Your AI Tools Won't Save You. Your Data Will. - GB Sharma, Mosaic Hospitality

GB: My career really started with Highgate Hotels and it started totally by accident. At that point, Highgate was just one hotel, believe it or not, back in 2003. The scope of revenue management was just evolving at that time. It was highly manual, and I got to learn under some really good leaders in that company.

As that company scaled, we were lucky enough to grow with that company. I actually moved to San Francisco three years later from New York, where Highgate bought its first asset, and then we started to grow into a lot more. Highgate Tech Ventures was also evolving, where Highgate's focus was a whole lot more on the technology side of it.

And our own draw was more driven towards it because tech is only as good as its user. So how do you unpack a lot of these things on how you really drive that adoption rate, and don't half it to a five on ten, but truly utilize the system to an eight on ten, nine on ten, and do it better.

That's how we started to really find our jam and make our own differentiation and our own identity in the market. That got us through a lot of those ebb and flows of the market that was happening from 2009, 2010, and driving that massive growth across all the hotels in San Francisco.

Then after that, the next legs of the career, right? You become a VP and then you chase what is that next evolution for you in terms of talent. You grow into a Chief Revenue Officer, a Chief Commercial Officer, and the question always comes, what's next?

Where do you find so you don't get complacent in a company, but you're always ahead of the curve? You're sharpening your pencil in a manner of wanting to stay ahead, but also wanting that ability to build a team underneath you that can do things better than you, faster than you. Because that's the only way a company evolves. So that mindset was always there. I have a student mindset, so I'm always learning. That's how Mosaic came into being of what Mosaic is today.

Josiah: I would love that. And then I'd love to go back and talk about some of the observations you had that led to founding Mosaic. But how would you describe Mosaic for someone who maybe is new to knowing about your company?

GB: Mosaic Hospitality is a commercial management company. We work across branded and independents across 22 markets for owners who are done leaving money on the table. How we do it is we go into the hotels, we fix the commercial engine, we run it, and we own the outcomes. We operate in really three different practices.

The first being commercial, everything that drives revenue. Call it revenue management, sales, digital marketing, e-commerce, whether it's button pushing at a property level or a complete fractional C-suite oversight for a portfolio level.

Second, we also do strategic assessments. Think of it as if you were going to buy an asset, what would all the things you would do in the asset that the incumbent may or may not have. That's a highly deep dive structured audit for highlighting exactly where the performance is leaking, where we have a competitive advantage, how we really stack up against the market, the headwinds, the tailwinds, and the high impact opportunities and the baby steps on how we get there.

The third arm was actually how we started Mosaic, which was technology and hotel tech stacks. Closing the gap between what owners are paying for and what is actually performing, from a marketplace to a core system that is truly more AI enabled.

Josiah: I'm always interested for people starting companies. What did you observe as being broken? You touched on this on a number of levels, but I wonder if we get into this a bit more. You've worked across companies, you've seen a lot of hotels, a lot of different contexts. We're recording this at the end of March 2026. It feels like if you look at the macro picture, there's headwinds, and I just get this feeling in all the conversations I'm having on this show that business as usual isn't going to cut it. We've got to do something differently. But you've been noticing this for a while. What were some of the drivers and some of the dysfunction that you saw that led you to start Mosaic?

GB: When we started Mosaic, the first thing was we live and breathe the tech stack. We know, and also working with so many leading companies, we know how the contracts are written, what is the right MSA, what is the right price point, which technologies have we utilized as a test case in an alpha mode, in a beta mode, et cetera.

Having that base as a foundation, we didn't want to just go and say, here's 20 different recommendations, go and pick one. What we wanted to do was first give them that scope. Hotel Tech Report comes up with this very good report of these are all the systems in the market. And that kind of shows the hotel technology ecosystem is a total cluster. So when you see that together, we know that disjointed technology hub is leading with outdated systems, broken integrations. It is expensive. Connectivity is off. Too many options. Revenue leakage. Product confusion.

How do you solve into those? By having a very clear way of, if I need a system and I need it to work with my existing ecosystem, this is how I need to reset my system. This is the price point that I would be paying now. This is what I would be paying based on the new contracts that we would sign. This is how an MSA should be written.

