How CitizenM Mastered Hospitality Automation with Mike Rawson (Formerly CitizenM)

Speaker A: Tech is often, for me, the least of my problems. And what I mean by that, it's people, process and technology and often it's in the people and process that you have to spend some time. And unless all three of those are, uh, working, forget it. It's not going to work.

Speaker B: My name is Jessica Gillingham and welcome to the Pillow Talk Sessions podcast. On this podcast I talk to leaders about hotelization, tech enabled hospitality and the convergence of lodging asset classes. And today on the show I'm really pleased to say that I have Mike Rawson, who is the former CIO of CitizenM, amongst other really fantastic roles. And in this episode we go under the hood of what made CitizenM so successful. We talk about AI and its role in hotel hospitality and then also what Mike's view is on people, processes and technology and how they really allow for a fantastic hospitality experience, but also operation. So I really hope you enjoy this episode and this conversation as much as I did.

Speaker A: Thanks Jessica and happy to be on Pillow Talk and uh, appreciate uh, you guys hosting me. So I'm Mike Rawson. Well, I was the last seven years I was the CIO at Citizen M hotels and that was a fantastic ride as we essentially did a very large scale out focused in the US and when I arrived we're at 13 hotels and uh, when I exited we're at 40 hotels and um, so that's been an amazing ride. And, and essentially at that point I was in charge of the global technology so everything in terms of plan, build and run hotels. So the scope was very big. But at the, that gave us a really huge opportunity to focus on the guest. And Citizen M is a very tech forward brand and, and so I was in a luxury position of, of having the team build on that tech specifically to make sure that the guest had the best possible experience. And I guess ultimately for us that was, that still is in the nps, which was one of the best in the world and was one of the reasons that Marriott was excited to buy the Citizen M brand because the uh, combination of the amazing ambassadors and the great tech gives us a great nps. So previous to that I was working for Heineken. I worked at Heineken head office for a number of years doing local, regional and global tech. And I've been in tech for about 35 years, corporate enterprise tech for about 35 years. And originally from New Zealand, hence the slightly, ah, strange accent.

Speaker B: You're not old enough for 35 years, Mike.

Speaker A: Yep, I will uh, definitely be the oldest one on this call. I Promise you.

Speaker B: Citizen F is known as a pioneer in hospitality for a number of reasons. So it must have been really rewarding to be part of that journey and seeing the shift that and the sort of the um, the innovation that Citizen F did in a number of areas, whether it is like the membership kind of bit or the kiosks, you know, the, like the brand as well. So what was it about CitizenM, do you think that was most appealing to you both before you started, but then also through that journey of being CIO there?

Speaker A: Yeah, well, um, what I was looking for to come on board was essentially a very strong brand that wanted to scale out, that wanted to basically have a strong tech enabler. And so honestly the job description was almost written for me in many ways. So that was just a no brainer and I was just super happy to uh, jump into that and enabling myself crew and then facilitating the digital transformation and essentially creating a really focused tribe of people who were uh, really good at the customer journey or the guest journey at this point, really good at understanding the employee part of the journey that supports that. And the shared services model that was running underneath, it was just a great opportunity. And at CitizenM, probably the biggest problem was not the lack of innovation. We just constantly had things that we would love to do. It was prioritizing and some sometimes doing the boring stuff and all those types of things. But from a high level the app was amazing because it did everything. So our app was the only one that would basically do pre book, book, stay, control the room while you were staying, check in after you've landed at the airport, check out while you're in the taxi. That was fantastic for me. And then other things that were also good, which probably obviously the guest didn't see. We did Mamba, which was the mobile ambassador app for the employee in the hotel that also saved a lot of, and did a lot of automation. Probably I'll give you the most unique thing we did that the guest never really saw, but I was just really proud of and that was whenever you check out of a Citizen M M room, essentially we run a room check. We check the room and we check the TV and we check the iPad and we check the lights. So we basically check that the room's still responsive. Um, if it's not, we automatically take the room out of service, we raise a ticket with what's failed. 25% of our guests pull the HDMI cable out of the tv. So that basically raises a ticket and all of that happens before the next guest gets that next experience so that we never give them a, uh, broken experience. And I loved that because it was a combination of the tech, of the people, and ultimately it gave a much better experience to the guests, which we did see in the nps. It just took us, a lot of everyone had to work on that together.

