and it's, you know, easy for us to run.

And they understand us very well across those three products.

Yeah.

The kiosk is special because, it just allows us to do such a easy arrival experience for the guest.

It still does all the localizations, so where you have to fricking get passports and all the ID and all that kind of stuff.

It can cope with that.

And then, the guest also has, the UX.

We put a lot of work into the UX, so it's a really easy and fast flow.

And then also on the checkout, it's a really easy and fast checkout.

You know, in your invoice it's clear, get your invoice on a QR code.

You also get it on a follow up email link.

So the kiosk is just taking the, you know, just the bulk of our transactions and it just works.

It's something it.

And, you know, and we, we see, we know the speed.

Everything's recorded.

We have all the data.

So we know the in ins and outs.

We know where the stuff fails.

We know, 96% of the checkouts are just within a minute.

Yeah.

So we're super happy about that.

And that allows the ambassadors again, to basically, so they, they're there.

Welcome you, to facilitate your check-in.

You do the check-in, they help you.

But it helps to then foster that, uh, discussion and that relationship rather than them standing there.

Yeah.

Like this, while you wonder what's going on.

Right.

Yeah.

So, so that, that is the very beginning of the, we call it the moments of truth, which is Casper, his term, our, um, CPO.

And you know, those moments of truth, we really have to, protect those and we really have to build on those.

So the kiosk helps us with that moment of truth, which is called arrival.

Right.

So you arrival, like you today, you're like, oh wow.

Yeah yeah.

That arrival, that's special.

And especially if it's the first time, you know, we really wanna make sure that that arrival for you is the same for you as everybody else.

That's nice.

Again, you mentioned when you first got the kiosk 15 years ago.

Yeah.

It was something that no one had ever seen before.

No.

It's pretty normal right now.

Yeah.

So how do you kind of tap into changes in trends?

Yeah.

And how do you prevent yourself from falling for fake trends?

So how do you like defend against like what might be hype and that disappears in a year from now?

Yeah.

Versus something that might be a durable, lasting change.

Yeah, well firstly you shouldn't rely on just one, let's call it one entry point.

So we already had the app, which we launched, you know, a couple, two, three years ago.

The app allows you to check in, the plane's landed at the airport.

You can check in and basically if you keep your CitizenM card in your wallet like I do, I check in, I touch my key card, my room's all done.

I can do all my meetings for the day and just come back and carry on.

So we constantly evolve the guest journey.

And in this case, we knew we couldn't have a kiosk in the plane, but obviously the app was the logical place to then can we, can we get you to check in multiple ways?

Yeah.

So by diversifying that, you've de-risked it.

And then lastly, because we do POC pilot rollout, so we do proof of concept one hotel and then rollout- that does tend to help us stop, or see things that are gonna basically not be fit for purpose.

Right.

Right.

And, but you don't have NFTs, for example.

Do you?

Uh, did you try that?

I'm just curious.

Yeah, we, yeah, we've been, we've been, I'm a, I'm a big fan of blockchain and I wanna.

Distinguish that discussion from crypto.

But I'm big, big, no, me too.

I'm a big fan as well.

But I just wanna understand how do you approach that new technology and validate if it's ready for the mass market adoption or not?

Yeah, it's not.

So we see the digital wallet, the EU digital wallet is here, that's ready and that's good stuff.

So that EU digital wallet, you know, we will and are already looking at that.

And I just yesterday talked about doing a POC around that.

So you have to keep an eye on the big things as well.

And when you do, uh, proof concepts.

Yeah.

How are you measuring ROI?

And how are you looking at payback periods in terms of ROI?

Yep.

So are you just going, because you, you guys own most of your buildings, does that impact your decision?

Because you own the buildings, are you able to have a longer term vision?

Yep.

In terms of ripping and replacing entire PMS system and building the whole ecosystem of applications on top of that, how are you measuring these new initiatives?

What's the timeframe?

Yeah, so I realize now that it is unique because we do plan, build, and run.

So we are unique in that hospitality space and the way we approach things.

So I can determine what goes into the build that we're just doing in London, Olympia.

Right from day one.

And that's super powerful 'cause then we can make sure the strategy's how we want it.

The second piece is the ROI.

So there's two things.

The POC is a, firstly a technical POC.

So we only ever start with a technical POC.

Does it do what it's supposed to do?

Is the supplier capable of doing it?

Is it just slide wear and vaporware?

It does the thing actually technically work for us?

And if that fails, that would game over.

And then that leads into essentially your business POC.

So it's like, okay, what would the business drivers be and how would we measure that?

And we can take that a fair way normally.

Um, and at some point you have to do a pilot.

So at some point the proof is in the pudding.

You do a pilot and that one hotel has to deliver it.

And either we refine the pilot.

Or we kill it, right?

So the pilot's either refinable or you kill it.

'Cause it basically, you do one, but it's not, it's not gonna cut it.

And roughly how many pilots are you running a year?

It's probably anything between, depending on which year and what's going on.

It could be anywhere from three to 12, right?

Yeah.

It's a lot of action.

Yeah.

So you have to be careful to try, you know, we only have x amount of things we can do a year.

And so that's the opportunity cost.

