#1 - How Staycity Group Went All-In On AI Before Everyone Else

Paolo, thanks so much for inviting me over to the brand new Wilde location in the center of Vienna.

Really appreciate it.

Hi Jason.

Welcome.

This is the latest and, uh, newest of the Wilde brand at Staycity.

So we're very proud of this property and almost a 200 year old property.

Such a historic landmark.

This is quite special.

Yeah.

Even for our standards, this is special.

It was built in 1850.

Yeah.

Is 22,000 square meters.

Hundred thirty seven rooms. Uh, it's a very historic building , we have a fabulous courtyard, some, spectacular rooms. It's the crown jewel of our estate at the moment.

Amazing.

And you feel it when you walk in.

You're just like, wow, this is cool.

It's very difficult to find in the Vienna City Center, uh, places that can host, uh, families.

So we have big rooms. We can host people.

Four or five people.

It's quite spectacular actually.

So it's a bit of a playful, uh, luxury boutique sort of experience.

Yeah.

Um witty comfortable we're very, very proud of this property.

Yeah.

And I just wanna zoom out, and take a step back because we started on this AI journey two years ago when GPT 4 first released and.

It's easy now to look back and say, oh, that was obvious, but it wasn't obvious back then.

In fact, I do remember when we went from GPT 3.5, that's where the world saw chat GPT first, and everyone's like, wow, this is gonna change everything to GPT 4.

And I remember looking at the cost of those tokens and everything that we did cost two euros and we're just like, is this gonna work?

Is this gonna scale?

Can we even build a business on top of it?

But I think where I want to go to is that you saw this coming, it was two years ago, and you had the foresight to just say, I'm gonna go ahead with this.

So I want to get into that mindset to say, what were you thinking?

You're very generous in your assessment.

I think, foresight wasn't really there.

Um, I believe as a company, we want to take, a few bets.

So it's obvious that AI and technology are gonna drive operational efficiency, competitiveness and hospitality generally has always been behind.

There's some statistics of mid-size companies between , EU and the US.

90% of them are on the path of AI adoption.

However, only 35, 40% have a live product or use, uh, a AI in production.

And when I looked at the subset of companies that are in hospitality, actually between 6% to 11% of hospitality companies are behind.

And I think as part of my job as CIO, to try to bring a little bit of, uh, of that competitive advantage.

So we decided as a company to place some small bets.

Here and there and you know, this being one of them that is prooving to be fairly successful.

Amazing, And in terms of like, just to go back to put ourselves in the shoes of the 2023 us without perfect information, how did you go about decision making?

Because what I saw at some other properties is they formed committees and these committees issued a RFP and they went out to like.

20 people with the RFP.

And what happens is then the committee tends to argue and decide among themselves, but essentially what they're doing is they're hedging their personal risk.

Because if the project doesn't work out, they wanna be able to distance themselves and say, look, we made a decision together.

It wasn't my fault.

In other words, so how did you, and you stepped up, you took leadership here.

And how did that work out?

Um, Staycity is a, let's call it a medium sized business.

We employ about 1500 people across 40 properties, and we're heavily centralized.

We have about 150 people in headquarters.

However, at heart, we're a startup that has grown a bit too big.

And so we don't reason, as a corporate entity, we value agility.

We value, you know, taking some risks, some bets, uh, as I said, so we don't really care about covering our, you know, yeah, we just wanna go do things, see what works, failing fast.

My past is in software development before, you know, joining Staycity so I'm very, used to that kind of agile fail fast, uh, quick iterative process.

It isn't as fast as a technology company in hospitality, obviously.

But we're probably starting to be a bit faster than the, the old guard.

Um, I would say so, yeah.

You know, there's a level of committee, there's a level of buy-in because, we start talking about, uh, decent size investments, uh, in technology.

Um yeah.

But we try to start to shy away from that.

So the ability to have a management team that doesn't blame failure.

I think is, it's very important.

So it's a management team that wants to know truly what works and what doesn't work.

And that fits true in our culture in being very, uh, metric driven.

So as you know, in big companies, there's a lot of opinions.

I think we should do X or Y, Z, and we try always to put a number against initiatives.

So, numbers don't lie.

So we try to stay close to the truth, close to reality, you know, and honest keep ourselves honest at every decision point.

Amazing.

So it sounds like a cultural angle as well, where kind of the leadership is looking at these new initiatives and saying, okay, let's try it out here, let's do this, let's do that.

And I, I guess to a certain extent, AI is coming from the top because it's a bit risky to adopt it from the bottom.

So you don't want your employees plugging all your data into any new AI tool that releases.

So there has to be some guardrails.

I guess.

