Latest / GAIN Momentum / Building an AI-Native CRM for Stronger Business Value | with Maxence Dussart

Transcript

Building an AI-Native CRM for Stronger Business Value | with Maxence Dussart · GAIN Momentum

• GAIN Momentum episode #114: Building an AI-Native CRM for Stronger Business Value | with Maxence Dussart === Adam Mogelonsky: Welcome to another exciting episode of the Gain Momentum podcast. I'm joined by our special guest today, Maxence Dussart. He is the co-founder of Familiar, a cutting-edge CRM. So Max, tell us the elevator pitch for Familiar. What is it? Maxence Dussart: Hi, Adam. I'm glad to be here today. So to put it in one sentence, I guess, Familiar is the AI coworker that's gonna bring a hotel guest back, and it's gonna prove the revenue it's making for you. So it connects to every system your hotel has, every system that might hold some guest data. It's gonna find who to win back. It's gonna draft the play. It's gonna run it, and eventually it's gonna show the money. Adam Mogelonsky: Yeah. I mean, you know, it's a good way to start in terms of, I guess, the goal of a CRM, getting guests to come back. And it's always the whole thing of saying, well, it's easier to keep an existing customer than it is to win a new one. And with the CRM right now, there's this whole idea of spray and pray, which is you literally just blast out and you're doing one journey, one campaign for your entire email list. So how is that spray and pray or that batch and blast model, how is it breaking down right now? Why is it breaking down? And then how does a AI-native CRM help with that? Maxence Dussart: Honestly, I think it was always broken, because batch and blast is basically you are treating 10,000, 20,000, different guests as one list, which is kind of the opposite of what we do in hospitality, right? We wanna really, tailor the experience to the customer. and it's also-- I think it's really breaking even more now because every platform out there is fighting for people's attention. and the only way to get people's attention is by being irrelevant. And you can see that, I think, in everyday life. e-commerce, platforms such as, I don't know, Amazon, Netflix, they have all understood that to get people's attention and to get them to buy their product, for example, they need to be relevant. and I think for hotels, AI kind of finally makes the per guest, decision, cheap enough to do it at scale. so I think, yeah, I think that's kind of the next big thing in hospitality. You had like some very small marketing teams, usually compared to the size of a hotel group, mostly doing communication. And the small time they had to do marketing was basically creating a newsletter, and as you said, sending and praying. And AI, for me, becomes that kind of like 10X coworker that is able to really look at the per guest level, do something specific for that person. Vincent Somsen: Yeah. How would you describe, you know, the difference between you and a traditional CRM that we, you know, we have a couple of them in the hospitality space, and also the difference know, a CDP, for example? You know, in the last couple of years we've seen a lot of CDPs becoming light CRMs, especially in the last 24 months. know, how do you see, you know, the difference between those guys and you? Maxence Dussart: Well, to talk about the CRMs first, I think, most of the legacy CRM, hotel CRM players, they were built at the time where it was more a matter of storing data, and then being able to send out emails, to those contacts. And they were not built for a system that is able to reason, and make decision and cross data points together. So I think that's a big difference for me, is like how do you architect your data model and your solution, to end up in an ideal world being autonomous or semi-autonomous, but to at least surface insights and so to compare a lot of data points and then, and then suggest some, like the next best action, that you should take. In terms of the CDPs, actually it's funny because we started wanting to build a CDP, because that was my biggest pain point when I was working in marketing for hotels, is I wanted my data centralized. That was like the first goal before anything, right? I think there are two challenges, if you're only focusing on data and not on the activation layer. First of all, you're talking to a very small subset of customers, that are technical enough to be looking for a data solution. Second, it's great to have all the data, but then comes the activation, right? And I think what's really interesting, especially with AI, is being able to suggest an activation that makes sense for a specific industry, so in our case, for hotels, and I think every vertical is gonna be different, right? so we started wanting to do a CDP, then we were like, "Oh, but we want to do the activation as well, otherwise we're not fully, masters of the entire workflow," right? And eventually we said, "Okay, but that actually makes sense because we're gonna tailor systems specifically for hotel groups, and that's gonna make the difference between a