Think of us like a mini Costco. We went out, we negotiated the price points at, call it X cents on the dollar. We have a very small markup, but with that we also have the ability to make a little profit, yet the end customer has the ability to buy it significantly cheaper than going with that technology system directly. And it comes with better and cleaner contracts.

Second, we can stitch different systems together, which no one can do right now. We are the only company that can do it. For example, you want a PMS company, an RMS company, or some other company. Together we can do opaque pricing where we have the ability to stitch different companies together under one cost per key per month and sell it to the end user.

The technology companies have the ability to come in a little lower on a price point that they won't want to show as publicly available. So we are an extension of the teams in those ways, plus stitching different technology points together that really adds up by having significant savings. Yet they're obligated or taking these X number of systems together, which feeds into all of these companies that even though they're not competitors and don't talk to one another, are being sold as a bundle package.

Josiah: It's so interesting because whether it's on the technology side or the commercial strategy side, it feels like one of the foundational parts of Mosaic is fighting fragmentation. In fragmentation, you have all these little silos, these teams, these technologies operating by themselves, and it's not nearly as powerful as operating together.

Earlier we were talking about how you're helping hotel owners with commercial strategy end to end. I was talking with Patrick Norton at Brittain Resorts and Hotels recently, and it stood out to me. He's Chief Marketing Officer but oversees revenue, sales, and marketing. That combination of all of these functions together allows them to outperform.

It seems like you've built your business on this theory. Tell us a little bit more about this, maybe on the commercial strategy side. Why it's so important to have sales, marketing, and revenue all working in concert.

GB: When revenue management started to evolve, it really evolved out of sales and marketing and reservations. So there was this thing of, oh, now we are going to do these things better in this manner. And each market trains a revenue manager very differently.

When you are in New York City, it was almost where it's 90% transient, it's 10% groups. New York runs the show and if Paul says no, then John is waiting in the line to take my reservation. So there was always that competitive side of driving that.

Then you move to San Francisco and it's a very different market. It is a 40% group house and you have to work with sales and learn to leverage a lot of these relationships in the market to drive the business your way. You still had that luck of having a very big corporate intake coming into the market from 2010 all the way to 2017, 2018.

Then you go to markets like Hawaii and it's very different where practically everything is driven through wholesale relationships and driven through a lot of leverages and static pricing that still goes into books that get published in Japan and other places, how people book packages.

So the mindset was very different and that started adding those little seeded dots that all of these departments need to come together for a unified strategy. One of the things that always came as a gap was how do you build a company or a department where everyone is seeing that same goal?

That was something we had to unpack again on what is that one goal that everyone should relate to and speak to and have that approach. We build that around NOI. Every person in all of our departments is speaking to that one unified goal of driving that NOI component.

Josiah: It's so interesting and so important to ladder everything up to that. I spend a fair amount of time at investment conferences and I feel like I've been increasing my own understanding of that. Everything in the hotel business ladders up. It's a business, it's a real estate business. Understanding how the money flows within the business, regardless of where you are in a hotel, helps you succeed. It's awesome to see everything laddering up to NOI.

Now, some of the structural factors that you described are somewhat timeless, or at least over the years or even decades, some of those factors have remained. We're also in a very dynamic world. Recording this in 2026, I want to help our listeners really understand what they should know and more importantly, what they should act on given some of the new technologies, some of the new opportunities that have emerged.

Maybe starting with discovery. At some recent conferences it's been interesting to see data from Focusright and others talking about the rapid quarter over quarter increase in people turning to LLMs for discovery, and then we have now AI agents acting on behalf of people. It feels like this space is evolving quickly. I've appreciated learning from you on this. Could you explain for our listeners, maybe starting with discovery, what's changing here? What should our listeners be thinking about and doing to succeed in this new world?

GB: The first thing is a lot of people get really taken away, oh my god, the discovery is being impacted. Where we are today, traffic across a lot of our websites are down probably close to 30%. But that discovery part or that decline on the web traffic is not uniform as a whole. The brand searches, the loyalty searches, those are stable.

