How to Grow Your Restaurant Business with Kelsey Verdier and Avi Goren

Speaker A: All right, here we go. Now that we're live, welcome to the Restaurants Grow podcast where you get tips, tricks and tactics on how to grow your restaurant business. Today in the show, we welcome Kelsey Verdier, VP of marketing at Marquee, and Avi Goreng, co founder and CEO of Marquee. Avi's been helping restaurants own their local search performance and now their AI search performance for more than a decade. Uh, with a hospitality career that began delivering pizzas, Avi continues to immerse himself in solving the daily challenges restaurant operators face an increasingly tech forward market. You could say that twice and still have not actually covered really what it represents. Uh, Kelsey has worked in bars, restaurants and breweries before moving over to tech and eventually at Marquis specifically, where she's dedicated herself to helping restaurants find the balance between tech enabled efficiency and high touch hospitality. I am very excited to have these two on the show. Avi and I have worked together for, I think more than a decade at this point. Uh, Kelsey is one of my absolute favorite marketers in the business. They always have great insights to share and are relatively just good, fun people to be around. Uh, some fun facts. Avi got his first hole in one ever at the Restaurant Leadership Conference. I'm not sure I was out of bed early enough to catch that one. Uh, and he also used to have pet chickens. Uh, Kelsey lives in Amsterdam and has almost learned to pronounce most Dutch sounding beach or sorry, the most Dutch sounding beach in the Netherlands. We're not going to test you on that. Kelsey and Avi, welcome to the show.

Speaker B: Thanks, Rev.

Speaker C: Thanks, Rev.

Speaker A: And I, I specifically wore my, um, Beer city bung hammers hat for you today. So I thought that would be fun with the bucky shirt. Okay, I'm just. We're going to go right into this. Are people actually using AI like Chat, GPT and Gemini to figure out where to get dinner tonight? Like, is that a thing?

Speaker C: So I'm going to jump in first because usually it's like, okay, uh, we're in the industry, we're in tech. Like, of course we're going to use it this way. And then I had one of, I'm not going to call them one of my biggest Luddite friends, but definitely not the most like tech for they don't have a Siri Alexa in the house. A Google home. And we were planning a trip together for two months from now. And she sent me a full itinerary by Claude. And that was one of those moments where I'm like, okay, this is real this isn't just in our bubble. Um, even my mom was like, hey, we're gonna drive down to Texas in a month. I can't log into chat GPT. What should I do? I want to plan my trip. I said, well, go to Claude. Step one, go to Claude. Um, but yeah, it's real.

Speaker B: And oh, I would say like, this is something I'm stealing from Avi. He makes a really good analogy where in the early days of online shopping there was this panic of is, is online shopping? Is everyone going to do all of their shopping online and in person shopping is going to go away tomorrow? No. Still the majority of search is happening in Google in traditional search platforms. And also people are using AI search. Um, and it's creeping up more quickly because we have a faster adoption right now for technology. So we're not going to see a five year long adoption cycle. We're going to see it probably over the course of a year. Um, so it's something that is catching on much more quickly.

Speaker A: Let me tell you what, if I never have to walk into a retail store again, I'll be very happy. But I think I inherited that from my father who died before Amazon was a thing. But man, he would be so happy to never have to walk into a Sears and Roebuck again.

Speaker C: Um, my sister in law makes fun of me because I get both my razors and my shampoo mailed to me. I don't care.

Speaker A: Like, like I'm not. 100 of the Hawaiian shirts I own and there are 72 as of this past weekend have been purchased online. Like, I do not go into a store. Um, that wasn't the case as of last Thursday, but I got rid of that shirt. We're gonna move on. Um, it wasn't a culture fit.

Speaker B: We won't ask.

Speaker A: It wasn't gonna work for this show. So I get that like people are going into like Gemini and be like, hey, I'm going to a road trip or I'm going to visit my family and I'm driving here to like build me an itinerary. Like that makes sense because that's like an exit. Like, I wish search would have done that 10 years ago. Right. But like, I'm hungry for chicken wings today. Like best wings near me. Is that also happening there or is that not happening there? What does that look like?

Speaker C: Yeah, that's actually a really, that's a fun and great kind of nuanced quote. Like I, you're right, I don't think it's happening. And this is all like qualitative, I have no idea. But my behavior isn't that. Which means no one else's is.