generalized CDP, a generalized marketing CRM, and so on and so forth." Adam Mogelonsky: So we have this activation layer above the data, which is the CRM. And now we have a AI-first platform that can suggest various actions to take for those activations. Could you walk us through an example of what those actions would look like, and how they add value above a traditional CRM? Maxence Dussart: yeah, I think the first layer is, closing the gap that exists between, let's say, usually what's in the domain of revenue teams. So what is my performance, for next month, next winter? What was my performance last year? Is my pace, on the right track? Because oftentimes marketing teams, they're-- some customers we're talking to, they don't really know what they should be doing except sending the 10,000 people email, right? So I think it's always great to connect marketing to how is it gonna solve my specific problem, that I have at this moment. So I think that's one of the first differences. Then I think it's an AI being able, to look record per record. So for example, I don't know, it looks at Mr. Jones, he's coming in five days. he hasn't booked any spa treatment, this time. He did his last two stays. the agent's gonna be able to first of all notice it, and then draft and send an email, with suggesting him to book a slot, and if the customer replies, then booking the slot for him. So all this action into looking deep into the data, then drafting the action, and so on and so forth, that would have taken a human a lot of time, multiplied by 10, 20, 30,000 guests, becomes now seamless through AI, I think. Vincent Somsen: Yeah. If they would even find the opportunity to do so, right? Because it takes the time to do those actions. But 9 out of 10 times, and, you know, having worked in hotels, I know that they won't have the time to actually look through, in-stay, reservations and say, "Hey, let's check if we can upsell these guys to a spa treatment or whatsoever." It just doesn't happen, right? Especially in this day and age Maxence Dussart: It never happens. It may happen for very small hotels, and quite luxury ones. But as soon as you start having scale, usually the marketing becomes a support team, right? And they just don't have time to go through individual data. They spend a lot of time on what I just named communication, so a lot of producing collateral for hotels. So they have very little time left to actually do marketing actions that generate financial performance. So when they do is, yeah, they do that newsletter because it's gonna take them a day to build and to send, and they only had a day in any case. So it does make one-to-one personalization, so the pinnacle of hospitality become feasible, I would say. Vincent Somsen: Yeah. And what, how would you describe your ideal customer? Right, you can, I saw that you can work with independent enterprises, but regardless, you know, what would be an ideal hotelier property for you? What, what would be the pain points that they experience currently? Maxence Dussart: Yeah I think the ideal one would be, a hotel group that has already started their digital transition, right? So they already have a few modern solutions in their stack. So they're using a modern PMS, a cloud obviously, but with hopefully a good API. I don't know, Mews, Cloudbeds, even Oracle Cloud is decent enough. So they already have this kind of cloud stack best in class solutions. They are aware that they could be doing better. Maybe they have a bit of pressure from the OTA's as well, as everyone does. So they're already at the point where they've made the first step, right? And then they're like, "Okay, but now what? How can we act on all these systems?" And they're looking for something. So for me it's a mix between readiness, in the sense that you've identified that you need to modernize and you've taken the steps first, and then urgency, meaning, I get pressure from the OTA's or AI is coming. I want to take out the best it has to offer. So there is a willingness to do something new in your approach to hospitality in general, I think. Adam Mogelonsky: Yeah. You know, one feature I love of Familiar is the visualization for journeys. I don't know the specific name for it, but it's just, it's a visual, shows you your branch trees for your journeys and your campaigns and your drips and all that. Could you talk about that feature and just how having that tool can really help hoteliers to see what campaigns they're running and see the performance on them, and then really optimize for that one-to-one marketing that we're all going towards? Maxence Dussart: Yeah. We call it workflows usually. Adam Mogelonsky: Yeah. It's a workflow, yeah. Maxence Dussart: A journey, a scenario, it can be called a thousand names. So yeah, I think it's a very interesting one, and we spent a lot of time and effort building it. First of all, we did it because usually a lot of hotel solutions, they box you into a predetermined scenario. For example, you want to do upsell, and you're just like, "Oh yeah, I'm gonna configure upsells." But as soon