Where we are seeing the shift is that zero click shift, which means that nearly 60% of all searches end without a click because you're getting that answer directly from an AI search. Think about it. In the past, you went to Google, you asked a question, you got ten different blue links, you clicked on it to get that answer. Now Google has the ability through its AI mode to give you that answer.

So it's not that it's being disrupted. It is more that there's more structured data available for AI to read through, and there's more data aggregation happening on the backend that you have the ability to read through it. The discovery funnel is being hit the hardest. Queries like "what are the best hotels in X" are seeing that guest disruption because Google is answering a lot of those questions for you. That battleground for the discovery shift has truly changed from SEO ranking to AI recommendation for its own visibility.

Josiah: In that world, with this shift happening, what does it take to win in 2026, 2027? What should our listeners be doing right now?

GB: One of the biggest things on our end is ensure that your contextual intelligence is very strong. What we mean by that, a lot of that starts with the FAQs. That is foundational for the websites.

Right now, if we talk about websites being glorified brochures, you're going to look back three years from now when every hotel website is going to be its own AI agent, and it kind of changes the whole look and feel of a website based on how we talk. How did we live like this for the last 20 years of just having a brochure and paying companies X dollars for doing so, with this simple word of schema and SEO on the backend? I think that part of it has got to significantly change.

Going back to where we believe that fundamental shift comes through, contextual intelligence is speaking to a customer in a manner that how you and I are actually conversing now. For example, what is your cancellation policy? This is the cancellation policy. No, you expand it the right way. If you're booking these plans, this is the cancellation policy. If you plan to come in this period and you're booking this, this is what it is. The more structurally available data we put out, the more questions we think through what a customer is truly asking, and we are answering those things and placing it on the website in that same manner.

AI is going to give a lot more citations into those and pull that data out the same way. What is AI today using to read a lot of this information? We talk about Reddit, we talk about Wikipedia. We are even seeing LinkedIn, where you're seeing different posts.

We did an audit for one of the properties as an assessment. It was weird. It under-indexed the property because of its reviews, but it took reviews from BBB. Then we had to question it, why are you taking BBB? And it gave an answer to say, because this hotel has had so many problems, it was an anomaly that so many people complained on BBB, and because of it, it ranked it lower.

So a lot of different things are happening. If you don't do anything, you are in trouble. And if you do something not right, you're still in trouble. I'll give you an example. If Josiah, you are planning to go to Hawaii and it gave you a recommendation of staying in Hotel X and for some reason you just told it, I don't like this hotel and don't recommend it. That property is not going to show up for you in your AI now, unless you have done something truly awesome on the backend as a hotelier to build that case for it, and all these things on how it responds to reviews and how it talks through it. Then you're in a better position to answer it. But otherwise, it is just set and done and it's over.

Josiah: It seems like this is a lot more content that needs to be created. I'm thinking about all the different scenarios, things to talk about, things people might be asking. As a marketer, that feels like a lot more stuff to create. Am I wrong there? And if that's the case, how do you create all that content?

GB: AI is not static. That is the one thing that we have to learn. It is changing constantly. What worked last month is probably irrelevant. So what hotels truly need is an AI champion. That's someone who can test those tools, learn fast, fail fast, build best practices, train the teams, and they're allowed to fail. But they also fail smart.

Things like those are so important. Foundationally it makes sense, but at an execution level it always fails. And that's what we go back to. The first step, before you talk models, prompts, or automation, your first step is to clean that data. Because if the data is bad, the AI will be bad. We go through it in that way of let's go back to the basics. The cleansing of the data before we can take these steps forward into building into this world of AI.

Josiah: Something else that comes up for me when I'm thinking about this is brand and how do you escape just being a commodity? If these agentic AI tools are just searching on criteria, what role, if any, does brand play as part of that process?

GB: One of the things, where it happens, is what we get asked in that same context is where do we start? Because oftentimes it is so theoretical that we miss that knowledge of what do I do. We go back to the foundational layer of how should we be working with AI, and that just starts with auditing first where your system and your data lives.

Learn to understand what your data is made of and clean that bad data. Standardize that structure. And all of these, while it's easy to say it, it is a very tedious task to go through. Then you're connecting the systems, all of that data coming in. Where do we store it centrally? That's a very important part. Who controls that access to that data thereafter. And then finally, then you have AI ingest that data.