Speaker A: Right.

Speaker C: Like, I don't. I don't think that's happening. It feels like it's much more in the planning stage, but also. Or rather and also maybe on your phone, it's different. I mean, we see it in searches in general. Right. Anything on your mobile, especially for restaurants, has over an 80% chance of happening in the next hour. I don't think we know enough yet on AI search in mobile versus desktop and that behavioral change. And again, I'm a millennial, so big things happen on a computer, so I haven't, like, made that transition yet. Um, but I'm hoping to. But, um, yeah, I don't know. I think that's a great. That's a great thing to think about over the next year and try and identify, like, the immediacy of an AI restaurant search.

Speaker B: I think what's happening now is it's happening in Gemini because Google is building Gemini Search into Google Maps and into Google Search. So I don't know yet if people are using ChatGPT or Claude in the same way they're using Gemini, but because Gemini is baked into the systems we already use to find things. Yes, of course. And people we know that search behavior is changing because of this. I would hazard a guess and say people aren't searching best pizza near me anymore. They're searching, hey, Google, what's the best pizza in the East Village that doesn't have too long of a line and that I can get a vegetarian piece? That's what we're having the conversations with our search engines and through voice search in a way that we weren't doing even two years ago. I think we're going to see it first with Gemini and then it will be adopted as Claude and chatgpt become more voice search. I'm in my car and how do I enable this other engine?

Speaker A: I totally agree. I think, um, search in general has become more contextual. Like this me still doesn't get me what I want anyway. And I've noticed that, like on my phone, Siri, which my Siri is a boy, not a girl. And, uh, my ch. My son chose that I'm not have any presidential, uh, prejudice against the sex of my AI assistant. But my kid liked the Australian man. Anyway, my theory will ask me, do you want me to use Chat GPT? So like I say, you know, if I ask a question that's not just like, hey, what time is it in Brazil, which I guess I've probably never done. Um, it'll, it'll go. Do you want me to use Chat GPT? So my agent actually goes into Chat GPT for type those types of things. And then Kelsey, I think you're right. I think Gemini is going to be sort of the AI search winner because it's getting forced into our search feed. You know what I mean?

Speaker C: Yeah. Well, they're the, they're the, they're the queen of it. Right? Google. Google does all that really well. And, um, I don't know, it's interesting to see even like in our household, my wife is all in on Gemini. She uses it for everything. Um, every step of the planning, just general questions. All in versus on my side of the house, we do. Claude is my doer versus Gemini is my searcher, if that makes sense. I use Claude cowork for things. But then Gemini is where I built my restaurant preferences and I have an AVI restaurant bot, or rather AVI restaurant gem that will go and find things for me in different cities. Um, but we're still. It's funny, my brother in law is going to Japan in September and which means for the last eight months he's been sending us restaurants and bookmarking and he has crazy lists on Google Maps. And literally yesterday he's like, I don't understand why I can't ask Gemini to plan my day based on my Google Map list yet. And we were trying all these different. It's not in. It's like data isn't allowed to cross over yet. I was telling him we need to export and upload a gem and he's like, I just want to push a button and it can't happen yet. So I think we're not all the way there. Um, but it's getting pretty cool pretty quick.

Speaker B: The development's happening much more quickly for sure.

Speaker A: Amen. Um, I want to come back to this, um, because I'm going to ask you some how to questions, which is what I typically do. But look, we all just came from the restaurant leadership Conference. Um, for those that haven't been, it's kind of like the show. So if you're an executive, I always joke with people like if you're the CEO of Domino's, uh, or Applebee's, uh, or whatever. Like that's the one show you go to. That's where executives go to like learn research and it's where the rest of us go to see where the smart people are doing. Um, what are restaurants talking about when it comes to this, like to AI search. Like, do they care? Do they not care? Are they doing something about it? What are they doing? Like, what do you take away from rlc?

Speaker B: I think it depends on. I'm jumping in here. Avi?

Speaker C: Um, yeah, please.