It's a cultural journey more than a technological journey to a certain extent.

So we're rolling out, paid ChatGPT licenses to all knowledge workers.

Right?

And this is a funny, this is a funny thing.

I was in, in talks with the, the openAI sales representative for, big number of licenses, and they said, we see X hundred of your employees already have OpenAI licenses not paid for.

They're obviously go and train their model and so on.

So you see that, you know, people, curious people, great employees will try to be modern and to be efficient on their own.

So you just need to create the culture and the infrastructure in the business so that you leverage that rather than trying stifle it.

Yeah.

And I just wanna take one step back here because you mentioned something very interesting a while ago.

You said, my background is in software engineering.

Yeah.

Can you tell us a little more about your background?

And I wanna know how does that background factor in this AI journey right now?

Yes.

So I started as a software developer.

Um, 25 years ago.

I had, a software as a service, company myself, which I've grown and sold in the , language school, kind of sort of business.

I then joined a venture backed company, called homestay.com, which is sort of Airbnb, sell your Room.

And that's how I got into the sort of hospitality, uh, space.

That company grew a bit and then didn't, you know, quite grow as much.

And that's when I joined Staycity, where my interest for travel and hospitality comes from.

In terms of where play, where that plays a role in the AI journey.

I think just as any techie, you know, you see problems, you see processes to automate.

And when you see the ability, of AI to, uh, understand patterns from large, quantity of information that we do have access to in Staycity, you just see the possibilities of all the processes and all the, the things that could be automated and make us faster in decision making and so on.

We use AI, uh, in places in Staycity think, um, forecasting of housekeeping costs.

Housekeeping is generally a high cost for hospitality businesses.

And being able to forecast that correctly and to model how much it's gonna cost, it's an important factor in keeping your company very profitable.

And you know what better than machine learning and AI and just throw all the data and see patterns that humans wouldn't be able to spot.

Yeah.

And do you care about the difference between machine learning and genAI?

Because to a certain extent, the way I look at it, like sitting on the outside from a software developer mindset, machine learning's been around for years at this point in time.

Hotels have been doing revenue management.

for ages and machine learning kind of depends on the amount of data that you feed in.

And I think what's changing right now is genAI allows that language element to come in and speak to your data, not only with just like numbers models, stuff like that.

So what I've noticed, and this is going to all of the hotel conferences- Every single thing is an AI company.

I, was at HITEC in the US a few months ago, and I kid you not, they had an AI rodent catcher with dashboards and stuff like that.

Yeah.

Right.

So you're like, I don't need AI to catch rodents, but so how, how do you filter that out?

Because obviously on hype now, and everything's called AI, between heuristics, machine learning, genAI, the different categories.

Do we care?

To a certain extent, I care as technology leader in the company, uh, because I wanna make sure that I set up the correct structure for a given problem.

If something, if a problem that we're trying to solve is very deterministic, we go towards heuristics and machine learning, obviously.

Uh, if it's less deterministic, uh, you know, genAI, I might be applicable if there's a lot of text.

And, um, so even the setup of having a human in the loop.

Or not having a human in the loop and what, what you are tackling, you just need to know what is the correct representation.

You need management and the different departments to understand which flavor of AI is best suited to a certain problem.

So for that, I think the education part, when everyone has got access to AI, is very important in the business.

You know, you get someone saying, oh, can we run payroll with GPT?

You know, until you have a level of hallucination.

Or you don't have a human in the loop infrastructure set in place.

You just need to be a little bit careful.

Know what you mean.

And in terms of like just.

When do you start building something yourself as well?

Because like you've got projects that you run in-house somewhere, you work with external people.

Do you ever think like, okay, this is a small little process, I'm just gonna use ChatGPT and get it done, versus saying, okay, this is something bigger, maybe we should work with someone to do it.

Where is that line when it comes to build versus buy?

Build versus buy is, uh, is always a big, a big question.

Uh, let's start from.

Let's call it first principles, right?

We're a hotel business, we're a Aparthotel business.

We're not a software development company.

So we would have a preference in buying services from someone that does that for a living, and that can spread the cost of building the tool or, you know, the, the system across a multitude of customers, unless we identify that specific problem to be specific to us or to be, uh, the ROI to be high enough to justify the cost of building.

That said, it's very hard to find suppliers that both give you the solution to the problem that you have and integrate with a multitude of systems that you have.

So one key attribute for a supplier for us is to be able to plug into our ticket management system, our property management system, into all the things because an isolated system regardless of how good it is, if it doesn't talk, if the plumbing is not there, the use of, you know, drops very quickly.

What you mean.

Yeah.

There are areas of the business where.