as you want to step out of it and configure it in a way that suits your property or your group best, it becomes impossible. So that, that was the first reaction. And then it becomes very interesting because it allows organizing agents visually as well. So it allows you to trust more what an agent is gonna do. It's not a black box. You're gonna be able to say, for example, "Oh, I'm gonna trigger when someone books a room. And then I'm gonna have a research agent that's gonna look at the guest profile, the past stays, everything. Then I'm gonna have a content agent who's gonna draft an email, and then I'm gonna have an outreach agent who's gonna send the email at the best time for this customer on the best channel for this customer." And you can see all of that visually, and you can build it, you can tweak it, you can look, you can investigate run by run as well. So for me, it's like a great interface between the user and the agentic workflows they want to build. Instead of just like typing in a chat, "I would like this to happen," and then, you know, you're not really sure it's happening the way you want it. Adam Mogelonsky: Cool that you're able to see it, but typing into a chat, you do have a conversational layer within Familiar. So tell us how that works. Maxence Dussart: Yeah. So it kind of does a number of things. First of all, I think everyone in Familiar hates dashboards. And every time a customer was asking for analytics, we were wondering how can we make analytics a bit more sexy, and how can we make sure that my main problem with dashboard is that people don't look at them, except the person that built the dashboard. And so our first use case was we want to let our users ask questions about their data in natural language, and then obviously produce visualization on the fly. And we want to give that knowledge to everyone, without having a fixed dashboard that is built by someone, it's never updated, and so on and so forth. So that's the first thing, like, how can I ask about my guest data, about the performance of my workflows, the performance of my campaigns? Second is obviously the agent can do everything that I could do in the platform for me. So the agent is able to build a workflow, is able to draft a campaign, is able to schedule a campaign. So he's able to do everything I would be doing as a human, except it's gonna be much faster, right? And finally, what I was talking about before is, the insights that we're gonna surface, the performance insights. So for example, your on the books is trailing same time last year. We're gonna make that connection between the agent and the performance insight where the agent is gonna say, "Okay, that's what I noticed. That's what I would do if I were you in terms of next best marketing action. So here's my plan. Do you approve it? Do you wanna tweak it?" You approve it, then the agent goes and does the work. And all you have to do is, well, make sure everything's done the way you had envisioned it and then press send or approve whatever and it's done. So it lets you do a lot of things in a very short amount of time and things that are relevant to your current situation as well. Adam Mogelonsky: So yeah, I guess that's a, to use a buzz term, that would be human in the loop to say, 'cause we're, you know, the agent's doing all the grunt work and then, but it still is the manager, whoever's in control of the CRM who's being held accountable for that workflow. Maxence Dussart: Yeah, I mean, there's different levels. You could ask the agent, "Draft me a campaign about this, this, and that," and that's like level one, you know? But you could also build a workflow where an agent's gonna look at next week arrivals, search every profile, draft an offer for each, post the upsell to the folio in the PMS, and that's like level four. So you have all these in-betweens where you can get more comfortable using AI, up until point where it's kind of semi-autonomous, and eventually, I think in a few years, maybe months, fully autonomous. Vincent Somsen: Okay. So that's actually, that was going to be my question. What do you think that's going to look like? And what do you say to marketeers in the hospitality industry, marketing directors, for example, or, you know, DOSs, that are going to, that are using CRMs right now? Do you think they're potentially even afraid of implementing your tool in, you know, afraid to lose their job in a few years or even a few months, like you said? Maxence Dussart: Really. I mean, that's probably not what's gonna happen. First of all, because 90% of the time that's a job that's not fulfilled at the moment, so it's not as if we were replacing someone. We're bringing in a new coworker who's gonna be able to fill in that position. Second, I think it allows humans to, on the marketing teams, on the revenue teams, to spend more time on strategy, on brand, on guest experience, on key moments, and less time doing the grunt work, right? I don't think there is any value in spending half a day drafting a newsletter. You already know what you want to say in your head. There's