Getting more sales analytics manpower with AI w/ Jeremy Hart, Somm.ai
38m 21s
Jeremy Hart, co-founder of SOM AI, shares his journey from working in wine service and wholesale to creating an AI-powered platform that transforms how wine is sold to on-premise accounts. Inspired by his experience mentally cataloging accounts for wine placements, Hart partnered with a coder to build an app that originally targeted wealthy travelers but pivoted during the pandemic to help restaurants sell wine retail. The platform now uses AI to scan and tag wine lists from over 100,000 on-premise accounts in the U.S., Canada, Europe, and Singapore, updating data every two weeks. Users can filter by wine attributes (e.g., producer, region, price) and account types (e.g., steakhouse, sushi bar), enabling suppliers to identify market share, track competitor placements, and target accounts based on geography—such as near NBA venues or retail chains like Total Wine. Hart emphasizes the platform’s ability to provide “air superiority” by eliminating legwork and allowing sales teams to focus on direct, data-driven outreach. Long-term clients like Jackson Family Wines have used SOM AI to reinforce partnerships and boost placements. Hart notes that the pandemic accelerated the need for digital tools, and SOM AI has become a leader in wine data and AI, helping an archaic industry adapt to modern sales strategies.
Hey listeners, have you heard of the term "brand differentiated commerce?" Probably not, since I made it up chatting with the folks at Offset about what they do. Offset is a proudly independent, wine commerce platform and brand studio rolled into one based in wine country. Their platform, Offset Commerce, has been shaped by listening to smart wine businesses over the last 16 years. And unlike other platforms on the market, it empowers individual wine business models, rather than a one-size-fits-all approach. That's why iconic wineries and merchants, like Larkmead Vineards, kill Cita Creek, and Kermit Lynch-Wine Merchant, trust Offset Commerce to handle the unique ways in which they sell. Go to OffsetCommerce.com to learn more. That's OFF-SET-Commerce.com to learn more. Welcome to X-Shatto. X-Shatto. The podcast that navigates the business of wine with unique perspectives and insights. With your host, Robert Vernick, and Peter Young. Welcome to this episode of X-Shatto. Today, we are talking with Jeremy Hart, the Chief Strategy Officer and co-founder of SOM AI. And we're going to be talking about leveraging data and AI to sell to restaurants. Can you please give me and Peter a brief overview of your background, Jeremy? Yeah, so first started in the business as far as the service alcohol business. Started in the late 90s. I did that for quite a while. And then I got a wine into '04. And once I got into wine, it was kind of game over for me. That was going to be my only pursuit. It was just going to be which avenue to kind of follow. Let's work on this little small wine bar that just did insane sales in Houston. It was way ahead of its time. We did like 50 wines in the other class, 16 flights. This was in like '05, right? We used to do almost $300,000 in sales a month. And the place was 5,000 square feet. It was just bonkers. We always had a glass of wine that's $50 by the glass. I got to taste a ton of wine constantly. And so that's where I got to cut my teeth on wine fell in love with Burgundy, Borolo, all those typical things. And then from that point on, it's in the service industry for a little bit longer. And then finally in '09, I moved over to wholesale. And then my first job in wholesale. I kind of fell into a really good role because it was way ahead of everyone's far as kind of wine knowledge for the time. The internet, what it was for wine at the time. I had this really phenomenal route and then I covered a major retailer for the whole state of Texas for a mid-size wholesaler. And so I did that for four and a half years, moved out or to be a supplier. And I worked for really a nice procedure-send porter for seven and a half years. And in that time period, I managed about 56 brands, about 430 SKUs across six states through 10 wholesalers. And so that's my background. At that point in time, we're predominantly on premise. And so I always kind of had that in my mindset of how do you sell X amount of wine and you know, manage allocations and all the other things, covering so many accounts with so many SKUs. So I'm just a quick tangent. Like you said, you had a $50 by the glass. What was the craziest thing you tasted? I'm just curious back in the day, what people put on the day? Like it was wild. Just crazy. Like we used to have a contest there. And this like one of the craziest contests I think back on it. And I didn't think I was a salesperson at the time, which was really wild. But we had a contest if you could sell $10,000 in retail in one sitting, to sit down to have a glass of wine and order one plate of food, you got a $50 by the wine off the shelf. And I would hit that probably three times a week. And I think back on that, and I was just like, you know what I mean? And so it's just like crazy. Would