Speaker B: I think it depends on the size of the restaurant and the restaurant group. If you are a multi unit with a marketing team, you are probably thinking about this in much different way than an independent restaurant owner or a small business owner with five locations or six units. Um, we're seeing a lot of proactive behavior from our restaurant partners and our clients that are a larger unit count and have that marketing team, they're getting ah, ahead of it. They know that this is something they need to incorporate or they will fall behind in search and discovery. What we're finding as we talk to other restaurants that might have slightly smaller teams is it's not top of mind yet, it's middle of mind. Where if we bring it up, if we say, hey, how are you approaching AI search? They go, oh my God, you're right. This is a huge problem for us and we don't know how to address it, but they're not asking. And so that's something we're working really hard on coming out of a show like RLC where we're talking to more restaurant groups and higher level leaders for whom AI is really top of mind. Yeah, that's going to be top of conversation. But how do we help the restaurant operators that aren't able to come to a show or that are learning things by searching online and figuring out what they need to prioritize? We're trying really hard to make sure that this is something we can educate about and can be top of mind for those that aren't actively searching for it yet.

Speaker C: What's been really nice is there's already this built in knowledge of, okay, I need to be in the local pack, I need to be found on Google searches. And so that transition to just understanding that Gemini is really important and making sure you're structured to be found on Gemini searches because again, when the, the leader of search puts together a new tool, that's where people are going to flock to. Right? And so being able to make sure that you are structured the right way is a really easy transition for people. And it's also, we want to make it easy for you. Like, hey, think about an FAQ situation, right? It's not point and shoot search anymore, it's conversational. So having that structured data isn't a huge leap to make because people already understand the importance of being found.

Speaker B: And I'll add another layer on this where we're in a more difficult economy than we were in a few years ago. We're having a lot of conversations about retention. Rako, who's not here, but is another person on our team whom we love and trust. Um, you can't retain guests you don't have. If we know that most guests don't return, you have to keep getting new guests in the door. You have to keep performing in that discovery, or you won't have any guests to retain. And as we are in, like, a place where people are struggling and they're maybe spending dollars, uh, in fewer places than they were, that becomes even more important, because if you have an opportunity to catch a new guest, you want to be the one they choose. You want to be the one that shows up in that search.

Speaker A: Um, I'm going to double tap all that. Um, and it's funny, I think back, Avi and I used to work together at yext. I think back to our Yext days. Like, we had to convince people that search was important, which is funny to me. So it's nice now that when we go to what Kelsey said, like, how are you coming up in Gemini searches? They're like, oh, yeah, that is important. Like, I don't have to educate you. Just have to show you how to do it.

Speaker C: Well, it seems to follow me in my career. When I was at Yelp, people didn't believe that reviews were important. It was just this angry feeling about reviews. And now in wonderful 2026, Search Matters reviews. What a time to be in restaurant tech. Took forever.

Speaker A: Forever. Um, uh, the thing I'm. I'm thinking about, right? And so, like, my. My. My search AI takeaway from RLC is that I think some of the smarter brands are getting ahead of figuring right out. But I think some of the smarter brands are actually going into AI agents and asking how their brand shows up in searches. Like, they're literally like, hey, AI, how do people find me? Uh, and so I'm fine. That was sort of my big unlock. I was like, oh, I guess I could ask AI how to use it.

Speaker C: Oh, yeah.

Speaker A: But it should tell me, right?

Speaker C: What's crazy about it, though, again, is even just one. One word change in your search from time over time will lead to, again, drastically different results. Right? Just, you know, from the. From the. Give me, you know, three pizzerias near me. To give me three pizzerias near me with a playground and just all of a sudden, all these different things, right? And whereas on maybe a traditional Google search, you would still get one without a playground in there, because maybe in a review somewhere, it mentioned, oh, I wish it had a playground, or whatever. It would be. Right. You could still get that in there. You're not going to see that On a more refined conversational Gemini search quest, it can really lead to such different results that we are still in the wild, wild west of, um, tracking AI search analytics. Right. And the only way to really do it right now is what you just said is just hounding these LLMs to try and find how many times you came up and how you came up. Um, so we're starting to see more, you know, need for LLM tracking, um, which will be huge for us when we can get there.

Speaker A: Um, yeah, we're off this show. I'm gonna go search my Gemini for pizza with playgrounds, because that sounds fun. And I want to know what's going on in Central Texas that you have these that I don't have in New Jersey. But sorry, Kelsey, you must say something.

