RLC Talks - Revolutionizing Restaurant Operations with Conversational AI

In this episode, Jeremy interviews Joe, the head of partnership for Revmo AI, a conversational AI platform designed to improve customer interactions for restaurants. Joe discusses how their AI agents provide basic information, manage reservations, waitlists, and even take phone orders. They delve into the rapid advancements in voice AI technology, the importance of training the AI model for accurate responses, and the benefits of automating customer service to enhance human interactions in restaurants. Joe also highlights Revmo AI's capacity to handle up to 100 calls simultaneously and continuously improve through learning. The discussion emphasizes the potential of conversational AI to alleviate common operational pains in the hospitality industry.

00:00 Introduction and Guest Welcome

00:07 Joe's Role at Revmo AI

00:31 The Evolution of Voice AI in Restaurants

01:07 Challenges and Solutions in Voice Ordering

03:26 Training the AI Model

06:17 Future of AI in Customer Interaction

11:11 Conclusion and Contact Information

Transcript

0:00

Well, Joe, thanks for coming on the show. Yeah, thank you so much. Why don't you introduce some, yourself to our listeners, who is Joe? Yeah. And then we'll talk a little bit about what Joe gets to do. Yes. Sweet. I appreciate it. Thanks Jeremy. so I'm head of partnership for Rev Mo ai. we are a conversational AI platform, so, we basically help interact with restaurants, customers on the front end. So if someone's calling you, our agent will. Interact with them, provide them basic information about the restaurant hours. and then we have the ability to then transition the conversations into a reservation or putting'em on a wait list or potentially then putting'em into maybe even a phone order. love that. And I, I know it's funny'cause it feels like the voice AI space has exploded. I know we were talking about that prior to hitting the record button in that, I think it's interesting that you talk about kind of the, the, the one side of it, which is helping guests get where they, where they need to get to. For me, part of my challenge with it is everybody goes straight to ordering. Yes. And I think ordering is complex and I, and it, it has a huge opportunity. I happen to be a family of six, so a lot of times I'm calling restaurants to see what the way it is. Yes. I'm calling restaurants to see if they've got a table available. I'm calling restaurants to see can I bring my dog to the patio? Yes. Like I'm looking for basic information. Are they open on the holidays? Mm-hmm. Those kind of things. Talk to me about why you guys chose to tackle that in addition to Yeah. The voice ordering.'cause I think voice ordering. Is awesome. Yeah. But some guests just don't want to, but they need to be able to solve that other problem. Yeah, that's, that's really good. So it's funny, like kind of the journey of the, of the business is, we kind of got into where we're like, Hey, you know what? We're just gonna kind of be the first line of defense for the restaurant. We're gonna pick up the phone for them every single call nonstop, twenty four seven, and we're gonna provide all the information they want. And what naturally happened was, is that was great. but then they started thinking like, well, my host does this too. And my front of the front of office or front of restaurant, front of operations. Does this, you know, they, they actually, you know, book the reservations as well, and they're, you know, managing this, managing that within the systems, right? And they're putting this through the POS system. And so as, as an init and as it started, it was great to pick up the phone and provide these customers the information that you're looking for. And then it was like. Wait, can you put'em on a wait list for us? Yes. Like can we automate that? Oh my gosh. Yeah. Great. Right. So many other places I go eat at need your stuff. So I'll, I'll let you finish. And, and so, and then the next stop is then wait. If we can go from wait list, can, can they just book a reservation over the phone? And then from there, then of course the common online ordering conversation starts, you know, because it's like my host does all of this for me. So like, if you're really like. Going to try to alleviate my pain points as a restaurant owner. I really need you to kind of take on this full responsibility. Well, and, and unfortunately, the, the host stand oftentimes is the busiest place. So people are calling Yes. The restaurant. I heard a statistic recently that, that this, it was a phone provider that. Said 70% of phone calls don't get, or they get answered. Somebody picks it up and then hangs off. So I guess, is that really kind of what you're trying to solve for that 70% of the people that didn't get picked up? Yeah.'cause once I get a human on the phone, typically I can get what I'm looking for. Yep. I, you know, but unfortunately, if 70% of the time I call it just ring no answers. Mm-hmm. Or it rings and I get put on hold and I'm not getting what I want. Yeah, ultimately I don't go there. Is that kind of where you guys are started? Yeah, and I would say like that's one of our strong points of our tech is we can handle up to a hundred calls at one single time. Right? So, and with that 24 7 type of approach, you know, it's being able to provide a consistent information, a consistent interaction to a customer, a friendly interaction. and someone that's willing to show up every single day and