Speaker B: That feels like the whole DNA of Citizen N. The tech, the people, the experience.

Speaker A: Yeah. And that's a very good way of thinking about it. So that's. And it's hard to do so certainly wasn't me and my team, we worked closely also with, uh, Casper, for example, who, who was the uh, chief experience officer would be. He brought those products on, he understood the product managers and putting them into the organization. And he was so close to the customer journey and all the, what we call moments of truth that we knew that really well and we knew where to focus. But the tech is often, for me, the least of my problems. And what I mean by that, it's people, process and technology. And often it's in the people and process that you have to spend some time. And unless all three of those are working, forget it, it's not going to work. So that would be my. If there's one takeaway for the people already today, it's people process technology. You have to combine them, um, to be successful.

Speaker B: Which feels very, very refreshing today when we're all hearing everywhere, you know, AI agents can take away all the people, et cetera, et cetera. It feels very refreshing to know that people's first there. Then you've got process and you have tech.

Speaker A: Uh, yep. So we literally draw it that way. So the tech is on the bottom and the tech also comes. Once you understand the requirement, once you understand value is once you understand a lot of other elements, then we can craft the tech and, and then, you know, we'll, we'll roll it out together and, and then you'll measure, uh, the results and you also do a retro, learn from what you did and then, you know, iterate. So that doesn't go away though. I think the people part of it, it's obviously going to change, but it's absolutely, in my book, not going to go away. I've been in tech for a very long time, probably 30 or 40 years of corporate tech. So. And I've never seen, this is probably the biggest generational technology introduction, um, certainly in my lifetime. And I guess just to give you guys an idea of just how fast it's going, the current rollout of AI, so let's call it three years, is Already well ahead of the PC and the Internet combined. So we are uh, seeing, uh, it's not just us thinking, it's a generational shift. We're seeing an uptake of a technology at a rate that's never been done before. So that's kind of, it's fantastic, but it's also scary and daunting. And because the technology is moving so fast, it does make some of the decision making a little tricky.

Speaker B: And do you think that trickiness is it, does it mean, is it almost like a paralysis of decision or is it making poor decisions? Like this sort of speed of what we're seeing of, you know, tools coming out onto the marketplace, new opportunities to do things differently. What kind of decisions is it making?

Speaker A: Yeah, so this is, you know, on kind of your usual shall I buy it, Shall I make it type decisions. At the moment you're also seeing whole platforms jump in front of each other. So this week it's Claude and next week it's Gemini and the week after that it's OpenAI and then Microsoft jumps in. And that rollout of those platform features is also rolling right across a whole lot of AI startups. So if you're an AI legal startup, I think you've got a lot of problems because essentially they are scooped up into the platform. So you've got what are called point to point AI doing specific things that are getting rolled up into the platforms. You're getting platforms with a lot of features that are just suddenly appearing. And uh, you know, video AI is an amazing example. Sora was 18 months ago was amazing and everyone was just like, wow. And that's already been canned and wrapped up and onto the next thing. Well, that's unprecedented probably in terms of the technology movements that we're seeing. So yeah, that makes it difficult to attach your wagon to something and essentially start making some decisions to roll out some tech to get some of the benefits that everyone's seeing with AI reliably or without the risk, or basically without making with a provider who's not here in six months time, which is going to happen to some of these providers, I believe.

Speaker B: So Mike, where is the um, two bits where you're seeing or where we're seeing technology and AI in particular impacting hotels and hospitality?

Speaker A: Yeah. So the way we thought about it, or think about it when I was at Citizen M is to try and help you land on the right technology. We think about personalization at scale for the customer or the guest, and we think about automation at scale for the employee. So everything that you do should at least be hitting the guest or the employee. And that's kind of like your starting north stars. Now in terms of the decisions of which area to go after, there's a few things to think about. So do you really want to put your AI directly to your guest at the moment? Does it have your right brand and does it have the right values? How do you know that it's doing the right work? Is it meeting all the compliance? So one of the things that we did is we put our AI initially against our employees and went uh, after automation at scale. And we learned a lot from applying the AI to the employee scenarios before we trusted it, if you like, with uh, the guest scenarios.

Speaker B: So where do you think that in today we're getting it wrong? You know, where are we getting AI and tech enabled hospitality wrong, do you think?