Most of our hotels, as of today, go into, "I'll report all these things and I'll just put it in a shared drive and that's it." That's not how AI works. You have to think cloud data warehousing. You need to work with companies like Snowflake or Redshift. That becomes your central home for hotel intelligence, which is really powered by human intelligence.

All of that, where you go to the foundation to store it centrally. It is secure, it is scalable. You have some level of governance on it, and then you can connect the AI tools and models to read into it. We unpack a lot of this first before we even tell hotels, this is your roadmap into 30, 60, 90 for AI. Because the reality is your 30, 60, 90 is just cleansing of the data and building that foundation.

Josiah: I'm going to put up on the screen for those watching us on YouTube an illustration that you posted that stood out to me. I really appreciate how much you educate people on LinkedIn. You had this notion that your hotel's data is the moat. Could you walk through, maybe for our listeners on audio, what is the difference between generic AI and private AI and what do you see playing out here?

GB: Great question. On a generic level, it is all your publicly available data. Every LLM, your ChatGPTs of the world, your Claude, your Gemini, whatever you use. Every LLM today is trained mostly on public data. So if everyone has access to that same level of intelligence, the tool is not the advantage.

Where that advantage starts to come in is who owns better data? Who is structuring it better? Who controls it and who can truly activate it faster. That is the moat. So what is our core foundation to it in the future? It is not about who's using AI because everybody will. The future is going to be truly about who wins the intelligence behind the data.

If you're only using generic tools, you're really renting that intelligence. But if you're building and protecting your own proprietary data layer, you're truly building an asset. And that's the differentiation. Why that matters as a whole is because the same level of public data gives everyone a baseline. The private data creates your separation.

In hotels, that private data can be broken into multiple arenas. Think from your PMS, CRS, CRM, RMS, all of these BI systems, your own guest behavior, your booking pace, the cancellation patterns, the spend behavior, room preferences, pricing elasticity, the length of stay trends, off season versus peak season, upsell conversions.

All of this is coming together and hotels, even today, still treat technology as just an application versus where it is heading tomorrow.

Josiah: I find that fascinating. I want to dig a little further into this. For audio listeners, you have a great line in this image I have up on the screen that the structured proprietary data a hotel owns becomes a secondary valuation layer at exit.

This notion you mentioned before around everything ladders up to NOI, everything ladders up to valuation. This is a really interesting perspective on that. For the investors listening, you need to make sure the people you're working with, the organizations you're working with, are building this into the workflow.

There are clear operational applications of this and then commercial applications as well. But I want to ask you, I've heard people talk about data giving you an edge, it could be an advantage for a while. What specifically have you found that hotels need to work through to get to this point? Everyone says data is very valuable, but what have you found needs to be worked through to get here?

GB: The first thing that comes is we talked about the cleansing of the data. Once you have it, once you have structured it in a manner, the value is not giving away your data. The value is controlling the sharing of your data. What does that mean? Your data stays yours. The access is permission based. It is used in a secured environment, and it can train itself on your own internal models. It can also power outside partnerships from your own team.

Your private data then becomes the leverage that you have even on an exit. Why that matters financially, think of it this way. This is not just a tech play, it is a value creation play. No one can come today and say, oh, we just know this because everyone is learning.

It's like Claude, right? Just two months ago, no one was using it. And now you have got so many people utilizing it and it's doing insane things so much faster that you have the ability to create true value for what Claude can do by just simply training and make a lot of money off it, because there are no true champions. We are in that first phase of adopters.

A clear proprietary data strategy can help hotels forecast better, price better, personalize better. You can retail better, retain a customer better. You can generate a lot more reporting better and yield better. Across the board, that shift in performance equals a better performance that lifts the asset value.

That first mover's advantage in this arena will become very important because no one is doing this. As smart hotel owners start to learn this, they will create a new tier of valuation that does not exist. Because a lot of people just continue to talk about AI. Very few continue to build that foundation that will be worth anything.