be going the back and go get the hang cart and you're just willing all these like crazy, bordeaux, everything and the people's trunks. And we had a locker program there. And then a weight came and the big crash and all that went away. But that was another kind of interesting thing that happened for the first time Houston really, or Texas really got a knowledge with all the allocations. Because right up to the way crash, New York, California can take them. And then all of a sudden people saw Houston as their Texas in general. It's this place that could take everyone's wine and they did. And so then from that point on, the wine industry in Houston and Texas at least changed quite a bit. You know what I mean? It wasn't just all that oil money. All that oil money. That's super interesting that that opportunity happened around that time. And that starting point for it. It's almost like, you know, the removal of taxes in Hong Kong. I mean, Hong Kong take off right. And so it's funny that Houston and Texas take off at the same time. Yeah. So to some AI, what is some AI and why and when did you found it? So when I was on the import side, with what I was doing, I was all over the place all the time. Right? I was traveling about 65% of the year. I saw that I had to touch all these account constantly. And every time I go into the any account, the first thing I would do, you know, because half of my job was taking people to lunch, which was fantastic. Three hour lunches, bags of burgundy, it wasn't bad, right? But the first thing I would do if I ever walked into a new account, I would sit down there and I come through the list. Sometimes get there early and just look at everything. And in my head, I'm just like thinking, all right, cool. I'm going to talk to this guy about this. You know, just kind of pre-setting what I was going to position because I had a lot of toys to play with. That, you know, 50 plus different brains to sell. A lot of Italian, a lot of French, whatever, right? And so in the process of doing so, it's just all those things always stuck in my head. And anytime we get, we had a new part that we picked up or if we got into a new region or appellation, I'd already have 15 accounts on my head that I would sell that, right? When I first saw that this could become a thing, I was in the 2019 buddy, David sent me this app that he was working on. And I kind of looked at it, and it was super rough. And I was like, what is this thing? And the whole original idea of somebody I was for like super rich guys, if they're traveling, you know, and they're Miami and never been in Miami, they don't care where they go eat. They just want to drink rapinoe or who has rapinoe, right? And this was still when apps were like, oh, yeah, I want to make an app, you know, no one wants to make an app anymore, right? No one wants another app on their phone. So anyways, this was in 2019. In 2020, the last major trip I took with the former importeries to work for us and it was in New Zealand for like 12 days or something. He sent me another new version of it. I start playing with it and I start thinking about all of the ways in which I would use this, right? And so I would reach out to him, I say, this looks incredible. Can you make it do this? He would have recoded it, send it to me the next day. And so we just kept doing this the whole trip as soon as I get back, pandemic happens. And I'm really like blown away of how much this could change the business engine, right? And at this point in time, no one's talking AI, no one's talking data and why not really, especially not in imports, because it's all like a delk, it's no flake industry. On that side, like we're way above that, like, you know, we're just know, okay, right? When you want to know where every ludiza, okay, cool. So anyways, when we get back, pandemic happens, reach out to David. The first thing we did, I don't know how long of a story I want, but I'll get through it all real quickly. Anyways, reach out to David and we come up with the idea to essentially turn wine list into retail shops. And so we reached out to all these restaurants, the first one literally on St. Patty's Day, the day the world shut down, or reach out to landries. And I did a lot of national count work as well in my background. So I reached out to them and I talked to them and I'm like, hey, I instantly I can turn on your restaurants into retail shops. And so in the process of doing that, we were taking everything that was on their menu. These weren't done as just general tags. We knew that these were the burrullos. We could add nebielos as sortable features, so on and so forth so people could shop them very, very easily. Within 60 to 90 days, we sold about 700 K in retail across Maastros, Commanders, Palace in New Orleans, Fyola, all this different stuff. So the next thing that we started to do was, you know, this thing should be commercialized as what the original tent was. And so we did some smart menus for a while for some clients. We did all the Delphuriscos and Maastros as well, but initially, or finally, we started getting into the commercialization of the platform for