Speaker A: I think the AI value proposition is just fantastic. So I don't think any. I think everyone's very excited. The excitement levels are probably 20 out of 10. So I don't think anybody's not seeing the opportunities. I believe that the problems are arriving in the AI governance and AI control, in all of the things that your security and privacy officer, what keeps them up at night, and also all the legislation that's arriving, like the EU AI act. And it's essentially also finally trust. And that trust is running in two ways, people worrying about losing their jobs and secondly trusting it against the guest. Right, so those are uh, areas that now AI is forcing because I would say in hotels at the moment, I'm going to go somewhere between 30 and 50% of hotels already have an agent or multiple agents running. But I suspect they're already walking into these types of issues and don't have good answers for them. And so that's for sure slowing down the progress.

Speaker B: How will it change, do you think? How, how do you see it changing?

Speaker A: Yeah, so, and there's obviously a lot of discussion about this and it's at the moment it's just, it's moving so fast that it's also the landing zone I think is not very clear. But if I was to have the CFO on this call today, just right next to you, um, and I said to him, right, we do the AI magic and you have no staff. I'm sure the CFO would go, yeah, wow, where can I click on that? That sounds amazing. But I'm pretty sure there'll be other chiefs in the call who will not. I want that. And so Klarna is a good example. So Klarna was one of the leading ones who went kind of all in on AI, ditched a ton of people and then hired back. As far as I can see, roughly 30% came back in. And I think that's an admission that an AI only model with no people I think will fail unless it's just really so simply transactional that it makes sense. The biggest thing that I think will come and I think that's something for your listeners to think about. Yeah. Dear business, where's your value add? So if we do automate some of these areas, where do you want to play? Where's your point of difference to the competition? Are you. If you're four stars, what does that mean when they arrive and how do you basically leverage that? What we call moment of truth. Right. And I think that's going to lead to decisions that will have to be made in the areas that you'll decide to play in. And that's a good thing. So from that point of view, making the business more defined about where they want to play and what their brand stands for. And ultimately that human to human interaction is going to become more valuable. So you have to value that up and then figure out where you want it to happen.

Speaker B: That's also about putting the customer at the center, isn't it? If you're thinking about what your value add, which lane you want to play in, what your brand is, you need to know who you're, who you're serving, don't you? And I feel that's something that you did really well at Citizenm. You, you had a customer in mind. That was your kind of. It was a Citizen M type person.

Speaker A: Yeah. And that also entailed choices. Right. So for example, because we had one room type, we couldn't support a family with young children. That just wasn't, that's just not in our model. So that, that's a choice. We decided that that was someone else could do that because we had a specific room in mind and we had a specific type of people in mind. So you know, there's one example of it and knowing your customers deeply. And I know that Casper would. He easily knew our top 10 customers by name and was regularly on WhatsApp with them and had just a uh, like say a micro feedback running the whole time. Yeah. That's amazing. We also knew for example, that we had a lot of business users. And the secret for the business user is repetition, repetition is what saves you. When you're a business user, you want to do the same thing every time M. So for example, if you check out at Glasgow and we already know that you stay at Glasgow every week, so we ask if you want room three to five next week that's available. You don't need a whole ton of brand new AI tech right now. You know, we could do that um, back then, but understanding that, that type of customer and understanding um, how to get to those moments with them, um, and recognize them. So that recognition of that type of customer. Yeah, that's the trick and essentially that's you start focusing on that and then you point your tech and your people and your processes at that and you get success.

Speaker B: I think you mentioned about the personalization that you were able to do at CitizenM. Um, we talk a lot now around personalization at scale and how AI is enabling that or will enable, enable that. Where do you see that changing and shifting and where are you already seeing uh, in hospitality operators being able to, or brands being able to do that? Personalization at scale.

Speaker A: Yeah, so I would say we had personalization at scale as a North Star and I would say AI will give you hyper personalization at scale. So I believe that it will go to a level of recognition that we couldn't manage at the time and that, that literally will. An example of that will be if you have your coffee with oat milk. That kind of hyper personalization is going to arrive with you at the hotel. Hyper personalization will also be your room temperature where you have the shade set, you know those types of things about your room. Uh, we could do some of that. We understood things like the average temperature for your stay and when you check in it makes sense that we'll, that we'll try and basically find you the room that's closest to your temperature preference is actually good for the planet. And again it's just recognizing you. Right. So the opportunities with hyper personalization are uh, I think substantial. I'll go so far to say it's not that far away where I think you could just walk in with your credit card and just tap your credit card and that essentially is going to be your check in your key and essentially with that will come all the elements of your hyper personalization. So the tech is really, I think probably getting is kind of ready for that. It's just there's a lot of other things sitting in the way of it.