Speaker B: Oh, I was going to say, like, I'm going to oversimplify this, because there's a lot more to marketing than this. But coming at this from a marketer's perspective, we kind of care about two things. We care about what can we control and how can we measure it? And right now, as Avi said, um, we're kind of in the wild, wild west of AI search. We don't really know what we can control. We can control a lot less with LLMs right now than we can in traditional SEO. And we're not quite sure the best way to measure it because it's so customized for every person that's searching. If you put in a prompt, you're going to get a much different answer than if I put in the exact same prompt. We are trying to figure out what's the best way to measure success. We can't think about it in the same way we think about how do we control and measure our traditional search performance and our SEO. We might not be able to control every piece of information that ChatGPT could think to surface in a search result. It incorporates so much more than traditional SEO does. This will get refined as the industry figures itself out and finds that balance. And what we start to measure will start to become what we want to control. But in the meantime, we are working really hard to figure it out the same way as everyone else is. We've got. Because we have friends like you and like the folks at Google, we've got a little bit of an edge and a head start. Bye to whoops. It still is, uh, an adventure. As we're going.

Speaker C: It goes into one of my favorite sayings when it comes to just search in general, but specifically now, with AI, it's like, you know, you need to have structured data. You need to have schema, you need to have your menus done correctly, you need to have FAQs. I don't know why I eat carrots, but I know they do something good for me. I'm not exactly sure what, but I know I'm supposed to eat them. And it's kind of that way with structured data and schema and photos. I cannot point and say specifically you were found because of this, but I know that one of the 12 things we're doing here is leading to success because we see it right, and we see it sometimes it's weighed heavier than others. But like, eat your carrots and have structured data. It's a straightforward thing. It helps.

Speaker A: I just finished a salad, there were no carrots. And I feel guilty. Um, okay. I want to know if I have this correct or not. So my understanding about LLM search or agentic or whatever word you want to use is a mix of listings management, reputation management and structured data. Um, is that correct? Is that incorrect? Am I off the path? Am I somewhere near it?

Speaker C: You're definitely on the path. And we can share some of the research that came out recently. When we talk about structured data, that's your website and it is heavily valued, right? We've seen certain research that shows up to 40% of LLM, um, searches will be impacted just by your website structured data. And I'll just give you an example. We were talking to one of our hopefully soon clients and we were going through their website and their menu was structured as a blog, which means it's not being found for restaurant or menus or items. It's literally a blog in schema. So when the robots speak to each other, it's not coming up as a restaurant. So you're not even getting that back. That's just one piece of the puzzle. But we know that it's important.

Speaker B: I'm going to jump in and do a definition of structured data quick because we live in this all the time and I didn't know what it was to begin with. When you're in a Google Doc and you start a brand new Google Doc, it gives you the structure of that document, it gives you the header, it gives you the Subheader, and then it gives you the body, and as you start to type in, it automatically structures it for you. That's what structured data is. And restaurants have a huge edge on this AI search because what is the most structured thing that you could possibly put on your website? A, um, menu. There's a header, there's categories, there's descriptions, there's items, there's photos that go with it. It is the perfect example of structured data. And it's such an opportunity to put that in your website as an embedded or as something that lives and is in a PDF document so that the AI engines, as they're crawling it and surfacing your results, they see that menu structure as something that they can read. And structured data just means between structured data and schema markup, which is the translation of that into the backend code of your website. That's what they can read. That's the language that LLMs know.

Speaker A: Amen.

Speaker B: Carry on.

Speaker A: So I went into Gemini the other day and I was like, do you know who Rev C Anzio is? And then we played this game, like, where I pretended I wasn't me. Um, and then I got a bunch of answers and I was kind of happy with all of them. Then I clicked the Sources button, which, by the way, if you're a restaurant and you haven't gone into your favorite LLM and asked, what do you know? And then click the Sources button. You can pause this conversation. Go do that. Anyway, none of the sources were like my LinkedIn blog, my website. Like, none of them were my properties. And I was like, well, wait a minute, how come then for a restaurant, my website is so important? Like, Gemini was not able to answer that question. But if you guys have an idea of, like, how it works differently for a business versus a person or an entertainer,

Speaker B: it's a good question. I think it probably depends on, again, that prompt, who's doing the search and what platform are you using? And there's the element that we forgot about in its structured data. It's reviews, it's listings management, and it's a little sprinkle of fairy dust that we're constantly trying to pinpoint.