no value spending hours doing it and then making sure you have the right segment, right? That you have the right translation and everything. So I think it's more about transforming the existing team into a 10X team, than replacing anyone on those teams. It just gives you a scale. It gives you the scale of a very large group even if you're a very small one. Vincent Somsen: Yeah. I could even see with this becoming more and more autonomous, GMs that are currently, you know, they're in boutique hotels sometimes, you know, they don't have a separate director of sales or a director of marketing or a director of revenue management. Sometimes these are combined, right? And usually a GM, you know, picks up the role of director of marketing or even the owner helps a little bit. So this is where I could see this being super valuable as well to those small scale teams, that don't have the budget of an extended team. Maxence Dussart: Yeah, it's interesting because no matter the size of your team, there's just gonna be a use case, right? If you're a single hotelier, usually you would either outsource your acquisition marketing guest engagement to an agency, or the GM would try to make some time to do a few things on the side. But either way, you don't have a lot of control, you don't have a lot of time to make it great. So I think that's where AI is very interesting. You can describe your vision, the AI then has the tools, to make that vision happen. And you can tweak it, obviously, you can improve it, but, that's the key difference for me, yeah. Vincent Somsen: Yeah. And in the future, do you think you'll stay focused on, you know, being a CRM, or do you think there's something more in five years that, you know, it could expand to different functions within the hotel? Maxence Dussart: Yeah. We already don't call ourselves a CRM, honestly, anymore. Because for me, CRM is a category of tools that are legacy tools, what I call usually newsletter tools. So you're already in a bucket of a tool that's not generating that much performance, right? So it's more of an overhead cost that hotels are reluctant on, than the true revenue driver. So for me, it's more interesting to build the AI coworker, starting with marketing and revenue. And then it can become very useful for operations, for example, where we're working on a Chrome extension that sits directly inside any kind of PMS, and it's gonna give front desk teams that usually don't-- are not willing to exit the PMS all the context they need about their guest. So you would have an AI overview of your guest telling you, "Oh, Mr. Jones has," to take the same example as earlier, "He has not booked a spa. He's a big customer of the spa usually. You should suggest him at check-in." And all of this lives directly into your PMS. So if you're on the front desk, you navigate to a reservation, we pull the reservation from Familiar, we bring in all the context from all the other tools in your stack. If someone has left a survey sent by Familiar last time, and they were mostly happy, but they had one or two remarks, then the receptionist is gonna be aware of them at check-in, for example. So it can go in a lot of different ways, so that's why we prefer to talk about really an AI coworker that's gonna work your entire guest base, and down the line make you more revenue and improve satisfaction as well at the same time. Adam Mogelonsky: Yeah. You know, this whole idea of a AI native CRM with a CDP layer underneath it, and you talk about implementing a Chrome browser as the latest thing. AI agents in there, that's on the, I guess, the user side for the hotel, but being AI native also allows you to implement new features much faster than other systems of coding or object architecture. Could you talk about just what that means in terms of being a next generation platform and how you're able to roll out new features and new integrations very quickly? Maxence Dussart: Yes, I think in terms of on the business side, it has made, and even on the technical side, it has made everything a lot quicker, right? A feature that would take months before now takes weeks, if not days. So it's more a matter of knowing what should I build instead of can I build it? And I think that that's kind of a trap, you know? You can build anything right now. Should you build it is the question. And we have this framework internally where if we consider Familiar as an agent, we wanna be adding skills to that agent that help the hotelier make more revenue. So that's kind of a filter for new features. Is it gonna be a new skill for an agent? If not, it's not gonna be built. Then in hospitality, one of the hardest things, the most time-consuming things is integrations. It's one of the things that slows your sales as well a lot because everyone has a different stack. There are hundreds of solutions out there. Some of them are open source, great, but these are just a minority. The rest of them are a bit more complicated. So, we built an internal integration agent that has learned from everything we've done before, our data