Speaker B: I don't know because I know you're based in the Netherlands and I don't know if you have Amazon stores in the Netherlands where you can literally go in, put the goods in your basket and sort of almost. And um, I think it's like you can even just walk out of the store and your credit card is charged and that feels like the. Almost the ultimate in being able to have very frictionless retail experience. Do you think that we're coming close to that or we can be doing that in hospitality. Have that really frictionless where you can literally just walk in and go straight to your room and all sorts of other things that all those needs that are anticipated and met M. In a way that. Like an example of an Amazon store.

Speaker A: Yeah. So Amazon would be an example of what we would Never do at CitizenM because you've taken away the human interaction. Right. So we, we are uh. But it's a very good example of dear business. Where would you like to play? So that might be if it's purely transactional when you just want to make the margin and get the customer out of the store and your brand is so strong that the customer's gonna come back because they've kind of got no choice. I suspect you're gonna see that kind of automation. Automation, yeah. Yes. For us, you know that we knew the checkout, we wanted to make it super easy. But we also recognize the checkout as a moment of truth and people who are social. Yeah. Value saying yeah, see you John, thanks for the stay, see you next week. Especially the regulars because that's what they know. So again it's a very good example of that contrast of let's say the transactional value, um, and that moment of truth versus the brand and the moment of truth when they're leaving. We're also in their mind they, they're like goodbye and I had a great day and see you later. Then we know that that's valuable and we know that that's also good for the review and the nps. So um, and I would say to you the technology can do all of that. So again it's dear business. What do you want? So I'm, I'm in a luxury position these days of being able to say yes, what would you like? But be careful what you wish for because essentially you need to really understand your brand house, your offer, uh, your moments of truth and where you want to play. And I think this is going to start to really uh, to really land. And by the way the hyper personalization we'll also figure out probably if you do want to low key check out your business person and you just want to leave and go. Yeah, we'll probably figure that out. But that would be my thoughts on the matter. And it's such a simple example, but you see how this leads to what should be these meaningful conversations between the business and tech and essentially landing these customers the right way.

Speaker B: What are the risks, Mike, around personalization at scale?

Speaker A: There are a lot, you know, an example would be the guy who travels, uh, regularly with his wife and all of the TV and all that stuff is basically set up that way. And then at some point he's not traveling with his wife. Uh, that can be an example where the personalization at scale can be a problem because the stuff that's probably lots of things could be different. That basically could lead to some pretty awkward conversations. So one simple example, I do think that people's privacy tolerances, and it's also related to their age and it's related to their nationality. Um, it's quite a big mix that essentially makes people decide where their level of comfort is. And we also spend a lot of time talking and thinking about that. So the way that we address it is we think about it as recognition. So we're just trying to add value, remove friction and, and focus on that recognition. And we hope that that doesn't go too far wrong. But if it does, you also need the guest to have a way to reset that. So if we've got the hyper personalization wrong, the worst thing you can do is keep repeating that the next visit and the, and the subsequent. Right. So you have to have a way to. It might take you five visits to figure out something, but if you've got it wrong, you need to be able to fix it it um, instantly. Now it might be the ambassador needs to fix it so that human in the loop will get in there and go, yeah, we thought this guy liked having a room on a high floor, but actually he hates it. Well, okay, let's. We need to be able to fix that. So it's got a lot of cool stuff, but it needs to also have a human in the loop at the right point if it uh, if it goes bad.

Speaker B: And then in terms of where technology is changing, hospitality, where else are you seeing big change or. And big opportunity at the moment?