So you have to make sure you make the business aware they can only put a certain amount of chips on the table.

And that's, that's always a challenge because, you know, there's always more stuff to do than what we can do.

And I also know you've spoken in the past about allowing anyone on your team Yeah.

Kind of just bring in something as a POC.

Can you speak a little bit about that model where you're just kind of sourcing ideas from your team?

Yeah.

So the philosophy certainly for me very much is freedom within a framework.

So we set the framework up so you understand the north stars.

So we set the north Stars personalization at scale, automation at scale.

We set the framework up.

We have a number of gates, and we also make it clear that you do POC pilot rollout and anyone can POC anything out.

And I'm just talking outside production here, just to be really clear.

No POCing in production, just to be, uh, okay.

So we're all clear on that.

So yeah, they all are empowered to basically put something on the team chat, come to me, come to one of the seniors, and go, Hey.

I just ran this, uh, NFT and, and checked this out.

And so, that is a really important innovation element.

It's also, there's a collaboration piece of that where they might do that together with someone from the business or someone from the hotels, and they can have quite impressive POCs that they'll basically bring up.

And that's the beginning of it.

So then we are like, okay, what is it and what's it doing?

And what are we proving?

And essentially, you know, is it adding value?

Is it removing friction?

You know, all those kinds of things, and they come from everywhere.

And I think once people realize that they're empowered, that I want them to go and do that.

Even if they need training to go and do that I'm fine with it.

Amazing.

And I know we are running short on time here, but I I have one question that I really wanted to ask you, and this is a little technical, so that's No, no, go for it.

Technical.

Hopefully I can answer it.

Apaleo just released this MCP server.

Yep.

And MCP is like a model context protocol.

That's right.

As a travel adapter that you plug into your PMS system just for people who are listening out there.

Yep.

Uh, what do you think about this, and especially in the context of privacy governance security trust.

How are you approaching MCP servers?

How do you think you're gonna use it in the business?

Yeah.

What's gonna happen with this?

Is it hype or is it gonna be, is it gonna last five years?

So firstly, Apaleo, one of the reasons we chose them is their tech stack is fantastic and, and they're super forward about these types of things.

People are kind of like, what's an MCP?

Why is it like, I've all the APIs there, why do I need an MCP?

And your question already, the privacy security.

So I've already built some, i've built an AI native app, so I'm in love a bit at the moment with a product called Replit.

Yeah.

And, with that I was able to build basically a full AI native Apaleo app doing a whole lot of interesting things in their test environment, just everyone knows, and, already that was super functional, right?

And for example, it'll look at all the events locally for that hotel, forecast where those events are.

If there's a convergence, it'll automatically, ask you to change the rates, which you need to approve.

Off you go.

You already have that privacy and security.

Right now with, just without the MCP.

The MCP allows the AI to have a really much deeper understanding of, let's say, the total customer journey and how to apply it and how to orchestrate all of those APIs.

Because currently the APIs are only as good as their documentation ultimately.

Whereas the MCP is able to be far more holistic and able to, let's say, take the technology much further.

So for me, it's just a leverage point and you still need the biggest, or the first thing I would build with MCP right now to your answer, your question would be a privacy and security wrapper.

And the first thing it would do is be checking the stuff that wants to go to production.

And what do you think the use case would be for that?

Or it could be anything.

Is there something that fascinates you?

Yeah, I think for the guest journey, there is a number of personalizations at scale that we can apply for the guest journey.

And let's take the usual, one of the best drivers getting more direct bookings.

Okay, so everything that you can take away from the OTAs and get direct bookings, I believe that you can, you could put an MCP server up whose sole job, not the server, but MCP based AI app, whose sole job is to basically help take you to 60% direct.

That's a really good use case.

It's really valuable.

No one can do it currently.

And you would use that as your north star for the ai, so you don't necessarily, you need to change the way you think 'cause don't give it the I want A to go higher.

You want to give the AI these north stars now and say, you need to drive my direct bookings from 30% to 60%.

That's your metric, that's your success.

And let it find things that you never thought of and let it basically run.

And you think it's smart enough to do that?

I think, it will take 12 to 18 months and then I think it's going to start to have that capability, and it might be that you need three or four other agents.

To do that as well, not just one agent.

You might need to orchestrate three or four other agents.

You're already living in the future Mike.

Yeah.

I really have forced myself to go AI native and I'm trying hard to make that move.

Yeah.

Yeah.

And I think it's gonna be some kind of competitive advantage if you get there before everyone else, because you'll have a few years then in your, home markets to kind of..

You'll have, um, you'll do things for the guests that no one else can do.

And that's what you have to, so if everyone's got AI and everyone's got MCP, so that's all automatically a huge lever.

So then it's the ones who got in there a bit earlier, we'll get some advantages.

But to be honest, if you take all of that, the piece that's gonna be the biggest leverage point is the human.

So the ambassador and the guest is still where you're gonna win.

So ultimately don't get, don't get sucked into, just this tech only world.

The more that we give these guys the, the ammunition, and the ability to recognize you and to basically support you, the more we're gonna win.

Right?

So you have to bring it back to the ambassador at some point.

Thank you so much for your time, Mike.

It's been a pleasure.

Went quickly..