Latest / GAIN Momentum / The AI-Adjacent Executive and Using AI for Senior Hotel Roles | with Vincent Somsen
Transcript
- GAIN Momentum episode #51- The AI-Adjacent Executive and Using AI for Senior Hotel Roles | with Vincent Somsen === Adam Mogelonsky: Welcome to the GAIN Momentum podcast, focusing on timeless lessons from senior leaders in hospitality, travel, technology, and food service. Today's episode is titled The AI Adjacent Executive. And I'm here with my special guests who are fellow GAIN advisors, Vincent Thompson and Michael Goldrich. How are you both doing? Michael Goldrich: Doing Vincent Somsen: good, Adam. Thank you for the invite. Adam Mogelonsky: Awesome. So special episode today, AI Adjacent. And in a nutshell, that basically means using the latest tools to free up your time and automate. So that way you can do ever more creative or important things with your time. I'd like to start off by digging down into what this term means. And I'm going to ask Michael Goldrich because [00:01:00] he here has published a book on the topic called Too Many Hats, Too Little Time. So Michael, what does the term AI adjacent executive mean to you? Michael Goldrich: Well, to me, and thank you, Adam, for having me on the program today, and Vincent, uh, to me, AI adjacent just means leveraging the technology as a tool to be a more effective decision maker. I mean, that's, that's all this technology is. It's all about facilitating Uh, Empowering Individuals, Executives. So, um, to me, that's what it means to JSON. It's just a, an assistant to guide you. It's another opinion. as, as we shift to more data driven decisions and move away from intuitive based decisions, something like this is very helpful. Adam Mogelonsky: cool. Uh, Vincent, over to you. Your focus has largely been on [00:02:00] using AI and other tools to help specifically in hotel operations, whether for executives, managers, or frontline teams. Can you describe what the term AI Adjacent executive means to you in a hotel operations capacity. Vincent Somsen: Yeah, for, for me coming from a hotel background, um, having studied, you know, hotel management, having worked for Hilton and Marriott over the years, I always had these thoughts that I knew were going to take a lot of time, are not going to have a large impact on the business itself. And now with AI, I can speed that process up. I can't take it away entirely. But as a hotel executive, whether you work on the front line, as a front desk agent, or on a corporate level, I can MIGHT more use of my time. Adam Mogelonsky: Okay, so we're talking about artificial [00:03:00] intelligence, and I think it's important just to lay sort of a ground, the groundwork for what we mean here, because artificial intelligence, there are several different types that have specific applications. The two that I want to bring into focus today are machine learning, whether there are various subsets within that. And then this newer one that gets all the buzz, which is generative AI. So can either of you, let's go back to Michael here, give a very simple explanation of what both those terms are? Well, I Michael Goldrich: Sure. So machine learning is basically taking data. and making a essentially a prediction as to what's going to happen from it. So it's taking what happened in the past, looking at that, and then learning from that, and then helping to make decisions in the board. So to me, that's what sort of like machine learning is. And, uh, it's sort of like, um, Netflix. [00:04:00] In terms of you watch it, you know, you watch certain shows, it sees that you like these different types of shows, so it can make a recommendation for you. That will resonate with you. So that's machine learning. It monitors your habits. Now, generative AI is a little bit different, but what it is all about, it's about generating new content, new text, new video, new music, whatever, based on what it's been trained on. So it, and it's all about, predicting the next word, predicting the next image, it's predicting the next sound. It's all done on probability. It's all math, but it doesn't really know what it's doing. It's just making predictions. So to me, those are the two differences. What about you, Vincent? What do you think? Vincent Somsen: Yeah, no, certainly I agree with that. I think most of the people, uh, it's important to understand, um, you know, that these, you know, Technologies, except for generative [00:05:00] AI that has been relatively new. The AI in general has been around, you know, for the last 20, 25 years plus, right? Only now that we have seen some transformations, particularly from, and, uh, revelations from open AI in the last, uh, in the last, uh, 18 months, basically, it has really picked up and made tremendous, uh, strides, so to say. But it has been around already. Is this not something coming out of the, of the blue sky, right? Adam Mogelonsky: is this new term that is getting a lot more buzz now is co piloting. And can you, Vincent, describe what that means in the capacity of a hotel executive that is working alongside an AI? Vincent Somsen: Yeah, certainly. So think of CoPilot as an assistant next to your daily tasks. One of the ways that CoPilot, the term [00:06:00] CoPilot, one of the reasons why it's so popular, this term, is because Microsoft published this product out, right? It has been integrated already with most of Microsoft tools, such as Microsoft Excel, and it's going to integrate more and more, right? you can do your everyday