Speaker A: Um, I think the reviews are probably more important than a lot of people are giving credit to. You know what I mean?

Speaker C: Reviews are important. And then some of the fairy dust is citations. And so we have this geo scanner tool we'll run to help people try and identify the inputs. And it's incredible what happens when you have a mention in Eater, right? A mention in Thrillist. A mention in infatuation, what that does, um, to your sprinkle of fairy dust magic. Um, because again, when you look in the sources, you'll see, like, I'll give you a real life example. We're going to San Diego at the end of the month and I've got two kids, five and two. And I'm literally like, besides Legoland and ice cream, what am I going to do for five days? And the amount of. I don't know. The only way to say is the amount of mom blocks that I have been going through as sources on Gemini are wild. Right? It's mentioning restaurants, it's mentioning activities, it's mentioning things at the, at the hotel we're going to be at. Um, and so that citation is part of that fairy dust and can't be ignored. It might be 5% of it, but that could be a big 5%.

Speaker A: So the answer is SeaWorld. Um, let's move forward. I think we're all going to the same place this week. We're all going to Megan nra. What do you think is going to be the big conversations at Meg? And for those who don't know, that's the marketing executive group. It's a small, intimate trade show for marketers, by marketers that happens the two or three days before National Restaurant association show. But what do you think is going to be happening there, AI wise?

Speaker B: I mean, on the agenda for the actual planned roundtables and the speakers, about 30% of the content is about AI. AI search. How do you use AI in your restaurant? We're focused on AI search because that's what we help with. There's so many other ways for restaurants use AI and I think there's going to be a lot of conversations around how marketing teams can. It's kind of what we're focusing on at this point is how do you maintain your humanity and your hospitality that only you can do with high touch service and high touch hospitality? And how do we choose what to use AI for? That doesn't remove that humanity. And I'm really excited to hear from the other people in the room how they're planning to approach that, because that doesn't change from a restaurant marketer to a restaurant tech marketer to somebody else that works in marketing. I'm really excited to learn from, uh, these other very smart people in the room who are trying to find the same balance.

Speaker C: We've come a long way from the last few years where it's like, here's an AI grill system to real ways to move your business forward, move your restaurant business forward, and take some of the, um, repetitive, mundane tasks that can either happen quicker or with less lift. Um, we were at an operator dinner back in March and I was blown away by the three or four operators I was sitting by. And at one point everyone's talking about how they're using whichever LLM they're in somewhere on Gemini, ChatGPT and Claude. And there's one operator saying that he has every morning his toast revenue data from the day before pulled into Claude then goes out to Excel and he's looking at his own graphs with the things that he cares about. Um, and you start taking that one step further. And can you pull in your review data? Can you start tracking the way that your menu items that are mentioned in negative reviews correlate with revenue and vice versa? What happens when you have a, uh, menu Item mentioned in 104 and 5 star reviews over the previous 60 days? Do you see that item start to track? And that's where I get excited and where I want us to start thinking about and is how can you start? We. I've always said hospitality groups don't have a data scientist, but now you can. And that is exhilarating because those are where those little details is where you can start to really move mountains as a restaurant.

Speaker A: So my, my new movie, you'll like this dovetailing off that is, um, I'm taking three to six months worth of reviews for a location or brand from Google and I'm literally cut and pasting them and putting them into an LLM that I've sort of trained to understand the business. And I'm asking it to bucket it into five to eight personality types. Like, can you look at the way my guests talk and bucket them into how they see themselves so that I can understand, uh, how my guests act, think, and the volume of which they do. That's been super helpful. Like, how do my guests see themselves? That's crazy. Again, I am taking data from one place and putting another. But it was really helpful.

Speaker B: We've seen you pull some of the data.

Speaker A: Yeah, I mean, can I get this? Yeah. Amen. That's good. Um, all right. So, Meg, small, intimate, lots of marketers ask questions I'm not sure everybody is listening to. This is going to go. But NRA on the other side, completely the opposite. 14 football fields. I don't know if you knew that. Stat. It is the size of 14 football fields. What are you expecting out of NRA?

Speaker C: 30, 000 steps a day in A stomachache. Next question.

