AI That Works: How Hotel Leaders Move From Experimentation to Results - Steven Moore & Joseph Benjamin, Actabl [Sponsor Bonus]

Josiah: All right. Hello, everyone, and welcome to today's webinar on hotel AI that works. I'm your host, Josiah Mackenzie. Today's session is about helping you understand what you need to know, think, and do to make the most of this moment, make the most of this exciting new technology.

This is going to be a conversation that is focused on things that you need to know now, and I wanted to bring together two incredible leaders in our industry for this conversation: Steven Moore, who is CEO of Actabl, and Joseph Benjamin, who is our chief technology officer. Both of them have a really unique perspective because at Actabl, we help more than 90% of the 50 largest hotel management companies in the US. More than 180,000 hoteliers every day at every level of hotel businesses are using Actabl technology to delight guests, to make life better for the people they work with in the business of hotels, and ultimately generate more profitability for owners so that more capital moves into the hotel business, and then we can fund all the things that we love about the hotel business.

This is a really unique moment. Many of you are preparing, going through budget season. You're looking at the advances of AI capabilities. You may have spoken with us at the recent HITEC conference, but a lot has changed in the weeks since then. So we wanted to bring Steven and Joseph together for a conversation on what you need to think about and what you need to do.

We're going to start with some data from AHLA talking about GOPPAR being below pre-pandemic levels, 90% of what it was in 2019. You have hotel operating costs that are rising faster than revenue. The US Bureau of Labor Statistics showed that employee quit rates in hospitality are two times higher than other private sector employers.

And under all of this, there are many challenges, but there are opportunities. I think more recently we have seen top line growth show some encouraging signs. We ran a poll recently at Actabl of hotel leaders, and we found that the number one challenge they were facing was taking that top line growth and making sure that it flows through to the bottom line.

So with that, Steven, I wanted to go to you first. Thank you for joining us today for this conversation. What are you seeing and hearing as you're talking with other hotel leaders out there?

Steven: Yeah. So I think the first thing that we're seeing is excitement, which is great. So people are very bullish on the potential of AI to help deliver hospitality to guests, which is great. There's a lot of momentum on experimenting with things and testing what could be beneficial for associates, for guests, above property, on property, so that's great.

And then there's certainly a sense of risk, right, with any new technology. What's the security look like? What's the implementation look like? What's real? What's a hallucination? What will it cost? What's the structure of these things going to be?

And then I talk a lot about the fragmentation that's existed in the industry for a while on the technology side. I think AI has the potential to only accelerate that in a negative way if we're not careful, just the proliferation of people running wild with their Claude dashboards and unique data sources and things being out of sync. So there's some risk on the fragmentation 2.0, but overall a lot of momentum and a lot of excitement.

Josiah: A lot of things to be excited about. And Joseph, I want to go to you in a moment to talk about that. But Steven, just staying with you for a moment, I've heard you describe this moment as a unique opportunity to reset, as you put it, the profitability floor across a portfolio of hotels. I wonder if you could share with our listeners a little bit more about what you mean by that.

Steven: Yeah. I think it's a really exciting opportunity, and not just to recover the compression that we've seen through COVID and post-COVID, but actually reset to a higher point than we've ever been before as an industry.

And I think if you consider profitable efficiency, it's this equation. So you need timely insights times quality insights times ability to execute against those insights. That equals profitable efficiency. And so if you think about what AI can do to each part of that equation, you can have AI monitoring and elevating insights 24/7 so that you're always getting the right insight at the right time. So that's the timeliness.

It can just ingest more information than any single analyst could or any single piece of software could. So it's constantly ingesting and thinking about what the quality insight is or the unique insight may be for your hotel, and then it queues up the action so you can know that, hey, across all of your hotels, this is what's going on in this specific hotel or this specific category. Here are the drivers, here's what it's going to cost you, here's the recommended action, right? Would you actually like to take action on that and execute across the insight?

So timeliness, quality, ability to execute. AI is just gonna amplify each of those categories or variables to help reset the profitability floor.

Josiah: A lot to be excited about. And Joseph, I wanna go to you because I've appreciated watching you build, not only with our teams, but with our customers, and having an approach of always looking at the latest and greatest capabilities, and not just building things for our customers, but building with customers.

And I'm curious from your vantage point. Many people were at HITEC, the big technology trade show. We had a lot of conversations there. Feels like every day you're building new capabilities with our teams. From the technology side, what are you seeing that is exciting to you to unlock some of these capabilities that Steven spoke to?