Potentially, uh, a build, um, would make sense.

It could be, for example, we, do a lot of feasibility studies, so we go through 5, 6, 700 sites that where we could be in the hotel per year and only a handful ends up being, you know signings and, uh, and build.

So that funnel of saying, okay, we have a site, this is, you know, draw, a floor plan, how many rooms can we put here, how many floors, and get the iterative process.

And that could be a tool that is a competitive advantage.

Yeah.

Or picking the right location.

That would've a cascade effect, obviously, on how profitable that property is for 20, 25 years going forward.

So these are processes where maybe you say.

They should be, you know, just our own, yeah, our own IP, let's say.

Yeah.

And I guess you've also invested a lot into your booking engine because that's quite a slick, custom built piece of software.

Yeah.

We started about three years ago, and we rebranded both Staycity which is our , urban, brand, and Wild which is our luxury lifestyle , boutique, brand.

We wanted two different websites and we wanted to push for direct sales.

Now, every hotelier wants to push for direct sales.

One key for that for us was to have a very highly converting website that we can fine tune.

Uh, so we decided to rebuild the booking engine, uh, our own.

We did a business case.

It seemed to make sense.

If we could achieve a, a 10, 15% improvement, in conversion on the website, that will pay off and would make the unit of economics, work.

Um, we've done it and we, we over achieved in terms of conversion.

So we're now pushing on the marketing side and performance marketing.

So the PPC, then above the line and all, the marketing team has grown in the past and is, proven a very fruitful, endeavor so far.

Yeah.

But I think you've even gone further than most people go because you also built a portal where people can kind of self service these, guest journey tasks, like online check-in and cancellations and stuff like that.

Is that something that you think is.

Core to the business will it feed nicely into the AI side of things because AI is not gonna be able to self check in, right?

Absolutely.

Uh.

If I take a step back, right?

As a company, we're customer focused.

Um, we want to provide the customers with all the tools of what they need.

And sometimes, you know how stressful it is.

You book a hotel, you want to cancel, you want to add a day, you wanna remove a day, and you want those interactions to be uber quick.

At the moment, we have a customer care team.

You can self serve, but it's always quite stressful.

I wanna, at the night, want to cancel, I don't find the right login.

Now, AI and genAI are showing us how the interface between humans and computers are becoming.

You know, more seamless, like in now you use natural language, either voice or, or chat and, uh, ChatGPT obviously is turning everyone to do any task via, via chat as a single, uh, interface.

So the, the vision will be to be able to hook that interface, that chatbot, D3x in our case with, with abilities with maybe MCP servers or tooling so that guests can self serve and can do whatever they need to do as quickly as reliably as they can.

And that will be a win-win situation because the customer fundamentally will be able to do whatever they need to do seamlessly.

And it'll be an operational efficiency piece for, for the company as well.

We have a, our customer care team, many languages, 10 different languages in German, Italian, and French.

But what if you are from the Middle East.

What if you are, you know, South American , if you're Japanese, we don't have Japanese speaking colleagues at the moment.

Yeah.

So 24/7 multilingual, instant, uh, responses.

That's, yeah.

That, that makes a lot of sense.

Yeah.

Father, just to switch gears, like what keeps you up at night right now?

Like what do you, in terms of like a technology point of view or anything else actually, what worries you and what are you thinking about?

Okay.

Well, there, there there's a few things.

I'll mention some that, fortunately don't keep me up at night, but probably should.

AI is gonna replace jobs, right?

That's kind of the elephant in the room, and some people just don't want to, don't wanna mention it, and we're in a very lucky position in Staycity.

We're a growing company.

We're gonna grow 3x in the next few years.

We're signing a lot of new deals.

We're opening a lot of new hotels, so it's very unlikely that we'll reduce our workforce.

But what we will do is we'll try to not grow headcount linearly with the growth of the company.

Okay.

So doing more and more with the same amount of people, um, that we have.

But that's something that might keep other companies, awake at night because if you're on customer service team, uh, you know, you might be worried of, automation, taking, your job, your job away.

Um, the two things that are worrying me are actually cybersecurity.

Yeah.

The scams we see, we spend quite a lot in cyber cybersecurity.

But I can see the scams are getting more and more sophisticated in the age of AI.

So the education and training programs and making sure that everyone is up to speed with the new techniques.

That's something that, that worries me.

And the other thing is actually not technological it's cultural.

Technology is improving so fast that I don't think the culture of the people working in the business is adapting quickly enough.

So how can I, in my position, um, create training programs and foster curiosity across all the departments so that they can break down their jobs into tasks and then think where can AI be applied?

You know, replace a given task?