what it is now. Benchmark and Leach Generation #1 and #2. And we finally launched the platform you see today. A very rough version of it and basically at the end of 2020, we're so far ahead of the AI, we're the original AI company in line by a mile. So it sounds like there were a couple pivots in there all within 2020 because it started in the pandemic. So that was really fast pivoting, right? Like, you figured out in a few months that this wasn't the right direction and then pivoted to a different direction until you found this because it sounds like maybe pivot number three or pivot number two, I guess, if you know, it was retail first and then like the wine list for rich people. It was always a thing to have some revenue coming in. Like, what could we use this for right now? And the thing was too is it's really hard to say whether we would have done what we're able to do now if the pandemic didn't happen. Pandemic helped us in a lot of weird ways. It also sluttestowns significantly. One because on-premise was dead until basically 2023. That's when companies started reinvesting in it. But by that point in time, we probably had about 20, 30 clients, which is great. But the other thing was too is that you couldn't find the menus online as accessible prior to the pandemic. And so like both of these things, it's just weird. There was a poll in a push and like, you know, it shook out to be a good thing at the end. And by that point in time, we were pretty well set up. We had enough users on the platform to have it really tweaked out and understand use case scenarios that our customers wanted. And so by about 20th, 23 when people started talking really about AI and data and you know the industry is archaic as hell, you know, they don't want to change. And so they don't want to hear about data and AI for a very long time. And so two years of being a recluse to launch this thing, but yeah, it's falling worked out well and going well. The thing is really interesting because when you let off, you're like, you were explaining how the things you did when you first met a customer and how you kind of cataloged and things. And then where you ended up with the app. I'm wondering if that's all hindsight reflection. It doesn't sound like you start off with that intuition to try to make this like that. It's just, it's all once you got there, you're like, oh, I used to do this in my head. Now I have this app that can do this. Well, the funny thing was I was blessed to have two really good mentors in my life. I Deeter Cam, who is phenomenal. He was the one that got me into imports and prior to that, I got it named Whitmeyer's. And so he was in wholesale and we used a hammer and everyone's head. We're going to do a general sales meeting and we're going to sit here for eight hours today and listen to 15 suppliers and taste 200 wines. At least right down for a restaurant that you would sell a wine too. Just like if you're taking 30 seconds, these guys came here to travel at the end of this. You have your list of wines. You can look back on that and you'd be like, this would be great in this account. That was all when that started. That would have been like 2010. And so that was always like the little training for how I would always do it moving to
forward for us in my life. And then after seeing what this could do with AI and be able to sort all of that, I would outsell a lot of people because every time I would, oh, that would be great here. And the whole industry is built off of that. It's built off of relationships. You team up with other rats. You team up with other suppliers. Hey, you should talk to sevens how they're looking for this, right? But if you had that, like the air superiority of sight to understand what the market looks like and who to actually talk to, you cut all the leg work out and you can just talk directly to the end result of who you need to sell to. And so that was the thing that was like the aha moment. I was just like, if this can do that, like you don't have to train anybody anymore, you can literally just give them the list of lead. And so yeah, when I saw the detail of how far we could tag down, that was when it was like, oh, this is going to be crazy. What do you mean by tag down exactly? Like so essentially like anything that where we're finding a menu, we have pre-sorted intact. And so like the AI major piece that we use for the platform initially, and this is the thing, the precursor, my business partner's in Kredway, right? It's 10 years younger than me. Incredibly smart, right? And so he was the coder to this stuff. And so the archiving of information, so when we use AI to figure out what the SKUs are, right? Once we know what that SKU is, all the other things now become sort of. So we know where this part of our brand, whether that brand is part of a wine group, if it's part of an importer. If it's from this, you know, country, this region, if it's from this place, then that's this bridal. And like all of the individual sort of pieces to it. And so like, and there's been some really kind of like aha moments that have been really funny to look at. Like, and I'll do a