Joseph: Yeah, Steven brought up a lot of great points there. I think the biggest change for me is that we're moving beyond using AI to be something you ask a question about. The models have gotten much better, but the bigger development and more excitement for me is everything around the models. So a lot of the reasoning, a lot of the tooling, the structured workflows, evaluations, and the ability to connect AI safely to enterprise systems. Those, to me, are the biggest changes that are happening at the foundation level that are really enabling us to build a lot more capabilities into our products and with our customers.

And so it means that we can start building AI that understands a question, determines what information it needs, goes and gets that information from a trusted system, and applies the correct business context to reinforce what Steven was talking about, to find the insight to be able to take action with confidence.

So I think we're seeing this really firsthand. A year ago, a lot of this was prototypes and ideas, and today we're putting a lot of these real capabilities into production systems like Altitude and like HE Insights and in front of real hotel operators who are learning from the systems that we're putting out there. So that, to me, has been the most exciting and why I think the opportunity is so, so strong right now.

Josiah: It certainly seems that way, and if we go to the next slide, we actually have a visualization that I think will bring this to life. And I want to go to you first. There are many different types of AI out there. You talked about these different components of how we are thinking about driving results for our customers. We've framed it around this category of operational AI and all the different areas of a hotel business, all the parts of a hotel business that you could apply AI in. I'd love your take as CEO talking to other CEOs, other leaders, what is interesting when we think about applying AI to the operations of a hotel business.

Steven: Yeah. So I think there's lots of places you can go with that. There's lots of opportunities. So I talked about the always-on nature of insights. There's budgeting and forecasting accuracy, which helps you plan better, helps you schedule better. There's execution efficiency. So when you have those insights, are you scheduling your labor effectively? Are you maintaining your assets? Are you coordinating different departments the right way at a hotel?

I think one of the most underappreciated opportunities of implementing AI is allowing hoteliers to finally spend more time with the guests or spend more time on things that serve the guests. And I think it's underappreciated just because it's been kind of a tired thing the tech vendors have said for a long time: "Oh, our tech helps you spend more time with the guests or on things that impact the guests."

And theoretically, that's true. I don't know if we're honest with ourselves if we've quite delivered that. In some cases, I think we have, but I think in some cases, technology can be somewhat of a burden, right? You're trying to track things to measure against them, track for accountability, track for improvement in how you may operate. All important things, but it doesn't necessarily give boots on the ground, people delivering hospitality, more time to interact with the guests or serve the guests.

And AI can automate so much of the digital workflow because it can ease so much of that burden to keep people from actually having to be in the system as much. I think it frees people up to, again, spend more time doing the things that got them engaged in the hospitality industry to begin with.

And so you think about that from the associate experience. That leads to higher retention, that leads to a better work environment, that leads to a better guest experience. All of those things lead to higher profitability at the end of the day. So I'm really excited about the retention aspect of this for the people that are working in hotels today, so that tech vendors can finally deliver on that promise to give people more time back to spend with the guests.

Josiah: It's exciting to think about finally being able to deliver on that promise, because the promise has always been exciting. Just the technology capabilities haven't been in a place where you can deliver on this.

And Joseph, you spoke about this a little bit a few moments ago, but I'd love your take on this too, because you've led very large technology organizations across industries, and you've seen the promise and sometimes the potential of some elements of technology deliver, but not always, and not in the direction maybe that Steven described. As you think about the different moving pieces here, what's possible with today's generation of AI-powered technology? What excites you when you think about this flywheel and this dynamic being able to be built?

Joseph: Yeah. I think what Steven brings up is that maybe the promise of technology hasn't really resulted in giving back time. It's created more work or overhead in some cases.

And I think one of the challenges, especially as it relates to our customers, what I'm seeing, is that the expertise and the attention and time don't scale equally across every property, right? At every property, there's a lot that's similar, but there's also those pieces that are unique to that property. And I think with AI, there's the ability to personalize that a bit at the property level in a much more efficient way. And again, it goes back to the data and the right context at the right time.

And so I think with this generation of technologies, the promise has real potential to pay dividends in that way of freeing time, because the technology can continuously watch the operations, the numbers, the data. It can automatically identify where something needs attention, that needs to be acted on. It can give the person enough context so that they're not spending time searching around trying to understand what to do about it. Eventually, that all leads them to take faster action with higher confidence. And so all of those things coming together really gives us the opportunity to close the gap and really deliver on the promise.