Josiah: I love it. I would love to bring this all to life with a case study. A story of maybe a hotel or group that you've worked with. You can share the name or not if you want, but I want to understand what this has looked like from the first touchpoints you had, the first steps you had in working with them, some of the process you took them through, and what were some of the outcomes.

GB: The simplest part, I think a lot of people also talk about this on LinkedIn, but it often goes, even with AI, when we talk about best practices and things. I'm happy to share on screen what we shared in our presentation so you can see what that looks like.

What happens is we did this same thing on what is happening. We talked about what are the seven moves that actually matter. We talked about website reviews. Who is your champion on property? Who's allowed to make those mistakes? How do we go in and address the FAQs? How do we measure that traffic? Because AI is not binary. It's not like ROI, I spend this high because of this. You have to create your own ROI. And how do we challenge every partner? Because everyone today is AI enabled. Just for demo purposes, I put this little star on Mosaic and it says AI forward. I'll buy mosaichospitality.ai and I'll call myself AI forward. So know the art of differentiating between what is there.

Photography is what I want to focus on. Photography is what sells, and that is your first impression outside of what AI is going to give you. Most of these hotels, think of seasonal properties, they have a very different snow season to a mud season, to a summer, spring. Yet all of the photography in most properties is static.

How do we use AI to make those smarter and partner with agencies? We did that in some companies. We said we have that space and how do we bring it to life by doing something different and partnering with agencies. We had another property and we said, why don't we build a better package for Valentine's Day and ideate what that could potentially look like. A lobby may look empty, so let's fill it up and showcase what that social proof looks like versus an empty image we have.

A property that doesn't have a story and it has a pool inside with rooms facing inside. Let's bring the elements of everything that city has to offer inside into artwork and partner with local artists to bring that to life and ideate it. All of these things are simple iterations of what can be done.

We wanted to say our rooms don't have a separation between the living area and the suite, and how do we do it? Well, we can build it very simply with AI. I've got all this space outside, but we don't use water fountains anymore. How about we put a garden space outside? But then the question comes, well, how do I go about doing it? AI helps you ideate a lot of these things faster. It bridges the gap of speed to efficiency a lot more versus thinking on your own terms. Not everyone has the ability to ideate to that same level or that expertise level. Better prompting can help you kick in those places.

That is something we feel very important as a foundational starting point. How to start training your team to be better at prompts and use this to your advantage, whether it's for renovations, purifications, giveaways, promotions, upsells, and over utilization of dead spaces that currently have no value to you.

Josiah: AI as an ideation partner is fascinating. It's also funny about screen shares, you kind of see how people bookmark stuff. I see all the AIs. I see Claude, ChatGPT, Gemini, Meta AI. So it sounds like you're just trying them all? You're just out there building everything?

GB: This is what's happening. It's nowhere perfect right now, Josiah. But I'll say it like this. Right now we are entering this new phase. Think of AI like a meta engine coming up where I use, for example, ChatGPT. It knows the way I talk, so oftentimes it can write things exactly in the manner how I like it with the personas I like, and the people who I look up to and how they talk. It understands it very well.

Claude helps structure a lot of things, especially in presentations, especially in the way that we need to articulate it to owners and how those can come. It can do it a lot faster. I use Gemini really for a lot of the image creations. I use Meta AI when I want to create images to videos and add a story to it.

And then finally, as another layer, now you have got these agentic AIs that you can build yourself, your own agents. It's very simple. You just have to take an hour or two in a day for maybe the next two weeks and you can be really far ahead from T minus 14 to where you are now. A better version of yourself 14 days from now. You can have Gemini, ChatGPT, and Claude argue with one another, come up with it, and the output that is coming has been vetted with your different personas of what you like. The output is further refined because each challenges one another.

Josiah: I love that. Before we go, I would love to get your thoughts or advice for two specific audiences. One is the operators. I know we talked a lot about commercial strategy, but before off air we were talking about how you'll often come into environments. You're handling the commercial side, but then there's an operations side, guest services, the whole running of the hotel.

I wonder, we talked about data and some data that operations might be able to collect for you to operate in this sort of data enabled, AI powered, commercial excellence mode. What, if any, asks or needs do you have from the operations side of the business? And then I want to ask you about some advice that you would have for owners and investors to think about this holistically. But for operators, what would you like them to be thinking about?