little tangent real quick. If you look at Nebula, right? Brolo is by far. It's like, I can't remember the numbers off the top of my head, but it's like, it's almost four to five, right? Brolo to Barbarasco, okay? What do you think the number one Nebula producer is as far as my placements in the US? It's Gaia. Yeah. Got to ask more placements than everybody. It's Barbarasco, right? Like number three is like Proditori, Barbarasco again. But if you just chop it up and you don't think within the categories of Barbarasco, you're looking at it's Nebula, it gets really, really interesting, right? Are those the same customers? Pretty much. So just interesting stuff like that pops up occasionally. So it's just, it's been fun to see these things. And sometimes we'll look, I'll look at some data or whatever I'm putting together. I'm like, I'm the first person's ever seen this. That's pretty wild. So yeah, it's been neat. So you focus on pulling on-premise data and building it into a database. I'm curious, is that include all on-premise sites or is it mainly focused on restaurants? No, restaurant bars that tells. So basically anywhere it's consumed on property. So kind of all the above. We also sort by what all the restaurants and kind of like everything category that they are, you know what I mean? Whether they consider themselves to stay-cow, sushi restaurant, cocktail bar, champagne, you know, whatever. All those things are all sortable as well. So the same way we chop up wine, we also chop up accounts. So that includes some of the bigger like a stadium or something like that, you know, like chase center or whatnot. That one's hard to do, but there's been really interesting applications for how we use the platform. So like one of the really interesting ones we did, so Jackson's been one of our longest clients. They've been with us for basically six years, which is pretty wild. And so when they did their big partnership with the NBA, we threw a one mile radius around all their renus, right? And so we target all the on-premise accounts around them. And then we pre-sorted all the wines that they would have when they're menu that are USA under $100 wholesale or $100 on their wine list, right? And so now they and they also have the categories of the restaurants are next to it. So they're going in there, they're dropping table tints, they're making new placements, they're taking out their competitors. And once again, they're like re-incorporating that were with the NBA, you know what I mean? Because all those restaurants on-premise accounts get the business from the NBA when those games are happening. And so they're happy to support it. So there's been interesting things like that. We've also done it with like retail as well, where we can drop our radius around all the totals across the US, right? And then if you're having a meeting with total, you can walk in and say, "Hey, I'm your number one sob you on block." By the glass around all your locations, we have a built-in customer base, like we should be all over your store, you know? So there's all kinds of neat applications that we've done with the platform besides just using it for just general on-premise. And so how many accounts do you represent in your database at the moment? Total in the US, we target 100,000 accounts. And our frequency is like the really wild thing that's happened is that our data has become so rich, right? We were by far doing this way before anyone. We started targeting accounts in 2019, commercialized it really in 2020, right? We started launching and training in January of 2021, okay? And our frequency is that we update every two weeks. So we're resurvey all of these accounts every two weeks, track all changes. And so now when you look at trends, like our data is so rich. And that's been really kind of fun to watch. Like percentage of market share has been a really cool one to kind of look at to evaluate companies. We're even doing it by placements or by accounts or whatever. But over time, as we've grown, we've pulled in more accounts. Like when we first launched, we're like 7,000 accounts. You know, we ate it now. We're at 100,000 accounts. And so that percentage of market share, our original size database was still good enough to have a good sample size. And then by like year one, we're already like 23,000 accounts. So it's always been big enough. But now that percentage of market share is super telling. If you want to chop it up, I supply and understand for by the glass and only Italian wines, where do you sit in that? Who's your competitors within that? What's their average price? Is it changed? And then you throw in another filter, you know, what's it for? Whatever, it doesn't matter you see what I'm saying. So yeah. And how many jar fees do you cover? I mean, obviously you mentioned the US and your base in Texas. Like what's your coverage like? So we do the US, right? We also do Canada. Canada was really important to a lot of our clients before the tariffs. Hate that. But as soon as those tariffs go away, that piece is going to be very valuable again. But we just include it with this description. Okay? We also do