Speaker A: Yeah. So stepping away, if you like, a little bit from the operator model, there's also a lot happening in the, in the real estate or the owner model. So at CitizenM, we also plan and built a number of hotels. And um, that's fair to say, I think five years ago there was a lot of promise around the digital twin, but actually it was, in reality it was fragmented and it was lacking and I now see that there are just some really powerful tools and platforms um, around the real estate construction, not just construction, actually the planning. So real estate planning, um, and building are now really converged and I think have a lot of advantages that we didn't have a number of years ago. And as an example, you can now run just really powerful simulations about your building that you're planning and how it, how many rooms that it's got, the number of keys, what you have to do to meet the local, the localization requirements. So for example, building in San Francisco came with just a ton, like 500 pages of compliance that had to be met. So now this thing can basically really simulate and run and give you a lot of options, but also just give you a lot of accuracy. So previously there would be a just ginormous spreadsheet running around with a lot of stuff on it. And of course you're only as good as the formula that's in it. So I, I think that uh, we're seeing a lot in the construction and actually during construction as well sort of move into that during construction. There's also a whole lot happening there. And you can literally have guys who are able to pick up an iPad, point it at the room that's half completed and it will show the plan of essentially what that room needs to look like, where the door will be, where the air con unit would be, for example. So those types of um, things are pretty cool. We, we did use one when we were building which, which the team introduced, which I really liked. And that essentially had someone who would walk around the whole building site every two or three weeks with a 360 camera on. And that gave us an absolute walkthrough tour uh, of every element of the building. And during COVID for example, yeah, we could still basically see what was going on in the building. And not only that, you know, five years later, those recordings are actually now amazing for insurance claims or any. So we can go back and go, well actually yeah, this is a, this is a, it's still within the 10 year warranty and it wasn't built right. And here's the uh, walkthrough that we did associated to that. So I think there's a lot of really good stuff happening in the, in the plan and build elements of the real estate industry. And I think that's um, actually happening.

Speaker B: So this actually brings me to a question that I'm curious about. If you had a piece of paper and you could write out or draw out what your ideal hotel would be Knowing everything that you know now and with everything that we now have available to us, what would that be? What would it look like? What would. What would be the sort of the essence of it?

Speaker A: Yeah. So I would copy with pride, uh, a lot of the Citizenm elements because we just know, um, why they're good. But that relaxed lobby and arrival space, I think, and done in that style, I think that's. It works really well. It makes the guests feel comfortable. They like spending time in there. And also the Citizen M room, it's small. We know that that's okay. You get a big bed, make sure you have a good window with lots of light, and we make sure you have a great shower. And the offset to that is people see the money spent in the lobby. So I'm like, okay, I've got a small room that gives me a huge number of keys, 30% more than any other operator. So I know that from that point of view it's going to be profitable. I do love the idea of the bar on the roof. So in New York, for example, it's just, it's amazing having a bar on the roof and even the pool on the roof, which is kind of like what we did in Miami because like, you're no one if you don't have a pool on the roof. So my. I would have a bar and a pool on the roof and I would have the rooms set out the same way and I would have the lobby with that, with that buzz feel and that comfort and great interior design that from Claudia, she did a great job with that. And essentially that is also really strong. That brand is really strong. If you walk into any Citizen M in the world, if you walk into another one somewhere else on the planet, you'll know you're in Citizen M. And I think that's, uh, you know, ultra good brand. And of course the tech, well, the tech in the rooms now actually has never been. It's never been cheaper. Probably were easier to put tech in the guest rooms when we did it. It was not cheap and it required quite specialist building skills. Um, and many building companies did not like it because it was just so different to how they, how they built. And I think these days, um, essentially the wireless element and those areas have just really come on to the point where I think they're commercially strong enough to do, um, a guest room with, uh, I don't know, I'm going to go for 30% less cost. That would be what I'd say.

Speaker B: I've stayed actually in the New York Citizen Air and the Tower Hill Tower Bridge. And I love them both. And actually it's not just that the lobbies are the same, they smell the same. You know, you can recognize it by the smell. But both, both. You know, I am actually a big fan of Citizen Air. Um, what I. What I'd love to know that in this sort of, this new scenario of hotel that I'm asking you to like build from scratch from up to down, what would your workforce be? So we're hearing, you know, loads about agents, AI agents and agentic workforce. What would be your split there and how would you manage your work, your workforce, whether it was an agentic or a, you know, human one as well.