GB: I love it. And I have a screen for you to share on that too because what we did was the same thing. So instead of me just saying it, let you also see it.

In this same thing, what we did was the next step. If you're an owner or a management company, what is something you should be doing today in this whole world of AI? We talked about data as the moat, and then we go through what a good data sandbox actually looks like for you and how it's structured.

Ask the right questions. It starts by asking the right questions and not just saying, what are we doing about it? Is your data being pulled in real time? Is it just monthly reports? Is your data also all of the contextual intelligence that we talked about?

From a team perspective, we really push on it a lot deeper. If you're an owner of an asset, where is the next great idea going to come from? It is not going to come from all these presentations. It's going to come from team members who are allowed to mess up and fail. That is very important. Are we allowing that culture to fail? That's the alignment that we start at a foundational level.

Does your team even have that permission to experiment? Because oftentimes what happens, we are in such a budget mindset. Oh yeah, we put $500 a month to do AI. But then what do you do with AI?

What lands up happening in a lot of these places is truly, I'll show you this slide. What happens in the world of an owner, month one, we bought the tools and it's a big launch and it's one day of training. It's awesome. On month three, okay, now the reports are generating, some of the leadership is reviewing, and then a lot of talk about where the data says and what it's saying. But it's turning into not more something like revenue management. It's more of revenue justification, or this is just data, but there's nothing quantifiable that you're doing with it.

Josiah: For audio listeners, this is a big problem you see. The set it and forget it. There's a lot of energy, we'll buy the technology, some reports start coming out, and then staff start going back to the old way.

GB: Because we are wired that way. Think of every system we buy today in hotels. It is set and forget. You buy a PMS system, you buy an RMS system. There is some little upgrades that come. But AI is changing every week. There is something so disruptive that even for this presentation, we joke about it, we did Delaware, I have to change this presentation three times in two months because stuff I put in there becomes so irrelevant. I can share with you. It's only a week old now. So at least that is still relevant to your listeners.

But what happens is in month nine, the vendor is collecting the checks, things go silent, and then the hype is just gone. Budget season comes, then the hotel tells the owner, oh, there is no ROI, it failed our expectations. And what does the owner say? Well, let's just cut it and move on.

This is the mindset we have today, very much a set and forget model, because that's how we are trained in hospitality. In the world of AI, it is very different. We are today what a modem used to be for the internet and how we used to connect to the internet. For those who know the modem from the 1990s, that's where we are in the AI space of what the internet is. There is so much to come. There is so much to evolve. Those who are going to be ahead of it are going to truly capitalize on what this has to offer.

The last part, you wanted to know from an owner perspective, what you should do. Back to the same things of what are the requirements when we take a commercial strategy that we have the ability to do. How do we position the website at a very basic, fundamental level that is AI ready? How do we talk to reviews? Because that is a place that most people don't understand.

How big is reviews for AI to read through on how you respond, whether it's generic, it's going to hurt you drastically going forward. Whether you have solved into a problem. Because if a hotel has a problem but you're answering it on how the hotel is solving it, it is going to give AI the credibility to say, we understand this hotel has issues, but we see from the reviews, from the management responses, they are addressing it in seven different ways because we have seen it across multiple platforms.

But everywhere, if we say, "oh, we are so sorry you had a problem, we take these things seriously and we have shared it with the management," you are going to be in the dust months from now and you won't even realize what hit you.

Josiah: I feel like we could talk for hours. What I want to do is point our listeners to where people can learn more about you and your work. I feel like I learn so much from you on LinkedIn. Is LinkedIn the best place? Would you point people elsewhere?

GB: Yeah, or our website from Mosaic Hospitality. It has a link to just schedule a demo, just talk to us. We are not in a place where we are like, oh my god, we want to win every account. We want to have good partners that are invested in the mid to long term. But if you're looking for a short term solution or consultants, we are not that company for you. But if you want an owner centric mindset that is powered by human intelligence on what is going to drive the future of hospitality, talk to us.

Josiah: I'll include some links in the show notes. Really good to see you, GB. Thanks for making time to chat with me.

GB: Thank you, Josiah. Take care. Cheers.