Europe. That's typically separate. And so we do six countries in Europe, Fraun Primus, the biggest country that as far as coverage is Germany. And so we do about I think it's almost 3,000 accounts now in Germany. And it also Switzerland, Austria, England, Great Britain, you know, whatever. It's about 1,000 accounts in London that we cover. And then we also do Singapore as well. Interesting. And so you mentioned you have a partner who's the technical lead. How did you guys build some AI and how exactly is AI leverages in the system? Yeah. So he's the architect to it. So you know, he's best to describe that. I'm more of the this is exactly what I wanted to do. And please give me every, you know, Bell and whistle on it. It's so funny. We have like an agreement now to where customer has to ask for something before he'll code it. Because he has to do other cutting. And I'm like, well, I'd love it for to do this. He's like, I will not do that. You know, so funny. You're a customer. How the customer right? Yeah. So it's so funny. But I guess sometimes, yeah, sometimes I'll, you know, email of, you know, it always comes from conversations with customers. So when that gets come up, I'm like, I'm going to need you send me an email. So yeah, we've incorporated a little bit more recently, you know, as to help with the coding. You know, that's all him though, but I don't do that piece of it. So yeah, I'm about to architect of it. Yeah. I would assume that plugging in some kind of front end, because like you're thinking the space already now, you've been like breaking up stuff. I would assume having some kind of large language model in the front end to be able to ask the questions. A lot of, yeah, so that they can just ask a natural language and it'll figure out how to then make a query to the database and call that data. Because I think a lot of people don't think about it. Like, you've already so you're so enriched in this that it's kind of natural for you to like parse all this data. And you're looking at it from the wider view. But it's like, I think that like, hey, this is my whinemist. Like, what gaps do I have compared to this in this customer? If you were to be able to natural language that into an LLM and if you able to spit something out, I think that would be super useful. But you got this is not like you guys are there with that user facing for your customers yet that they can do something in natural language. I don't want to add anything to it as far as another AI platform. Any kind of piece to it, but I understand exactly what you're saying. The funny thing is is that we do like with all of our customers, it's unlimited training and unlimited seats to them. So like, and we have massive clients, where some really, really big clients have 500 users on it or whatever. No one's ever ran me out as far as like too much. It's still basically mostly me and David is the meat potatoes of getting everything done. I still handle all the training, all of that stuff as well. But the funny thing is to exactly what you're saying, I've now moved in the past two years. I don't even care if people use the platform. Just ping me. Like, it's I'm still like I'm that bot. If you want to say it like that or like just give me the quarry because I come from I came from their world. You know what they need. Like literally give me two sentences of what you're talking about. I'm fast on the platform. I'll knock it out in 30 seconds. Here you go. And so that's been a really interesting thing. I've had now I think four clients that haven't logged in. Just because it's like, all right, this is what I need. Here you go. Boom. Out of how many? Just to give that up. Wait. So right now, I think we're probably have we're almost close to 70 clients in total. And so, but some of these are massive, massive. Like, country or towards a client. Jackson or I mentioned, it's her lot of ventus in all size clients. Like, we have food clients as well. We have water clients. Most of our clients are in alcohol. Most of them are in wine or at least wine focused. Most of them are either, you know, supplier wine group importer. We do have some wholesale clients as well. It's a hodgepodge. Yeah. Okay. It sounds like most are on the supply import side. And then like, what about like large hotel groups? Now we don't work with any of those people, which is really funny. We do do. Yeah. We do help some of our like, so our clients are national accounts. And the national accounts people have questions. Like, we're super happy to build those reports, right? And like, we can comp against like one of the funny ones is like a, you know, a bariloui or something like that. We know that that's tagged as like a $2 gas drop up. All right. Cool. Let's sort all of those. This is what their wine list looks like. Or this is how many skews they have for beer or whatever else, right? And these are their average prices that they're doing for the locations you have. You're in these states or these cities. These are your cops. They're, these are their prices there. And so we're happy to do that stuff. It only takes a little bit of time. But it really helps the national account teams get placements because they're doing a lot of legwork for their clients. Can you give me like a customer profile? Like, is it the 10,000 plus