Josiah: It's exciting. I have spent my whole career in hospitality and most of it in technology, and I think the promise of working in hospitality is to spend more time with your guests, to delight them, as you mentioned, Steven. And then what goes into actually delivering that is a lot of the elements of what you described, Joseph, right? Whether you think about it as a leader, how do I encourage people to serve guests out of the way that they're wired, how fulfilling it is to act and operate in that way, and then when you think about the business impact as well, the financial impact.

Steven, I know at a recent conference you were sharing some remarks about how you look at the economics of a hotel business and where the largest cost centers are, what are the biggest drivers of financial returns. And a lot of it does come down to the operations. So if I hear you both and think about the opportunities that you raise, there's the feel-good side of it, how do you inspire people, make their life better, but also how do you generate better returns? And a lot of that exists on the operations side. So I think operational AI is one of the biggest areas of opportunity for us.

And so I would love to get into how do we get there? What does it actually look like to deliver on this? I know we have a quote up here from Dan from Loews Hotels talking about AI being only as good as the quality, the accessibility, and the connectedness of the data and systems underneath it.

Joseph, I've heard you talk so much about this recently, about anything you wanna do, it all starts with data, and this is often a starting point when you are working with hotel leaders who want to stand up these capabilities. And I wanna get your take on this. Why does it all start with data? Why is this the foundation?

Joseph: It's a great question, and I think it gets talked about a lot without enough of the why. And so you can really understand it for anyone who's using chat today, right? The variability of answers you can get. You can even ask the same question twice and get different answers.

And so the criticality of making sure that data is of high quality, cleansed correctly, it's normalized, that the relationships between different data sets are clearly articulated in the database, if you will, becomes even more apparent as you start turning AI to start scanning the data and looking and giving answers.

And so the difference is, if I'm asking a general question to an LLM, which knows an extraordinary amount about the entire world, it may be useful to get a pretty good answer. But if I'm sitting with an owner and discussing why a property missed budget by $200,000, approximate isn't useful. You need to have the exactly correct answer every time you ask the question, or you start creating lots of new problems.

And so maybe this idea of operational AI versus general AI: operational AI has to be constrained. It has to be grounded in the data to be able to produce the right answer, not an approximate answer. One, to make sure that you're getting the correct information, and even more importantly, when we go to start taking action against those answers, we need to make sure and have high confidence that the data and how the AI is acting is 100% accurate. And so hopefully that's just one example of why this is so critical and even more important than it has been in the past.

Josiah: It's great, 'cause we want that end state, right? But you have to build the foundation to get to that and enable this.

And Steven, I'm thinking back to what you mentioned at the beginning of this conversation about the excitement about AI. There's a lot of people, whether it's creating dashboards or using it in other ways. As CEO or as a business leader, how can you think about balancing wanting to make the most of the moment, capture the excitement, move it in a positive direction, but also thinking about what do we need to have as our foundation to help us win today but also move forward? It seems like that's a tough balancing act. Any thoughts on that?

Steven: Yeah. I think there is a lot of goodness in moving quickly, right? So there's this illustration. I forget where I first heard it, but this is not original. It's the pottery class paradox. At the beginning of the semester, the teacher broke the class up into two groups, and to the first group, he said, "Hey, you need to make one perfect pot by the end of the semester." To the second group, he said, "Hey, you need to make as many pots as possible by the end of the semester."

And probably not surprising, at the end of the semester, the most creative, the most beautiful, the highest quality pot came from the second group that had to make the most pots by the end of the semester. The idea just being that you learn by doing.

And so I think there is a lot of value in moving quickly, in testing things, in seeing what works. And at the same time, you've gotta validate. We're out of the party trick phase of AI. You need to make sure there's some ROI here on the things that you're testing, and you need to make sure there's no irreversible decisions, right?

You need to set up some guardrails around security, around what you're allowing AI to do, the access that you're allowed to give it. So I'd say yes, full steam ahead, move quickly, but think about the ways that you're going to validate those things. Think about the guardrails that you're going to set up as you start to move from experimentation into a more architecture way of thinking.

Josiah: I love that. Let's go to the next slide. I want to bring this to life with some examples of how we're thinking about this at Actabl. We have capabilities at every level, and our vision is to see AI working for you every day, at every level of your company, doing work, helping your teams do work in the products that you use through new capabilities, and with that data foundation underneath all of it.