model, everything, and is able to build an integration in a very, very fast way. So the only barrier that remains now is getting the docs more or less. And so I think that makes a big difference when talking to a prospect. We would usually ask, "Can you tell us your five or six top systems, PMS, booking engine, spa, restaurant, housekeeping, whatever, POS?" And you end up always with a different list, right? And so now we're able to tell them, "Oh, it's gonna be ready by the time you get on the platform," which makes a lot of difference. Adam Mogelonsky: It will makes the world of difference. Yeah, 'cause that's make or break, and the future of CRM is not just that import from the PMS, but the ancillaries, as you said, that the key drivers why people go to a lot of different properties is hidden in the details of, they were treated well by a server, so it's recorded in the restaurant system or, you know, they had a great spa treatment or something else, so it's recorded in the spa system. Maxence Dussart: Or someone is browsing the website, they have cookies, they're looking at, I don't know, stay dates. They don't go through with the booking, you want to know it. And every time you need to connect to a new system. So, I usually say PMS maybe has 30% of the guest view. The rest is scattered everywhere. Adam Mogelonsky: Yeah. Vincent Somsen: Interesting. And particularly in the, I know you don't want to be called particularly just a CRM because it's more, but for simplicity's sake, you know, we have some CRM systems in the industry that are not specifically built for the hospitality industry. They're generic CRMs. And usually these pieces of softwares, they're quite huge, right? And you need three different consultants to implement them. It's gonna take half a year for everybody to get trained on this. I see your platform as a much, much leaner platform. So should hoteliers be aware of some limitations that you guys might have that, you know, some of the other guys in the industry do offer? Is there anything that they should know before, you know, contacting you and really understanding, okay, what is it that you guys offer? Maxence Dussart: Yeah. I think with generalist CRMs, all the Salesforces of this world, right? You can get a lot of customization, but it comes at a very high cost in terms of configuration, consulting hours, and then in terms of training for your teams. We know we're in an industry where turnover is pretty high, right? At least in Europe. And so getting someone trained on a very complex system, only for that person to leave, I don't know, two months after, it's a hidden cost of that kind of implementation. I've seen groups spending close to a million dollars implementing such solutions only to down the line make a different decision a few years after that. And it's also kind of a trap, I guess, because in the end, you're so invested in the solution that there is an opportunity cost that prevents you from switching to a new and better solution more suited to your needs because, "Oh, I've invested so much time and effort in this, I'm not gonna let it go away." It was the case with PMSs even before, like, "Oh, I've invested so much time and effort into Opera, that I'm not gonna go with Cloudbeds or Mews, even though they look more suited to my needs." Then in terms of why us versus any other, I think hospitality is very specific. You have a very specific stack. You have a very specific way of doing things. You have guests that you engage online during the stay, after the stay, they come back. They can go to your hotel, they can go to your restaurant, to your spa, to your shop, to your event space, if you have some. And I think there are some very specific workflows inherent to that industry. Especially when you consider AI. If we forget about CRM entirely and you're like, okay, what is AI gonna help me do? I think having the knowledge of how the industry works for an AI is gonna make a lot of difference as well. So any system that is specifically built for hospitality is gonna outperform a generalist one, in my opinion. Since in the end everyone is using the same models, right? It's not like we're building LLMs in our backyard. So, everyone is using the same model, so in the end the difference is gonna be your data. And so that's our customers' data, what they make with it. And then it's gonna be about, is the system built for the way I work for my industry? And the thing we wanted with Familiar is, we want people to come because they're excited about the product, and we want them to stay because it's delivering value, right? So there's no point in spending months doing onboarding and charging enormous fees and so on, if then we're gonna have customers that are not gonna be happy and not gonna see value in our product. So, we're very product focused in that sense, in the sense that it's our job to retain the customer. It's the product's job to retain the customer. Adam Mogelonsky: Yeah. Vincent Somsen: Yeah. Adam Mogelonsky: Yeah. You know, one difficulty of the hospitality industry is we have cloud software and