Speaker A: Yeah. So for sure, I'm not a big fan of all automated. I'm sure in Japan you're going to find that where you can just do probably a motel or something and there's just not going to be any people. So I'm not a big fan of that. I do think that having a multi skilled ambassador. So our ambassadors do everything from essentially food and beverage, a cocktail, the check in, the check out there, and I love that and it gives them variety. So front of house is the ambassador and I would definitely have that. And there are, uh, I would have them enabled with a, uh, mobile ambassador app. And that app is all they would need I think. So shouldn't have to go to the actual big screen and PMS and all that stuff. Um, unless it's an exception. And the kiosk. Yeah, I mean you'll be able to do a check in remotely, no problem. But I do see the value of that arrival experience and having that kiosk and could be as simple as your phone's dead. There's lots of reasons to do it. So you have your kiosk. I think I would just have less of them because the overflow is you can do a QR code and do a web check in. So I think that gets rid of a lot of those areas. But that arrival, that's a big moment of truth there. I think the ability for the ambassadors to then also recognize you at the bar with the food and beverage and it's the second time you've been in a week and they know it's a gin and tonic for you, I think we get that gives us that recognition opportunity. And then, um, I also, I would not underestimate the housekeeping and essentially all of that scheduling and all of that, that will come from the phones as well. So that would be an element of it. Um, and that's a touch point with the guests. I think it gets overlooked. But the housekeeping, uh, can. Whether it's just towels or something like that, there's a moment there that they have a small moment of truth there that I think is important. And also, when your room's done and it's nice and smells good and you've got fresh sheets, um, it's part of the brand, it's part of the experience. So I would do all those elements rooms. I would have meeting rooms. I think that's a nice way to keep the buzz going in the lobby. And, and essentially, um, keeps. Keeps all the coffees turning over and all the F and B elements of it. And then, uh, on the top, yeah, you have. For sure, I've got the bar. And yeah, you know, look, it's not always easy to keep the bar full. And so you have to essentially then also have a bit of a combination of private functions and, and those types of things. But as a. As a just a wow moment and as you know, Citizen M M Bowery is probably the best, best, best view in the world of New York in terms of that bar at the top. So I would, I would have it because I kind of feel that they're iconic. So I think it's just a brand statement. So there you go. I think those, you know, and the AI is there again. I'm going to use it to get that MPS maxed out and essentially make that so that the ambassador and the guests still have the maximum amount of time that they need or choose to have, you know, to basically to have that moment. Right, right.

Speaker B: So you talk a lot about that moments, don't you? And moment of truth. So it's that connection that your team or, or elements of the property have to impact the guest's experience is really, really important.

Speaker A: Yeah, so. So those are really important. And it's important that you can recognize the guest and that you can add a bit of value to that moment and help them remember it. Uh, and sometimes it's helping in a different way. Sometimes the guest is upset and the ambassador, uh, is just there to help. And, and, you know, we. There are some great stories with the ambassadors about how they've helped in some of those areas. And we call them the random act of kindness, which is also another thing that we do just to do fun things for the guests, and it's fun for us. But there's also lots of stories where the ambassadors have just done things to help people when they've had a really bad day and something's just gone really wrong. And it can be as simple as, uh, you know, the guest posting. So even you know how good it's been. Like the guests, I can tell you one' story the guest posted. He came in and he, he said himself, I was grumpy, I was pissed off and I was really short with the person. And my coffee came and next to it had a big smiley face and it's like essentially it said something like, you know, hope those rain clouds go away. And he said, and that just changed my day. And he said, and it changed my day because someone cared and he fit. And he said, and I just had that coffee moment and I realized I'd been a bit short and, and they hadn't basically also clearly seen it, but they had then just put that moment together with that coffee. And a, uh, coffee is a simple thing, but the bit that went with it was the magic and the fact that the guests posted that, yeah, you know, we've done a good job. Uh, and that's what you need to do. You need to drive for those moments so that the ambassador's not busy trying to do some food and beverage, ordering of some milk and miss that opportunity to essentially recognize that guest and turn that.

Speaker B: So it sounds like there's no chance that you'll be replacing humans with agents or doing any of those things.

Speaker A: Yeah, no chance. So I would say that maybe in the economy, brands and down, down the food chain a bit, um, it probably is going to really happen and it's going to happen a lot. But I believe that four and especially five star plus, that's, that's going to be how you maximize AI is you've got to maximize that value of that human to human interaction and you've got to create as much of that in a meaningful way as you can. Not spam the guest, but basically find those moments and really by the time you have that moment to do the recognition, you need to know the personalization at scale. So you need to understand who you're talking to and a bit of the history to be able to say, I see you're in San Francisco last week. How was the trip, you know, leading into those discussions like that? We, um, know that's good for the nps. We know that that gets you ahead of everyone else in the nps. And that's what you're going to have to figure out. Right. Ultimately that's where things are going to

Speaker B: be to a topic that we talked about in our prep call, which is around the Dangers or the risks around compliance, security, things that we may not be really thinking about in the way that we probably should be in this sort of AI is everywhere world that we're in right now. What are your thoughts?