Speaker B: I know that there's a presentation I'm really hoping to catch from Lisa Landsman at google@nra. She's speaking on a lot of these things. She's speaking on the future of search. She's speaking on how, how AI search works. Um, some of the developments that Google has coming out for small businesses and for restaurants in general. So I'm looking forward to that. Um, I mean, we also just love hanging out at our booth. We like to learn from the operators that come by. What questions are they asking? We're going to have a free, um, AI search tool that Avi helped build that say, uh, it's, it's a scanner, so it'll scan your website location. Tell you or not your website location. I'm sorry, your restaurant location. Tell you how you're stacking up on reviews. What is, what does your menu look like? Do you have the right structured data in place? All of that. So, um, we're excited to see some of the results from that. But Lisa's talk is one I will not miss.

Speaker A: Awesome. Are you guys in a similar place as your booth was last year?

Speaker B: Yep. On the left side of north hall, right before you, right after you get to the bathrooms. Buy the cheese.

Speaker C: We're always trying to be by Dairy Row.

Speaker A: Yes. So if you need icy cheese or pretzels or to understand how your, uh, restaurant is coming up in AI searches, it's all in one area. So.

Speaker C: And I'm happy to walk you to the Dippin Dots and explain along the way the big question that I've been thinking about. And my, my brother in law is actually going to come this year to the show and I was trying to explain to him the hot dog situation and I just, I'm really eager to see are they back or was there collusion between those marketing teams and like we're just not going to spend the money. You don't go, I don't go. I really want to know. I'm very eager to find out Nathan's Vienna. What's going on.

Speaker B: I also want to know if there will be bags of pickles again.

Speaker A: Okay, I'm just gonna go to the Kronos because Kronos is not far from.

Speaker C: Oh yeah, yeah, Kronos. Chronos hits.

Speaker A: Okay, before we sort of wrap here, uh, I'm gonna do a little shameless self promotion. So if you are going to NRA and National Restaurant association show is called by its full name and you want to learn more about this for sure, you should Go stop by the marquee booth in Cheese Row, or Dairy Row, I guess we're calling it. There's also ice cream there. Um, I am co hosting a panel with Kenny Charlotte. Um, it's on Saturday afternoon, where we're actually going to share five foundational AI frameworks like do step one, do step two, do step three. And one of them is optimizing for Agentix Search. So if you literally want the playbook, come see me and Kenny Saturday afternoon and then go to the marquee booth and see if it' already working or not. So, um, sort of before we depart, do you guys want to tell everybody kind of quickly what Mark he does like how this solution works?

Speaker C: Yeah. The easiest way is just think about us as we're here to help you get discovered and chosen. And so that's going to be through a myriad of, uh, backend ways, such as structured data, making sure you have great reputation management, easy to respond to reviews, and a plethora of analytics to make sure you're going up into the right. So we're here to be your partners. Uh, ideally to help you get discovered, but also, whatever your dreams may be, we're here to help.

Speaker B: Couldn't have said it better myself.

Speaker C: Marketing. Is that approved. How to do. What would you say we do here?

Speaker B: I'm not here to get in your way.

Speaker A: Okay, so people want to learn more. Come, come. See you at Meg Kelsey. You'll be there, right? And if people want to. Or at the National Restaurant show, come by your booth over in, uh, Dairy Row in the north hall. Uh, they can also go to your website, LinkedIn, all that kind of good stuff.

Speaker C: Yeah.

Speaker B: Qi. There's no U.

Speaker C: And what's our. What's our booth number? I want to say it's 55.

Speaker B: 55. 22.

Speaker C: 22. 55. 22.

Speaker A: Awesome. Well, I will see, uh, both of you there. I'm super excited. I actually get in tomorrow night. I'm flying Tuesday. Me too. Are you really? Okay, God bless. We'll talk offline. All right, everybody, thank you for joining us today. Uh, you know, if you like what you heard, definitely go check out Marquis. Either you can do it right now, uh, or come by the, uh, come by and see Kelsey at the Meg show, or walk over to the booth at the NRA show, or just go check it out online. Uh, if you enjoyed the show today, please subscribe, uh, to the show on YouTube or iTunes or wherever you like to learn. And, uh, thanks for joining us today. We really appreciate it. Thank you.

Speaker C: Thank you.