now we have AI to allow hoteliers to do so much more than what they used to do. But the sales process is still largely local, in terms of meeting teams on site and offering 24-hour support of some kind. So how do you as an AI first cloud native CRM deal with the local aspect of selling to hotels? Maxence Dussart: I think, it's gonna become quite clear for us, as soon as we get into US and into the US market, for example, which is quite far from our home base, which is European. We didn't face this issue so much in Europe, mainly because, we have made the effort to have some great documentation, to have an AI agent that's able to handle a lot of customer support requests. I think that having an AI as well inside of the product that does things for you, eliminates a lot of the support bandwidth that you would need before. But then, yes, in terms of the sales, they want to know that they're talking to someone that is in their region, understand their specific needs, and so on and so forth. So, I'm pretty sure that, for example, selling to the US would mean, at some point, some presence in the US. And so that is something we are considering as well. Adam Mogelonsky: Yeah. You know, the US and then all the other markets. Everything is local in hospitality sales. It's just part of our industry. But yeah, I guess having the AI agent can definitely help. What percent of support questions do you think an AI agent can answer and fully resolve before it has to be elevated to a human? Maxence Dussart: I can answer statistics for our direct customers. I think, at the moment it's between 80 and 85% of product usage questions that can be handled by AI. It can go up to, I think-- I read an Intercom paper the other day, talking about how much they managed to automate for a client, so it can go up in the 90s, I think quite easily. There's a lot of companies specializing in this at the moment, and the results are pretty impressive. Then I guess for the end customer, so the guest of the hotel, it's interesting as well, how much of the one-to-one engagement can be automated, right? So the agent can do the first outreach, but what if a conversation ensues? Can the agent then handle it with the same touch a human would? I think up until now it wasn't that great, but I always like to not look at the past to predict the future, right? Because the growth curve of AI is very steep. And so I think that within a few months, honestly, AI agents will be able to handle all the non-face-to-face engagement. Adam Mogelonsky: Yeah. And I guess another aspect of all this, these agents that are embedded in a CRM would be the MCP of the matter. So tell us what is Familiar doing with developing and exposing an MCP to allow hoteliers direct access into other AI systems? Maxence Dussart: Yeah, so we're talking to more and more prospects that are very into this AI transition, and they're on the train. They want all their data points crossed and connected. So the first ask was, "Oh, I'm already in ChatGPT, I'm already in Claude. I want to be able to connect my systems and I only do my work in Claude nowadays. Can you guys get in there?" And so that's why we're building the MCP as fast as we can, because I think that it's not like we want to force people being in our dashboard, right? There's no value in the end in forcing them to be in our interface. I think they should be when it makes sense. Otherwise, I'd rather have them in their preferred tool, right? It's like asking the operation to get out of the PMS. For me, it's something that's never gonna work out. So you should be in the next tool where work is gonna happen. And so someone, a general manager, a CEO, a finance manager, whatever, I think tomorrow will connect mostly, most of their stack to their favorite agent, LLM, and be able to ask questions about anything, make connections between a lot of different data points. I'm pretty sure there's some pretty amazing things to discover between maintenance data with guest feedback, with I don't know, a discount. I'm pretty sure there's a lot of things to uncover once you connect everything, right? It's humanly impossible to do, but I think, like, a great GM, a great manager, will probably adopt this workflow in the future. And so we should be in there. And I think the solutions that are not readily available in those tools would be making a mistake, honestly. Vincent Somsen: Yeah. I do have one question on the data input. Adam and I, we sat during HITEC, we sat in a roundtable with a couple of CROs of some of the larger hotel firms, and one of them, they said, "Well, you know, CRMs, CDPs, and data, and AI, it's all great, but the real problem for me and my hotel group is who puts in the data and the quality of that particular data," right? So what would you say to a hotelier that actually is concerned about the who's putting in the data and how that's going to happen? Because it's easy to see, okay, the CRM will take everything from email, from surveys, from all online data points, right? But where the real treasures are and the golden nuggets are on the floor