tasks, um, let's say create an Excel sheet, uh, or a presentation, it's going to take you maybe one and a half, two hours, uh, whatever, 30 minutes, it doesn't really matter. It's going to learn from your previous decisions, from your previous designs, whether it's an Excel sheet or a presentation, and it's going to speed up this creation process, right? Uh, I believe that, you know, for me, I would have come maybe to the same conclusion or the same work, um, as I did without, uh, the help of a co pilot, uh, product or feature. but it's going to speed up that, that, that process. In some cases, uh, the more you train it, right, the, the co pilot on your decisions, on [00:07:00] your patterns, it can also, uh, get potential customers. You know, increase creative solutions. So, um, learn from your creative thinking and your designs, and create or make decisions, that really follow your patterns and your decisions in the past. Adam Mogelonsky: So getting into specific use cases now, cause we got to drill down to some specifics aside from just dwelling on some theory. Ken, Michael, back to you. This problem of too much information in a hotel, too much data, it's siloed. You have guest data, business intelligence based on capital assets, labor data, all these things that need to be inputted you Can you offer us some sort of specific examples of how AI is being used in the realms of on prem financial reporting for the property level, above property [00:08:00] portfolio analysis, or any other types of business intelligence? Michael Goldrich: Uh, it definitely. So, um, so I'm currently conducting primary research on behalf of the HSMAI Foundation, and specifically with this, I'm being asked to look at sales, marketing, and revenue management. How they're using generative AI in terms of recruitment, retention, and professional development. And as I'm talking to some of the sales leaders, they're talking about how they're leveraging the Salesforce Einstein that actually has a component of AI or dynamics. And what they say is that is act is able to collapse time because, uh, in the past where they might need to get an administrator. To kind of do an SQL query for them, And they actually have to have the business analyst go, and so the sales leader will say, okay, this is the type of query I want. They bring it to the expert that can do it, and then [00:09:00] deliver the content back, and then they say, well, that's not just exactly right. Go back and forth a couple of times. Now the sales leaders can do that on their own, and cut out that sort of back and forth. So there's a time savings here. in addition, there's, uh, apart from what you're looking for as a, uh, application only, maybe a business case, there is sort of a side case where it's about, as I focus on this retention piece of professional development, it's about, um, work, and how people like work, and what is their work, and you have work that's composed of two different parts. You have joy, and you have toil. And the BCG said if you can have 10 hours or more of joy a week, you're not going to leave that job. But if you have 4 hours or more of toil, you're probably going to think about leaving. So you can leverage this technology to maximize joy and minimize [00:10:00] toil. And most of the toil is in the administrative tasks. So if you can find a way to leverage this technology to minimize the toil, you'll retain the people, you'll actually have the cost savings of not having to go find someone else which is two times their salary. So that is a, a real, real case of euthanasia. Adam Mogelonsky: So essentially what you're saying is that Utilizing technology is a way to preserve your teams by making their employee experience better. Michael Goldrich: A hundred percent. It's actually called the joy effect. Adam Mogelonsky: The joy Michael Goldrich: If you make it, if you make them more joyful, uh, they're happier, less likely to leave. Adam Mogelonsky: So, Vincent, over to you. Any other specific applications in hotel operations for artificial intelligence that can enhance this joy [00:11:00] effect? Mm Vincent Somsen: Well, I think it's, what we talked about these, these mundane tasks, right? Uh, you know, whether it's an agent or a, you know, corporate director of a, of a region of one of the big brands, right. It's going to bring joy to your life, whether you're an agent or a corporate director. I've seen it firsthand and many people that we talk to, for example, at Hitech, are also wondering, okay, how can I use this as a, you know, certain executive? And it can be done in this, in a very simple setting, such as emailing, whether it's your emailing, um, you know, within your own team or to a guest, for example. Or, um, as an agent or concierge, reviewing, uh, responding to reviews, right? Um, I'm not saying take out the entire manual process of responding to reviews. Reviews, which you can, for example, enrich the quality of that review, make it more personal, make it more tailored to that specific experience, bring together certain data sets [00:12:00] from all different departments, as a general manager for example, and discovering patterns and Uh, from those patterns, uh, creating forecasts so you can, uh, you know, increase your resource building. it speaks to the various basic, uh, use cases of Gen AI. For example, creating a simple email, but also really high level strategic decisions that you can make as an executive or maybe a general manager, uh, as a hotel. so across all sections