Joseph, I would love to get your take on connecting the pieces. We'll talk about how you build your advantage, but just reflecting for a moment on how we've thought about this at Actabl, how do you think about all these moving pieces working together, from AI in existing products to brand new products? We have Altitude as a new capability, Actabl Data Services. How do you see all these pieces connecting for us? It might help the leaders listening think about these pieces interconnecting in their ecosystem.

Joseph: Yeah. So no surprise, I would start with the build-your-data-foundation piece on this, right? And make sure that you've got connected data so that you can reliably get the information. And you need to make sure it's understood data. So do you know what those numbers actually mean across different properties and systems? And once you can answer those, you can start to use AI for the decision making.

And again, if you're looking for a general, approximate answer, the bar is lower, but if you're actually using AI for decision making, that bar is much higher in terms of having that solid data foundation in place.

And then the other piece would be around digitizing the work. So can you use the AI, and then, based on the insight, prove the impact? So if an insight tells you that your labor budget is over, is it accurate? If not, why not? And I think that iteration cycle is really important to learn, improve, and scale. And as you increase the confidence in the insights that the system is generating, you're in a much better place to enable your team to take action based on those insights without having to second-guess, or worse, undoing actions that you took later that prove to be inaccurate.

And then that final piece on digitizing the work is, once you've generated the insights, once you've proven the impact and you're confident in those, can those insights reach someone where they actually can do something with it? As opposed to maybe having them go find it in a dashboard, can you push that insight to go take action, or look up this issue now, as opposed to maybe later or waiting for them to find it on their own?

Josiah: Steven, I'd like to get your take on this, because we talked before about this knowing, doing, proving the work. But this flywheel nature is interesting as a frame for understanding the current moment, because hotel leaders are always thinking about efficiency. How do we better support our teams? They go through annual budget season or revisit that multiple times throughout the year.

At the same time, it feels to me that there is this moment to not only do that, provide capabilities to help your teams win, to help your business win, but maybe build a compounding flywheel that generates advantages today, but also long-term advantages. I know you spend a lot of time thinking about strategy, and you talk with leaders about strategy. How do you think about this unique moment now?

Steven: Yeah. Well, you said it. I mean, the power of compounding. So you won't have an incremental differentiation from your competitors if you implement this well. It will be a step change, right? It will be an exponential differentiator. And so with the power of compounding, your best friend is time. Start now.

And if you think about what is going to limit your ability to execute well, it's not gonna be the tools, right? It's not gonna be the models. Yes, they're the worst they'll ever be today. They're only going to continue getting better. But if you're waiting for some breakthrough model, you have foresight that is better than mine in terms of, well, the model can't do this, this, and this, I'm just gonna wait until the model can do all of those things. You're going to lose, because I don't think the model's the issue.

I think the limiting factor will be your ability to implement it and your organization's ability to adopt it, and that's just a skill that you develop, again, by doing. You learn by doing. And so the tools in place today are great. You can make a lot of progress today. Yes, they will continue to get better, but you just have to go.

You have to build that muscle so that as they continue to improve over time, as you get more data coming into your data foundation that you're building all this AI on top of, as you extend the reach of your system of action to use AI to drive change at your hotels, you're going to get better and better and better outcomes, as long as you have the muscle to adopt and implement. So power of compounding. You need to start that now.

Josiah: Love it. Before we go, share with our listeners and viewers a couple things to think about. We talked about some of the exciting potential of AI and, when it's put to work in a hotel business, what it can do. Also at this juncture, it's important to keep in mind a couple things to avoid creating risk or for all this to go off the rails, and one of them is safety. There's been a couple safety incidents in the news recently with regards to AI, and I'm curious, Joseph, when you think about advising hotel leaders on how they make sure their teams are using AI safely, what comes to mind for you?

Joseph: You know, it's funny, because these have always been best practices for several years, but I think everything just becomes even more critical with AI, right? So before, we used to talk about shadow IT, and now it's shadow AI, right? Organizations really need to know which tools their employees are using, which systems they're trying to connect to, make sure that sensitive data is being kept out of the tools, and that you're not using any of that data for training.

I think keeping AI and the data in secure environments so that there's no exfiltration of data, that you've got the right guardrails and compliance controls in place, is super critical, because what the LLMs will go do and try and accomplish for you is pretty unbounded if you don't keep a tight set of permissions there.

Controlling third-party exposure is also critical. And again, the theme that I keep coming back to is: verify answers that matter. So really make sure that those answers are grounded in trusted data, show the sources, make sure that human review is still in there, especially where any decisions carry risk. So I think these are great call-outs and worth repeating.