So don't think about the whole job, like not being replaced completely, but which task can we delegate to AI?

Which decisions can we delegate to AI?

And I think that culture transition.

Is something that's very difficult for medium and big, uh, company.

Yeah.

And there are many jobs that humans actually don't want to do.

So for example, I was just chatting with another client of ours in Berlin.

They've got 12 properties and we are working on a feature where we can kind of, generate new invoices because German customers have very strict standards when it comes to invoices.

And they're like, my VAT number's not on the invoice.

My company address is wrong, my company name is wrong.

And in order to achieve that, we have to actually void the invoice, issue a credit note, create a new bill against the credit note, and then generate a new invoice.

And that takes someone a lot of time to do that.

Yeah.

From a human point of view and nobody.

Gains any satisfaction from doing that?

So much so that the customer was telling us like, if you get this done, my team is gonna send you flowers every Monday because they, they're gonna love you for that.

So I don't think anyone's worried about automation when it comes to invoice generation.

Yeah.

Uh, but something interesting that you did a few weeks ago, months ago is you created a new role in your organization for an AI product manager, program manager.

Can you talk a little bit about that?

That's the essence of that human in the loop structure.

People used to look after, let's say customer queries or you know, what time is your check in?

Do you have inhouse already?

Have this, like that instead of replying now those people need to be the managers of a system that replies automatically to the mundane queries.

So having someone that project manages the knowledge base, the review, the performance is key.

And in particular now for the, uh, for the D3x platform, but many other platforms. And what we're seeing is that, sub department is gonna grow out of my team specific for AI initiatives.

We're doing other little bets.

So for example, we have a large, data warehouse team where, as you said, we're heavily data driven.

We collect a lot of information for different departments in different city.

Systems, but it's very hard for a company of 1500 people to access that information.

Writing your reports and Power BI and Data Lake and ETLs takes a long time.

So how do you increase the data?

Lit literacy.

And we are now adopting a platform that is a chat to SQL transformation.

So you can ask any question in natural language and that automatically generates a query appropriate with a.

With a semantic layer into our data warehouse and star schema and spits out an answer.

So that's one of the steps, right?

Of lowering the barrier to access to data.

Yeah.

And I think it's another instance, and mark my words, Paul, is probably two, two years ahead of the curve here as well.

Like just thinking about an AI sub department is very interesting.

I think more organizations should be thinking that way and thinking about kind of bringing the team along on the journey as well.

Now in terms of your North Star, since you're like so far out in the open.

What, is there any hospitality brand that you look to for your north star?

What are you looking for in terms of when you build out this company, what, what is the end goal here?

I don't have a specific one.

However, uh, there are new wave of hospitality businesses.

There are even more tech forward than we are at the moment.

Uh, staff less operations like BobW, Numa, CitizenM have been for a long time ahead of the curve in terms of the adoption of technology.

So, um, we can't claim to be there.

We have less tech in the room, for example.

That's not something that we have particularly believed in, over the years.

But we want to be world class in efficiency, in profitability, and customer satisfaction.

Amazing.

And just as we wind up I want to ask you a simple question, like what new trend are you checking into and what trend do you think you're checking out of?

Because, Customer data platforms, I think for marketing that there's something we looked at.

We haven't really jumped on that bandwagon.

We believe that brand, will trump, the hyper personalization.

We, obviously we want to have a great view of the customer, great analytics, but we've seen over investment in customer data platform, they're very expensive tools on half a million, A million a year, not pay back as much.

Right.

Okay.

So that's something that would stay away from, um.

Mobile, uh , mobile door locks.

It's something that we wanna do, but not right now.

'cause adoption is still poor and experience is, is not as quick as seamless.

So we, we've done all those trials, so CDP, mobile keys, and I think just going back to having a chat interface to most features and most systems. So what my mention is.

That we, there's a chat bot for guest, but it could be internal chat bots.

So I give an example obviously to query that out.

Warehouses, I'll repeat that one.

But also, let's imagine you're in the finance team.

We have payroll for 1500 people.

There's definitely employees that have questions about their payroll and say, oh, what is this?

What is that?

And so imagine how many of those quiz going into the central finance team.

What if a chat bot could answer 50, 60, 70% of those questions without, you human intervention.

So, you know, you have internal customers for Gen AI chat bots as well.

Amazing.

Yeah.

Yeah.

So maybe we covered those.

Yeah.

We start a podcast on those three points that you mentioned right now.

Thank you.

No, but thank you so much for your time, Paolo.

It's been a pleasure chatting with you and, I look forward to continue working with you.

Yeah, looking forward to it.

Thanks.

Thank you Jason.

Thanks.