of hotels, AI can do something for you. Adam Mogelonsky: Cool. Well, you mentioned forecasting and the, uh, The term underneath that is labor, labor, labor, which all hotels or most hotels are having difficulties with in various departments. I'm wondering if you could discuss some of the ways that AI is being used to help with the hiring process. And, uh, we've, we've touched on the joy effect, which [00:13:00] is the maintenance of the teams, but now looking at the hiring process and how AI impacts that specifically. Vincent Somsen: Yeah. So for example, you can use artificial intelligence, carefully, right? Uh, with background checks, interviews, um, you can, for example, create schedules even after the staff is being hired. Right. Uh, currently and still at the largest brands, you know, in, uh, in our industry, creating schedules for, you know, hotels that have 800, 1000 plus rooms, right? Um, the ones that I worked and did, it was still manual, right? And you're talking about, staff, uh, schedules that involve, uh, I mean, perhaps for a banqueting event, uh, you know, 150, 200, 300 people on a, on a schedule, and that was still being done manually, right? Um, so Deploying AI on creation of staff schedules is a very clear use case that has been proven, uh, to work and there are specific vendors that you can [00:14:00] reach out today, that are available. uh, in terms of, uh, retention as well, I mean, it's one thing to hire staff, but retaining is really where the, where the gems are in our, in our industry, I believe. Um, and. Not only retaining the staff member, but also the knowledge that they create, right? So whenever, you know, Joe, the concierge, leaves, uh, you know, the, the Grand Hotel in New York, also a lot of the knowledge that he has created for the hotel, uh, usually leaves. The same goes for a maintenance manager or even a general manager. so there are currently ways to create digital twins of staff members. Uh, in hotels, that basically retain all the institutional knowledge that, uh, a staff member, uh, creates. So that the next concierge or the next general manager can actually build on top of the knowledge that, these folks, uh, create and not are, uh, that they're not left to, uh, to do this all by themselves, and start from the bottom. Adam Mogelonsky: I think that's a good point to insert another term [00:15:00] worth explaining with this AI revolution, which is, uh, the term is summarization. And I'm wondering, Michael, could you explain that term and how, how that would. It, that would result from or intersect with these digital twins of a manager or person in terms of their knowledge base. Michael Goldrich: Yeah, so in terms, so generative AI has a number of capabilities. It's able to generate text, it's able to classify, and it's also able to summarize. And what it does is it can read through the text that you provided and then based on however you ask it, it will summarize it however you want it to be by pulling out the key relevant points. But, uh, something that Vincent was talking about earlier, uh, regarding, uh, recruitment. So from the people I was talking to, one of the ways they're using this generative AI is in terms of writing job descriptions and making better job descriptions. And the other thing is part of the interview process. So in terms of you have a job description [00:16:00] that the people are hired for, you can actually have their LinkedIn profile downloaded. You can have their resume downloaded, and then you can prompt it to basically say, here's the role, here's the person's background. What are five questions that I should ask that will make sure that this person is the best fit for this? So you can actually use that to help to essentially sort of summarize their skills. And based off the job description and then say, okay, uh, let's look for the gaps. Let's look for the opportunities of where this person might be successful. So that's another way to kind of look at it. But in terms of summarization, it's greater than also categorization. So, uh, which is not something that Vincent talked about, but what, that's one of the great things about this technology is let's say you have a hundred things. And you say to this technology, put into two buckets, it'll put into two buckets, and it'll make sense. And you say, put into ten buckets, and it'll find a way to find each of those ten items into ten different buckets. So, it's able to look at the patterns, which we were [00:17:00] talking about before is the mean machine learning, and able to kind of find some sort of connection between them to kind of, Organize it. So it's all about, and when you think about summarization, it's all about organize, organizing ideas to the best way, possible based on what was provided. Adam Mogelonsky: Okay, so I want to be devil's advocate at this point because we're talking about pattern recognition and getting into the correlation versus causation issue and the term that I see emerging for those who are somewhat Fearful or apprehensive about artificial intelligence is generalization or the generalizability of, machine learning or generative ai. So you have the pattern from machine learning that is based off of a. Off of a database and based on what, you know, within that database, you're making a conclusion that [00:18:00] fits that data. And then from there, based on human assumptions that are inborn in that data set, it is actually fairly narrow cast and you cannot reasonably use it for, to make predictive