are when the front desk agent is actually talking to a guest, when the bartender is shaking the cocktail with a particular recipe or non-alcohol or whatsoever, when the waitress at the pool is asking about the dogs and whatever, right? So how would you... What would you answer to those hoteliers that are more concerned about the live data that is transmitted? Maxence Dussart: Yeah. I mean, with data, as you said, it's always the same issue when-- and usually in data science you say garbage in, garbage out, right? So I think there's two aspects to this. First of all, we need to make sure that wherever there is actually data today, all these systems, where we have that kind of middleware where we clean everything, we enrich the data when we can. I think that's an important aspect as well. And then we make it available to agents. So there's a lot of-- There needs to be the correct structure so the agent can access the data in a timely manner, create connections between it, and so on. For the on-site, for the human feedback, let's say, I think it's gonna be interesting to see what kind of technology arise, where the entire floor of a hotel, of a property, can participate in in improving our knowledge of the guests, right? Honestly, I don't have the answer right now. Is it wearables? Is it processes where everyone should once a day fill in some data in some kind of form? Is it with voice? I don't really have the answer. I think we're gonna see something happening and I think at this point it will make a lot of sense for a platform like us to use this data, because now with AI you can also extract a lot of value from unstructured data, right? So you could have someone just dictating something about a guest to his phone. We can use that. We can structure it. We can extract insights from it. We can generate tags from it. I think it's gonna come. I don't think we'll be the one offering the interface to capture this data. Maybe I'm wrong. But we'll definitely be the place where then the magic happens based. Vincent Somsen: Yeah, yeah, that makes sense. And I think even the ability to identify certain suggestions with MCP at hand as well, right? If you can look at the maintenance or HotSauce or maintenance tool or any other tool that something is being recorded in, and that your system is able to identify a certain suggestion, a certain personalization, certain touchpoint, I think there, that's where you could potentially find a middle ground. But yeah, I guess we'll find out. Adam Mogelonsky: Yeah. So, to sort of tie into a conclusion here, I actually want to touch on a personal CRM issue that I had within the past 72 hours. And I want to focus this around how, an AI-first system like Familiar would resolve this, or at least, you know, identify the pattern. So I logged into a website, this was a consumer packaged good, and a pop-up came up saying, "Hey, fill out your email, we'll give you 10% off your first order." Okay, sure. You know, simple mouse trap. But then from there, I gave them my email, and then it said, "Oh, we also need your phone number." Sure, whatever. Fill that out, and then literally within the first 15 minutes, I got four emails from them saying, "Here's your promotional code. You have to get this now. Oh, here's another offer. Here's another promotional code." And then 24 hours later, I was getting two emails every hour from them saying, "Last chance, you have to use your promo code now." And then I was getting texts from them as well. So, you know, that to me is... That's not just spray and pray. That's a Gatling gun of a— Vincent Somsen: Scam. Scam and pray. Adam Mogelonsky: This-- but this wasn't a scam. This is, like, a well-renowned company. I'll tell anyone who DMs me what it is so they can avoid this. But, like, unsubscribe, unsubscribe, stop text, all that. And, you know, I'm a Virgo, so I unsubscribe from everything 'cause I want to have digital, be a digital minimalist. But, you know, I imagine their unsubscribe rate on these campaigns they're running must be astronomically higher than another campaign. So Max, my situation, how can we detect that within a campaign to then have an AI agent say, "Hey, you're over-emailing someone," or, "You're over-emailing this segment. Here's how, here's what I suggest you do to adjust based on data"? Maxence Dussart: Yeah, I think the way we're gonna approach it with Familiar is not by reinventing anything. We already use obviously LLMs for internal work, especially engineering. And we have built for engineering hundreds of skills, internal skills, where we have some very specific guidelines on what should happen when, which systems should be consulted, how code should be written eventually, and what tests and so on. I think the same can apply to any consumer-facing agent eventually. Every property, every hotel group is gonna be different in the way they do these things. You might have groups that are okay sending these 25 emails just to create a urgency. You might have some that would never send a second one. And so we cannot forecast every wish and every behavior