do it every single day. Right. Well, that's the thing with AI and, and computers Right. It does not call out sick. It's there unless it, unless your system goes down, it's there. Exactly. And it's gonna gonna be what you need it to do. Yep. one of the big things that you keep hearing or I keep hearing Yeah. Living in the restaurant tech. Technology spaces that you have to train the model. Yes. So help me understand how you train the model. Again, think about that casual dining restaurant that, you know, I'm looking for a patio sale. I'm looking to see if this item might still be on the menu, looking to see if I can bring my dog. How do you guys build that in? Is it, is it. Somewhat scraped from their information. Is it, you know,'cause again, a lot of people that aren't in the tech space don't realize that the model has to be trained in order to tell it how to do that. And so I, I find that that's a, that's a challenge for people. They don't train it properly and then they don't get the results that they want. Correct. Yeah. And, and that's one thing I would say Rev mo really we pride ourselves on is like, we try to promote ourself more as like that conversational platform, right? So we're not just looking to, once again, answer a phone, take an online order book, just a reservation. Just tell, just put someone on a wait list. We really want to elaborate and provide additional information, right? Yes, you can bring your dog on the patio between this time of the day, right? we also provide water for your dogs, whatever it may be, right? In addition, yes, you know, this menu item is available, you know, or, you know, if you're asking about allergies, right? You know, having that complexity in that diversity in the, in the agent and how it, how, how you're providing that information. One is if we can integrate into the, into the POS or into the customer data platform, or customer data system. And then the next thing would be, you know, as we're building the agent, we're we're going through this onboarding process is what we call it, and we're intaking all this information from you. And then we're feeding the agent with this information. Right. And it can be, you know, ideally it's a web hosted or cloud hosted database, right. So that we can, we're not having to continuously update it. So like, because we can do something as simple as a spreadsheet, right? Yeah. You can put a spreadsheet in the agent. Let it do. But the thing is, you gotta update the spreadsheet, right? And you gotta keep, you gotta keep learning. Yes. As you continue to change. Yes. As hours change as how you, you know, services change. You gotta teach those things because the human gets that pre-shift. Yes. They get that in, they're noticing their scheduling platform or whatever else. You've gotta get the computers in that same place. Exactly. And so, yeah, so the. Key is, is gathering that information, that Intel and making sure that, you know, identifying is it gonna be automated or is it going to be manual? Right? And being able to understand that and being able to maybe then develop even, you know, if it is manual, like maybe we can help kind of push you in an automation for, way, right? Or, hey, you actually host this on your website and you update your website as soon as you update your hours on Google. Right? I do, well, let's just use your website as a data source. Right? Instead of you giving us, you know, a, a spreadsheet or something, or going into some portal and updating that. Exactly. And then somebody inevitably forgets and now they're stuck. Well, and, and that's awful. I, I, I've driven to a restaurant and had it been closed, even though the website said it was open, or I called and, you know, the IVR said they were open and then you show up and they're closed. Right. Really, really frustrating because somebody's job was to do that. They forgot. Yep. So now with the, with. With things like your solution that can be scraping and correct, getting that data, no longer does it become a human task that they have to do. Exactly. And you know, and then, and then the flow. Right? And then, and really what I want to take the time also to highlight is while you're implementing this inside your business, you're, you're really enabling the human interaction be stronger. Right. Because when they are engaging with those people Yes. They can, they can deliver the value and be hospitable. They're not so stressed out because they're 12 calls behind trying to deal with it. Why not? Why not? Even I, I've started, this is just one of my little ideas for a restaurant, but think about your hostess going up to your, your, your wait list customer that comes in and is expecting to wait 30 minutes and they walk up in, offer them the water while they're sitting down. Better at even a beverage. Right. Or potentially a menu to then look through it. Right. And even maybe prepare the order so when they sit down Yeah. Preorder, preorder, you're looking for, I mean, talk about that human interaction in that and, and I, and I go back to this too, as a restaurant, like when you think about the, the journey of a customer, right? Yes. Advertising's important, email, marketing's important. All of these sources of marketing and promotion is key. But the most valuable is a walk-in. The second most valuable is a phone call. So you have to answer. Yeah. And we all know how hard it is and we know how impossible it is.'cause I can only answer one call. You can only answer one call. So that host can only do that too well. And