analytics or insights or actions that are outside of that specific use case. And this, I've seen this come up as a problem because, you know, the whole idea of, oh, we're going to look at somebody's LinkedIn profile and their CV, and then we're going to make it, We're going to use that to make questions. And then I've heard people come back and say, well, the LinkedIn profile doesn't really describe who that person really is. It doesn't really show their passion for hospitality. So how would you respond to that? And Vincent, I guess I'll go over to you. How would you respond to this, this apprehensiveness over the generalization of patterns based off machine learning? Vincent Somsen: I think it's a good point. You can't base an entire [00:19:00] decision of, um, the decision of, uh, pattern that AI establishes, and that's why it's crucial to, to still involve and interview and still in to, deploy or employ critical thinking, uh, on these, um, on these decisions or, that the AI suggests. Right? Um, at least for now, right. I think in the future when, AI has already been, um, trained on a lot of data that is out here. but there hasn't been a time that, that there has been, the generation of data, uh, and the creation of data, uh, hasn't been more than, than now, right, than present day. so the more we feed it, the more it's going to be, uh, intelligent, hopefully. but it's always good to check these decisions and, not to only rely on, uh, on these tools. but I think what Michael also says, you're anyway going to do research. It's going to take you anyway. Half an hour, [00:20:00] an hour, um, at least, why not employ AI to do this and provide some backgrounds that you're anyway going to do, some research that you're anyway going to do. Adam Mogelonsky: Yeah. So we're saving minutes and Incrementally, that adds up to something very profound, whether that is in HR or another department. Okay. So, back to use cases, one area where we can segue from staffing and operations is into food and beverage, where staffing is so important all the way from the server or line cook up to restaurant manager, and I'll pose it to both of you based on your expertise. Where are you seeing artificial intelligence impact food and beverage operations and marketing as well? Vincent Somsen: I can start the, um, where, where I see it. Um, and it has been deployed for a number of years is really inventory management and waste management, right? [00:21:00] the majority, the vast majority of restaurants and F&B kitchens in hotels, um, they have no idea what they're wasting, right? and Using AI, uh, and, you know, normal, Technology that has been around for, uh, for decades, um, uh, putting a layer of AI on top of that, it can be very simple. There are vendors out here, uh, that the majority of the hospitality tech, uh, tech professionals know, using, uh, garbage bins, uh, Uh, with a, uh, with a skill, that you can, you know, weigh your food on, uh, throw it, uh, throw it in the garbage bin, um, and it takes a picture, uh, before that, and it uses image recognition, right? To understand what you're throwing away, how, um, uh, what it weighs, you know, five kilos of potato, peels, whatever, and establishes a pattern over time, uh, and recommends you certain, um, uh, procedures in order to, uh, reduce waste, what can you do with the waste, and, it also really [00:22:00] starts with knowing How much waste you produce and how much it's costing you, right? And measuring that is already going to provide you a lot of really practical insights to do so. But inventory management, the same. What do you think, Michael? Michael Goldrich: so I'll, I'll take a little bit of a different angle in terms of the human labor. So answering the telephone, but right now people have to answer the phone, take an order and, you know, say it's for room service. Uh, that is now actually being automated by AI voice agents and they can answer it pretty much immediately. They get it right and they can kind of collect all the information that people are asking for. Then they can automate it, send a ticket and actually, you know, work with another company in which a robot then delivers it to your room. So that could all be an automated streamlined process. So that is one way in terms of F and B, you can leverage this technology, but in terms of marketing. Uh, you have a lot of [00:23:00] tools that are all about targeting, say, uh, personalization and, but usually there's a little bit of a learning curve and sometimes people aren't the greatest writers, so if you're going to personalize it, sometimes you may not know the best way to do it if you have a junior staffer, so some of these technologies are using the machine learning, using this generative AI, say, okay, what is the person that you want to target? All you have to do is basically say, okay, I want a. Business Traveler or I want a couple and then based on that it'll just create essentially a full campaign for you that you can then deploy across your website. So before you would have to think about all the different branches, think about how you would write the text. Now it can essentially be all auto generated for you. And after it's done, the reports that are created where you stop to kind of look through it and you may not necessarily fully understand it. Now you can query these reports to understand how effective these solution was. Adam Mogelonsky: Okay, so we're getting into querying [00:24:00] and we're talking now getting back to the