Speaker A: Yeah, so I think let's call it all the boring stuff. So from that, from that point, you

Speaker B: said that, not me.

Speaker A: Um, yeah. And the business doesn't. Look, it's just, no doubt about it, the business is not excited about any of this stuff. But I'm sorry, you know, ultimately the biggest problem is that the security and privacy is essentially mandatory. So at a certain point you just can't dodge that bullet anymore. You have to deal with it. So I think the EU AI act is coming in August. I haven't seen hardly anybody talking about it. So what does that mean? Well it means as a minimum, you've got to have a register of agents, you've got to have some compliance from. If you've got third party agents, you've got to have some compliance statements from the suppliers. You can't do certain things. So there's a, ah, there's a number of things that just can't be done and there's a number of that around the biometrics that you have to be careful of, um, in terms of the guest and facial recognition. So yeah, how's that going and essentially how's that going with uh, that new AI agent that you want to put in next month? Um, and those are going to be showstoppers. So that's one example. The other example is essentially just the AI governance. So do you really. I think the most concerning thing, non compliance and non security is yeah, I see these agents in these hotels and um, yeah, what's the logic? So what logic is it running on? What business logic is baked into that thing and who's signing it off? And then what happens when you want to change it so that next Tuesday, you know, the selling rate is 10% cheaper. Uh, currently that whole thing is just like, yeah, where's tech? And that's a terrible answer, where's tech? Because it's going to be, yeah, where's tech? I need my discount on a Tuesday to be changed to 10%. And, and then we change it and they go, oh, sorry, I screwed it up, I meant Monday. And then where's tech? So this current run of AI is doing nothing but baking tech into the landscape. And I believe that there's a serious need to get the business logic back to the business to get control of the bots so that you understand what they're running and essentially ensure that they're compliant and that you don't get any surprises. So there is currently the biggest blocker I think for the rolling out of let's say AI at scale. Scale agents at scale is going to be a lot of this compliance, control

Speaker B: and governance, which really certainly my LinkedIn feed has nothing about that. It's all, you know, it's all open. Isn't it amazing? We're going to change the world and you know, all of these things, none of, none of that stuff. But it is. And it will be a headache that comes along. It'll be a rude awakening, shall we say that that comes along.

Speaker A: Yeah, it's, it's guaranteed because some of the stuff is legal and hygiene and compliance. Right. So it's like, why do you think it's not coming? And so, and part of this is also going to be, there's going to be a lot of finger pointing. It's going to be like, oh, it's July and this stuff all comes in in August. Yeah, who's doing the EU AI Act? And it's going to be everyone looking at each other and it's going to be our tech. Well, the security officer has to do it or it's the uh, you know, privacy officer. So that'll be a bit of a mini train wreck. And of course it just doesn't have to be that way, but I understand because there's so much changing that it's hard enough to go, okay, what agent do I want? What does it do? How do I get it working? How do I test it? Okay, wow. I've got through all of that, what people are associated with it, um, what processes need to change, how do I create buy in for it. Okay, now I'm almost ready to roll it out and I'm exhausted and then you're going to tell me, oh my God, compliance and legal and security and probably that person's going to run out of due. They're not going to be a champion at that point. They'll just be ready for some sort of medication or alcohol, I suspect.

Speaker B: Yeah, I believe you. You're building something, have built something that can help the industry with.