so for those naysayers that go, I call American Express and it's an awful experience. You get it? I'm certain you get it from people all the time. Yeah. to stop the myth that it has to sound like a computer, because I think that the models have gotten so good that there's times that if you don't know, you're talking to a, to a computer generated dumb voice. Mm-hmm. It could be a computer generated voice that you would have no idea that it was correct. It was that. And the interaction's gonna be better. It's gonna be more consistent. Yes. It's gonna deliver on the brand promise that you had. Yes. You're gonna have to train it one time and let it do a hundred calls at a time versus having to train a hundred different staff members. So, I guess to spell that myth that it sounds like the American Express line or, you know, American Airlines line that you call and you, you sit in the queue forever and you're agent, agent, person, person. Now human now human now. And then the human gets on the phone and I joke around and I go. We go from the human, then we ask for the supervisor. Right? And then directly, a hundred percent. So, yeah, let me, so I'll, I'll be honest with you. So we have the capability of actually using, this is gonna maybe sound a little crazy, but you could record as 32nd, like voice recording, and then you can take that voice recording, upload it into our agent, and you can use your own voice. Yep. Which is pretty cool, right? So, and I've, and I've seen the tendency in, in our, in our pilots and in our agents, where that really helps eliminate that robotic feel and that robotic. You know, experience, I would say. and then sometimes I also also have found with customers and my partners and stuff is like sometimes opening the conversation like, Hey, I'm an AI agent. I'm here to answer and talk to you like a normal person. So don't think of me as an IVR, like, joking around, even maybe, you know, and it's like, okay, I'm looking to see what, when I can come in to get the, you know what I mean? Like having a true conversation with a.'cause sometimes people think that they talk to it like an IVR and it's not that, right? It's not that at all. It has a, it has a brain, you know, ostensibly where it can, it can reason and give you back reasonable answers. That's the other part that I think is so remarkable and. I guess the last piece, Joe, I'd love for you to share is, is how much better it continues to get as you guys continue to feed. I mean, because that's the part that always blows me away. It's like where it is today versus where it's gonna be six months from now versus where it's gonna be a year from now is vastly different. Correct. And it's crazy. Like I'll, I'll, we have the ability to create like a demo agent. So like, you know, for a restaurant in like 30 seconds from like their Yelp listing. Basically we have this, we have a partnership with Yelp, so we have the ability to do this. So then the next thing, so then it's like, wow, that's not bad. And I'm like, that's literally a 32nd generation, you know? And then I go, imagine what it's gonna be in 24 hours and 48 hours a week from now. And then I said as, as we continue to fill this agent and it has more conversations, it's learning and growing. The LLM is learning and growing, right? and we put guardrails on it, is what we say. So it doesn't. Trail off and get lost in. He starts sending, he starts sending you, you know, ordering from Domino's. It's like, let me get you an order for pizza, hu. You're like, whoa. We don't want that to happen. Right. Yeah. And so, you know, as that's, as it's being fed, that it's becoming more and more intelligent and becoming a better employee. Right. Just like, just like we know as the longer that hostess is there, she's gonna become a better host. And the agent functions the same way, which is, once again, it can be a little like alarming or maybe even a little concerning, but at the end of the day, the technology has that capability. So like, take advantage of it is what we say, right? And continue to provide it more and more information and the more conversations it has. And the more interactions it has, it's gonna become a better, better operating platform for you. Yeah. We, I use a AI tool to do all of the editing for the podcast, and it's amazing. I, yes. I said somebody's name wrong the other day, and I, and I rerecorded it and typed in the words, and it rerecorded my voice. Voice. Really? Yeah. I re, I mean, it was amazing. I was blown away. I had a whole paragraph that I had to retype. And it sounded just like, like if you, if you weren't paying attention, yeah. It could have been me and it was an ai ai and it's, you know, it's cheap too. And that's the crazy part. It's not, it doesn't have to be that awful expensive. Yeah. It doesn't need as much compute, so. For those that wanna learn more, get in touch with you guys, kind of figure out how they can put tools like yours. Yeah. You know, rev mo into, into their stores. Where should they go? How do they get connected? Yeah. Yeah. You could literally visit our website. It's rev mo.ai. Um, you can go there and kind of even understand our pricing structure. Uh, we have our partnership information, which is what of course I focus on. Um, which is a big initiative for us as a company. Um, and yeah, so the website's the easiest and yeah, so the website's the easiest way. Awesome. Well, Joe, thank you very much for so much continuing to make the world a better place. Yeah. Appreciate you. Thanks Jeremy.