executive level and I want to throw another buzz term that's emerging, which is RAG or Retrieval Augmented Generation. Michael, back to you. Could you define what this term means and then give us sort of its, utility for hotel executives? Michael Goldrich: Sure. So right. So Just to kind of get back to what it is, so, um, understand the issue that derived this need for this. So there, this technology hallucinates in terms of what hallucination means. Some people just say it's lying, but you can't say it's lying because that means you're intending to mislead. This technology is not lying. What it is, is based on probability, it's determining the next best word, and it'll go across all of that. And It's using it based off whatever data set it has to basically come up with those sort of [00:25:00] calculations. So what RAG does is basically it looks at the answers that are given and fine tuning it to help it come up with more accurate responses in terms of just coming up with an answer. Because this technology inherently How it seems to be designed, it wants to please the person that's doing the prompt. So it, sort of understands your intent of what you want it to say. So it's going to kind of give you an answer based off of that. And so that way it could be misleading because it can validate your initial opinion. And what this rag does is essentially it's fine tuning. And basically it's taking the data, taking information and adding another layer. To it, in terms of before it delivers a response, kind of checking it out and making sure that, yeah, okay, is this accurate what it's saying, and it's sort of like a second level validation, so to speak. Adam Mogelonsky: [00:26:00] Okay. how does that help hotel executives? Let's say either those, the marketing automation we mentioned for food and beverage or something in the realm of financial reporting, you Michael Goldrich: you feel like the answers that you're getting have, Uh, significantly less value. Uh, Hallucinations, so they're more accurate. I mean, that is, that's it. It's basically, it's a trust factor. So, uh, as we were talking about a little bit before, Vincent was saying about being a, you know, you have to be critical thinkers on this. It's really important ultimately to have a human in the loop. you always need to have a human there to kind of make sure because I was talking to a data scientist. He said, even if we're the best at what we do, the data is never going to be more than 93 percent accurate. There's always going to be problems with the data, always, no matter what. So you always have to have somebody there to think about it. And this fine tuning, this rag will make answers better. But they won't, still won't be perfect, and so it's just [00:27:00] another layer of sort of, um, ensuring that it is a better answer, but it's still the best they can do, but it may not be the most accurate. This still may be a little bit, incorrect. What do you think, Vincent? Vincent Somsen: Yeah, I think you made a very good point because without data, without ensuring that the quality of data and integrations and um, You call it like, sure, you can do very simple things such as, you know, utilizing chatGPT to write, uh, you know, your emails or responding to reviews, but to really take it to the next level, what we're talking about, I think you really need to ensure that you have structured and enriched and, uh, high quality data sets. Uh, before that, it's going to be very hard to make, impactful, um, transforming. Impacts to your organization, right? that's, you know, my belief. Um, you know, and it goes for any technology, uh, to be frank, that you deploy on the property or [00:28:00] headquarters level. Adam Mogelonsky: know, the, the whole idea of quality data sets sort of goes back to my, my whole thought about this generalizability, there I said it properly, and that whole problem that makes people apprehensive. And. Before we move into the future, I'm wondering, because it seems like hotel executives, the future, They sort of need to know a little bit about AI, they should, but AI is based on statistics. So what would you advise hotel executives to know about statistics in order to help them make better judgments about using databases and predictive mechanisms based upon those databases? Michael Goldrich: I got a great answer. All right, so here's what I have to say about that. So last week I was in Nashville at the CIO Summit and I was talking to them about this technology and what I was specifically [00:29:00] saying to them, I was talking about the generative AI productivity paradox. And what is that? So you have all this promise of this generative, generative AI that's coming from Gartner. It's coming from Bain. It's coming from BCG. It's coming from McKenzie. You're going to see so much more productivity. You're going to have so much time savings. You're going to have so much revenue. then you see the news. You hear the hype, and you see what's actually happening, and very few of these projects are actually getting out of the pilot stage. In fact, Gartner says only 15 percent are getting out of that phase. And there's a lot of reasons why. It has to do with technology mismatch, it has to do with cost benefit analysis, revenue, all that. but the one thing that can actually help these projects to be successful is AI literacy. And that's to your point, Adam, like you talked about statistics, but it's more than that. It's understanding everything about this technology. It's not