Speaker A: So I have, it's, it's something that's on my radar. So I've been thinking about it and um, indeed got a framework running and kind of just got it up and running and, and it's called Value Driven AI and why it's good, why I like it, is that the biggest thing to take away today for Everyone with your agents is that they will have guaranteed a thing called a markdown file. And what's nice about the markdown file is it's in plain English. And that is the only genius thing about my framework. The rest is actually pretty much put together from existing stuff. But the markdown file is the one place everyone can come in and out of AI land. Essentially all I've done is build a framework around that. And in vda my markdown is substantial. So what does that mean? So it has the business logics. If it's the check in agent, it says that you must check in, that you must have a valid identity, that it must have a reservation, that you cannot check someone without X and Y. So it'll have a whole lot of musts and must nots. And that's all really expressed as business logic. And I can get the hotel operations manager to look at that logic and sign it off and test it and have an owner, crazy things like that. The other advantage of the framework is that because I've built the customer journey in there, so it knows automatically that we're talking about hospitality. So this is pre book, book stay, post day. That's the customer journey at a high level. The agents come because of that, because it understands what stage of the value value stream it's at. So the check in agent knows that it's basically part of the stay value stream and it is check in all the way to checkout. So because of that you will get automatic agents created for the right areas. They have the correct must and must not. And then finally, um, the opportunity is there to build the compliance and with the agent. So actually in the markdown file you'll also see a whole lot of compliance things that make the agent compliant. So you must log this decision, you must do X and Y. You cannot take crypto for payment. So there's a whole lot of compliance things. And whether that's SOC2 GDPR, the EU, AI act, that's all fine, that can all just go into the markdown files. So yeah, all it's basically done is create a framework of some really clear English that everyone can look and sign off and then that's what you put in production. And then you have control of basically what the agents are doing. And then lastly there's a human in the loop. So you decide what an exception is. So essentially if, uh, the discount is more than 25%, for example, it goes to the general manager of the hotel or goes to the regional manager. So those types of things can also be set up in plain English and all the escalations can be built. And if you're happy with that, you start with that new rule and you check the first week of that new rule and you're happy with it. Then you just change the exception to be normal and then the agent will do that. So it gives everyone a really clear way to onboard those agents in a controlled way with a human in the loop until you're comfortable with it and uh, then you can let it go. Then the last piece is because you don't trust AI ever is there's a witness agent and the witness agent is there essentially always looking for the compliance and always looking for the correct uh, activity from the agents and always able to get a human in the loop. If something happens, something strange happens or if, I don't know, if an agent fails or an agent goes down for example, for some reason then it will automatically escalate to in this case probably tech and go find it. So all I've done is converged, yeah the customer journey for hospitality with all of the ownership levels. So it's got a built in raki to start with, who's responsible, who's accountable, who's consulting, sorted, who's informed. It understands because of the customer journey, what agents to build, it understands what compliance to build and then it also understands currently how to do a literally a crawl, walk, run, rollout. So that's what I've called it. So then uh, you can just do baby steps until you're comfortable. So I'm hoping that's a, it's a pretty comprehensive framework but ultimately it's also very easy to explain and I'm hoping that will allow businesses to get the um, an AI operating system is how I think of it.

Speaker B: Mike, you've done all the hard thinking for the rest of us which is pretty amazing. So not just the hard thinking but put it all in a place for everyone.

Speaker A: Yeah, it's much easier. It's you know, waving your hands around isn't very useful but there is a working site so we two things. So essentially um, we've worked very closely with Apple AO PMS because that was who we selected after a two year process. And I do like Apple AO a lot, it's great tech. But just to prove my framework was not just theoretical I um, yeah built it on top of the Apple IO since their test system. But I have to say that if you go to that link you can actually run the agents. It's running on real data, it's running on top of Apple ao. It'll. You'll see the reservations, you'll see the folio charge. So it's not theoretical. It's running, uh, at now. Now. It's still a framework. It's still draft. I haven't finished it yet. We're still working on how to roll it out, but I do think it's already. The feedback is reasonably strong that it's fit, uh, for purpose. Yeah, it's exciting. I do hope, uh, you know, that it's going to be a lot of answers for the hotels who, especially if they don't have a big tech crew. You know, I was very lucky to have a big tech crew. I know a lot of hotels don't. And, um, I'm hoping this gives them a really, um, comprehensive AI platform to go forward with without, um, having to have, you know, 25 IT geeks running around.

Speaker B: That's really great, Mike. That's such a superb resource for the industry. So thank you, Mike. We are interviewing. It's been really, really a pleasure to speak to you. And I think there's so much more I wanted to ask you, but we have run out of time now. So thank you so much for being on Pillow Talk Session.

Speaker A: Yeah, loved it, Jessica. Great set of questions and, um, yeah, I'm super happy to have contributed. And thank you so much for, uh, having me on the show.

Speaker B: Thank you.

Speaker A: Cut.