of every prospective customer. So the way we're gonna approach this at Familiar is we're gonna allow customers to upload skills, right? And to draft skills. So you could have a tone of voice skill that first of all tells the agent how should I express myself. Then you would have an upselling strategy, or playbook skill that will kind of provide guidelines on how any offer should be presented, because you're not gonna say 24 hours left for your 20% at the restaurant if you're a five-star luxury property, right? You're not gonna have that language. So that's the way we approach it. And second, I was talking about workflows earlier. We're gonna introduce approved workflows. So it's like a training phase where before running an automation unsupervised, you're gonna be able to approve runs manually, and you're gonna be able to correct what goes out it's not to your liking. So that's the way we hope we're gonna solve the issue. And I think agents are pretty good at following instructions in general, as long as these instructions are well baked into the entire agentic workflow, so. Adam Mogelonsky: Yeah. You know, yeah, there's lots of ways to tackle that problem. I do feel sorry for the hotels that feel that they can Gatling gun their guests and think that that works as a long-term strategy. I mean, hey, maybe it does, but six emails inside of 24 hours is a little bit too much in my regard. Maxence Dussart: Yeah, I agree. Vincent Somsen: Is there any, for hoteliers that are just start, or let's say they're, they have a new opening, for example, and they don't have much experience with marketing or CRM, is there a library or best practices of workflows or agent flows that they could potentially rely on? Maxence Dussart: Yes, we do have templates inside the solution where you can pick, for example, the usual workflows, the basic ones, so pre-arrival, selling, room booking confirmation, option reminder, all those kind of usual workflows. I think the next interesting we're gonna do is that the entire onboarding is gonna be agent-led. So you're gonna have a discussion with an agent on how you would like to do things, and based on this kind of informal conversation, the workspace would be set up, exactly as you pictured it. So maybe sometimes you don't have the words, but we're gonna do our best to make this workflow, this platform tailored to your needs specifically. So I think that that's gonna help with first-time users as well. Adam Mogelonsky: Yeah. Vincent Somsen: Okay, makes sense. That's good to know. Adam Mogelonsky: Yeah. And Max, to close out here, is there anything, are there any other features or anything else that gets you excited about the future of CRM? Maxence Dussart: Yeah. I think the most interesting thing in any kind of business at the moment, but hospitality is very interesting for that is, the data you have, the first party or zero party data you have is gonna become more and more valuable. It's the only thing kind of out of reach of all these LLMs, and it's what's gonna make the difference between you and your competitors. So I think the database eventually is the most underused asset in a hotel. And I think hospitality, it's always been about knowing your guest, remembering your guest, serving your guest. I think AI will finally let you do it at scale, no matter the scale do you have. And it's very interesting for me, as I was saying earlier, to see this data finally being put to use, because since I started in hotels, a few years back, I've always heard, "Oh, we have a lot of data. We have great data. We know a lot of things about our guests." But I've never seen it used to effect basically. So I'm excited to see how finally our industry is gonna be leveraging this data, thanks to these new systems. Adam Mogelonsky: Yeah, the data is there collecting digital dust. Maxence Dussart: Yeah, exactly. It is sitting in different databases. No one's really using it, or when it's used, it's for dashboards or reporting that only a few people do actually read. So, I think there's gonna be pressure for hotels to accelerate on making their existing, their past guests come back. OTAs have been the subject for the past, I don't remember, 20 years. I think there's a window where it could very well end up being LLMs that are the next booking.coms and everything, taking a lot of your margin. I think there's a window right now where it's very interesting to see in which direction it's gonna go. Are the hotels gonna be able to adapt fast enough so that they are part of this new way of doing things? Or again, is it gonna be in the hands of another platform? So it's a very interesting time for, a very exciting time for me, and I think the hoteliers that put their data to good use will see great returns compared to their competitors. Adam Mogelonsky: Yeah. No, that's a great way to finish off, looking ahead to the future and there's lots to look forward to. Max, it's been a fantastic conversation, really, really insightful to learn about the future of CRM. Thank you. Maxence Dussart: Thank you for having me. Thank you very much.