just [00:30:00] prompting. It's not just, um, ethics. It's not just the, the tools, it is actually understanding it all together. And then ultimately, you know, a lot of these projects that are coming down are coming down from the CEO or the CTO. But if you take an AI literacy approach for your team and you show them how to identify these use cases, you show them the benefits of the automation. These new projects will actually come instead of top down and go bottom up. And by pushing up, they embrace it because that is ultimately a big challenge in terms of some of these AI projects is that there's employee resistance. Employees aren't engaged. But if they understand that this is going to empower them, that it's going to make them more successful, it'll help their career, uh, they will be advocates and they'll work with everybody. So to me, uh, you talked about statistics. To me, I say it's the whole thing is, I would call [00:31:00] it AI literacy. Adam Mogelonsky: hmm. AI literacy. So looking ahead to the future. And we're talking about the future of hotels, of hotel culture, which perhaps there is some form of continuing professional development that involves teaching hotel teams on AI literacy. Vincent, where do you see the day, a day in the life of a hotel executive five years from now, insofar as how they are AI adjacent or using a co pilot? Vincent Somsen: Yeah, it's a good question that I thought about, as well, and, you know, for the past months already, how, what, what is AI going to do in co piloting, um, for Hotelier and specifically, I think on property level, whether you're a GM or from desk agent, how I would like to see it and how I, Um, I think it's possible is constant suggestions, uh, in order to personalize a stay, in order to personalize, um, you know, [00:32:00] emails and experience, um, but also fast forward certain processes. I'm working on a forecast. I'm working on, um, on a brief for my, for my, uh, morning meeting, right? Uh, with my, uh, department heads, um, You know, that brief being enriched or even created for 80 percent by my co pilot, by my AI assistant. providing certain insights from the last week, right? That usually would take me maybe two hours to create, uh, gets now instant, I get instant access to. So I think in every touchpoint that I do as a hotelier, let's say a general manager of a hotel, um, I constantly get. well put suggestions, um, and certain automation, uh, suggestions, in order to free up my time and spend it more with guests, spend it more with staff, um, create more face time with people that, uh, that matter in the hotel instead of these, uh, you know, [00:33:00] relatively administrative things. yeah, that's how I see it and, and, um, already some, you know, it's already possible to some extent, but, not to do a great extent in an overhauling, flow, so to say. Adam Mogelonsky: So it's, it's a constant Evolution. It's not this punctuated evolution, you know, like X Men, like all of a sudden you're a mutant. It's, you know, you apply one thing, one thing at a time, and it's part of the culture. as Michael said, it has to be part from the ground, from the bottom up. It's just part there, and everyone is looking for newer and newer and better and better use cases. So are there any hotel brands that Have this, have embraced AI truly, not just as part of a press release, they've actually embraced it and are doing cool things under the hood or overtly to, to help guests and to help their teams. Vincent Somsen: There are certain brands out there. I wouldn't say there's one brand that, um, that, [00:34:00] employed at every department and, and, and whatnot. I think, you know, technology for brands such as Citizen M, um, you know, is, is, uh, is a chain or a boutique chain, I suppose, uh, that you could look, at where, where they. created some really interesting applications, uh, Accor is utilizing, uh, artificial intelligence, as well, for example, guest message, guest messaging, but again, there hasn't been a, uh, a brand that I would say, okay, they created an AI first approach. whether that's good or bad, uh, you know, it really depends on the applications. Um, but there are, there are definitely some interesting developments, uh, that I'm seeing, uh, you know, as a boutique hotel chain, you know, obviously make, uh, a lot of progress in a relatively short amount of time, because you, you know, you don't have as many decision makers and whatnot, so that, that also doesn't help for the larger chains, but there's definitely, uh, yeah, developments. Adam Mogelonsky: Okay. Michael, any, [00:35:00] uh, any brands or companies that you think are on the bleeding edge? Michael Goldrich: You know, from, because I've been talking to a lot of the commercial leaders as part of the research that I'm doing, and to me, there, it is still, this year is sort of new. So like some of them told me, you know what, Michael, it was so new last year, we didn't budget for it this year. So I think now people are going to be budgeting for it for next year. So I think right now there's, uh, people are kind of exploring the tools a little bit on their own, but there's no. consistent, uniform approach across the hotel companies that I've spoken to. I think each department's taking a little bit of a different perspective, and each manager's taking a little bit of a different perspective in leveraging these tools. Some of the tools that they already are using are you know, enhancing them with AI or this generative AI, so