Josiah: Yeah. It's been interesting for me to watch unfold over the past year, or past six months especially, the cost component to this, right? So it's not just a conversation of can the AI do this, which is an important thing to investigate, make sure it works for your business, but also just what does it cost?

And I think when you apply some work to these public LLMs, you have people across the organization maybe building different things, and the cost of getting something done might exceed the last generation of technology. And I know you and your teams think a lot about how we contain costs, and I wonder if you could speak with those listening about how you think about containing costs, and maybe some things for them to think about in this regard as well.

Joseph: You know, I think everyone's excited, rightfully so, to try the latest model, but there is a real cost to carry those. And so I think at the stage we're at, it's really finding the right model for the right job. Not every job or task that you do requires the most expensive model inference.

And so it's really about getting more sophisticated in terms of how we route different jobs to different types of models so that we can find the right cost-per-quality trade-off on almost a per-workflow or per-job basis. And so that's the next piece of this that we're actively working on as we build out our products, as well as internally with some of the work we're doing across product and technology today.

Josiah: Love it. Before we go, I want to give some advice specifically about budget season. A lot of people listening to us are navigating some things that we've talked about as a group before we went on air here. And Steven, I want to go to you. Advice for leaders listening: what should they think about this cycle versus maybe some prior budget season cycles, and what might be new and different things they need to keep in mind this year?

Steven: Yeah. So I mentioned it before, the move from experimentation to architecture. I think the experimentation is great to imagine what's possible, but it can quickly become the Wild West in terms of security, or in terms of the budget, the economics, in terms of integrating a team to accomplish something.

I think of the wisdom: if you want to go fast, go alone. If you want to go far, go together. So how do you integrate all of these AI efficiency opportunities into a singular direction? And so be thinking about the architecture now, not just the experimentation. What does the dream state look like three years from now? And then work backwards from there.

So think about the value of standardizing data and ways of working across your entire portfolio. I think you need to consider the quality of that data in order to drive those valuable insights, and then those valuable insights in order to drive things that you can actually change in your portfolio or at your hotels to drive more profitability.

And then think about the team and the partners that you need to go far in what you build. We've developed a lot of capabilities. We've worked with a lot of different hoteliers, and we have some great proof points and some really interesting momentum. So we'd love to partner and be a part of that conversation this budget season.

Josiah: I love it. Speaking of conversations, Joseph, I know you're in a lot of those. Any practical next step to act on what Steven shared? Is that just it? Reach out, let's have a conversation and dig into this?

Joseph: Yeah, I would. I think we're really coming at this with our customers around: what are the core problems they're trying to solve? What are the outcomes they're ideally looking for? We wanna meet every customer where they're at on the journey, and so different customers are at different levels, and we're trying to make sure that we can accommodate and provide solutions for all those.

And maybe before you contact us, one practical first step is: take one important operating decision and trace it end to end. It probably would illuminate a lot for you and your team. Where does the data come from? Do we trust that data? Does everyone agree what the data means? Who's making the decisions on that data, and where do they get made?

Trying to go through that exercise with one decision will help inform a broader strategy. And as we work through it with you, whether it's providing data foundations, whether it's integrating with some of the product capabilities we're building, I think it makes a much easier conversation and much more helpful to accelerate. Because once you go through that exercise, you can really see where your team and business is ready to apply AI and where your investment needs to go, again, back to some of those core data basics.

Josiah: Awesome. The place you can learn more about these capabilities is actabl.com/ai. We'll make some resources available where you can see more of the capabilities that we've built at Actabl.

But Joseph and Steven, thank you both for sharing your perspective. Across all the teams that I've worked on, I've really enjoyed watching you both lean into this process as the leaders of Actabl and making sure that we're not just creating things in isolation. I see you both in all of these conversations with so many leaders, so many companies, co-building with them, hearing from them what it is going to take to win, and not only sharing our perspective at Actabl across all the hoteliers we serve, but in the different components of what we got into in today's conversation.

There's so much here. So we'll make sure that everybody watching and listening has a copy of these resources, where they can review and discuss with their teams. This is a moment that is really important, and then let's keep the conversation going. Go to actabl.com/ai. You can learn a little bit more about these capabilities. You can request a conversation with us. We'd love to speak with you.

But thank you for joining us today. I've learned a lot, so thanks so much, Joseph and Steven.

Steven: Yeah, it's been great. Thanks for having us. Exciting times.

Joseph: For sure. Thank you.