they're just kind of automatically getting it. To an extent, but I think that, um, as I talked about before, [00:36:00] like outside of that, people that are looking into automations may or may not be working, may or may not be succeeding, as anticipated. But, you know, I, um, I think Atlantis has done a really cool job with their, their chatbot. Uh, they've built that themselves and it is really powerful. and there are a couple of other, uh, sort of, uh, technology companies that have really done a good job of it. Enhancing their tools with this. But, I think as Vincent was saying before, like an AI approach, AI forward leading company, I haven't seen one that stands out yet, maybe individual hotels, smaller hotels where teams are, can be much more nimble, but in terms of teams, uh, hotels that have more than two hotels, I haven't seen anything yet, doesn't mean it's not out there. I just haven't seen it. Adam Mogelonsky: Yeah. I mean, so it, it seems like there still is so much. We've looked at the opportunity here for hotels to [00:37:00] use AI to get ahead, whether that's on the staffing efficiency side, the team retention, uh, or any sort of new marketing or innovations on the guest facing side to ramp up revenues. So to close out here, one final question. we've looked at hotel brands, we've looked at the opportunity. Are there any other types of artificial intelligence that hoteliers should be cognizant of? Michael Goldrich: Yeah, so I'll go first. So, uh, not necessarily type, but agents. So we know that Apple is going to be rolling out Apple intelligence, sort of like Q1 of next year. What does that mean? That means that this phone of ours that we have it knows everything about us, right? It knows our personal life, it knows our professional life, it knows our emails, it knows our texts, it knows our calendars. So I think initially what's going to happen is people are going to be using these to do simple things, like helping me. Get [00:38:00] movie tickets. And people get comfortable with the kind of doing that and it would be, it's called an agent. Now once they get comfortable with that, they're going to give a more complex task. Schedule this meeting for me. Schedule this trip for me. Now when it starts to schedule the trips for you, that is when it gets interesting for us as hoteliers because where is it getting that information from? And what is the person reviewing to validate that? So what's going to have to happen as people get more comfortable with their phone? Doing all of the, uh, sort of like hotel booking. So as a hotel that wants to get my eye, eye attention, if I'm not going to the website, if I'm just looking at my agent, the hotels need to figure out how to get my agent's attention and that's going to book the hotel. So that's going to be something that's going to be coming down and that's something new that hotels need to think about Adam Mogelonsky: so we're talking about intelligent agents and there's a layer of conversational AI, uh, in there because it's essentially you're, you're talking to Siri, right? [00:39:00] And Siri's pulling all this personal information and then mixing it with what's out there in the web. do you see. Companies like Apple, and assuming Google will come out with a competitive version for Android. Do you see them monetizing this? Like basically saying, okay, well, if you want our, Apple agent to recommend things, you're going to have to pay to get to the top of the pile. Michael Goldrich: My opinion is that there'll be a new distribution channel in which it goes to look first. So if I'm looking to book a hotel, it goes to this, whatever this is first. It won't necessarily be Expedia booking, it won't even be necessarily direct. There'll be all the hotels that kind of line, decide to line up to be on that distribution channel. And I think they'll get a piece of that, would be my thought. Adam Mogelonsky: Yeah. I mean, it's yet one more channel that, revenue managers have to consider as part of the segmentation. Vincent, any, anything else that you see on the horizon? Vincent Somsen: yeah, I think actually to build on top of [00:40:00] that, it's, I think it's going to look much like the SEO optimization. You're gonna, you're gonna see AI optimization, uh, I think for your products, for your businesses. And if you want to be AI optimized, you're gonna have to pay premium or you're gonna have to implement some sort of first smart strategy in order to get into the spotlight of these, uh, these algorithms. That's why I think it's super important. That's why there's a lot of debate on open AI, right? It started as a non profit and now for profit. That's why I think, a lot of these technologies should be a public good, right? So it can be checked, it can be verified by the public, because it's going to have such a profound and it already has such a profound impact on our lives and even more as a consumer, right? and everything that you don't pay for, you know, you're the product. Right? So, that's something to, to always keep in mind, I think. yeah. Adam Mogelonsky: Yeah, 20 years of, uh, searching on Google for free and 20 years of Facebook likes and image posting. You are the [00:41:00] product. this has been a fantastic conversation. A lot to think about for hoteliers and any other technology company that's listening or any other hotel brand that wants to innovate. Vincent, Michael, thank you so much for your time. Vincent Somsen: Likewise, Adam. Thank you. Michael Goldrich: Thank you.