or is it even bigger? Like, what's the typical kind of customer that is using the platform? - It's varied, it's really, really varied. It's like the same thing when people ask me, like, where are the general use case scenarios? And without always a benchmark in these scenarios. - My next question. - See them, I found her. - And so when we first launched, one of the ones that we got a lot of were importers. Like importers saw the value of our platform immediately. You know, 'cause like where my world was from, you have a lot of skews, you probably have five or seven people across the country that need to, you know, sell hundreds and hundreds of skews across a lot of different wholesalers. And so this was a forced multiplier and a lot of ways for them. But as far as our clients, like Antica Terra, as far as production, it's probably a smallest client. We also have, I would say like the, at least 25,000 case client is usually a pretty good one. But also depends on kind of what they wanna do and how much money they had behind them and how they wanna get there. 'Cause we could help chop up where the best markets they should be launching in if they're in infancy. All of those kinds of things as well are like really important, you know? We help people speed up the process of selling. And so, and also what does that market look like? You know, do they need this? Is it up or down? You know, are you wasting your time watching their versus over here? We also do some really interesting stuff. We have a lot of friendship relationships in the business as far as like we don't have any of our exchange money. You know, we just help each other out when we can. I'm not doing any of this service by saying it's like Clawinger is one of them. Like I love those guys, they're fantastic. But I'm always intrigued of like the other applications that this platform can do. And so like one of the things that I do with them and some of the other marketing agencies that do a lot of winery stuff or spirit stuff is if they're working with a consortium and that consortium has come into the US. And like let's say it's Trento Dock, right? Where are they gonna go? Like the general answers, New York, San Francisco, Chicago, maybe Miami, right? Always every time, right? Okay, well with my data, I can chop up. Let's go look at Cromata Allsauce. Let's go look at like every other sparkling wine plus Trento Dock. See where the best markets are to go for that. And then plan the route for the tours to do that. And so maybe it's New Orleans, maybe it's Denver, maybe it's Portland, Oregon. And so to kind of change that to get to the right people, it's been a really kind of neat use of the platform as well. And another odd client I'll throw in real quick. We also have some clients that do like vending and airports. And so those are like considered monopolies. And so for them to do pricing on their menu, legally they can only look within like a three mile radius. And so we can help them find like the most expensive chicken nuggets within two miles. (laughs) There's lots of applications, the platform that are just like, it's funny, man. Yeah. - So you're actually pulling in food data. - Oh, that's a lot. - And I always order. - Yeah, like, and like, if you think back to the past of like for national counts and stuff like that, you do these LTOs. Or it's like, all right, well, I can put together this package and put together all the collateral for it. And it only makes sense because I'm talking this rest on group that has 20 restaurants that can do it. Okay, well, what have you put together all that collateral and all the things you wanted to do? And I can show you who has a $25 hamburger and a $30 class in Napa Cab, but they're all independent accounts. And now you can take that same piece of collateral and say, hey, we wanna run this program with you and we'll hop with your social media stuff. So there's like some really cool ways to like, you know, use the tools that the big guys get for the small guys, yeah. - That's cool. You mentioned that the general use cases are benchmarking and prospecting. When you say those words, what does that mean for a wider, important perspective? - Yeah, so like, let's say, and I'm gonna use this as just a general example. Like, let's say you're in Poso, right? In your ex-winery, it doesn't matter which one. And let's say that the comps that you only care about are gonna be, and you can be either one of these five, Austin Ho, Justin, J-Lor, Dow, blah, blah, blah. Right? Let's say there's five of them, okay? Within those five, now you can have tangibles. That's one of the things that we provide that never ever existed for on premise have been tangibles, right? And so across these five, like, let's say that you wanna go look across the whole country and across these five, you rank number two in accounts, across these five and you rank number three in placements, okay? And so if you're in one restaurant that's one account, you have four placements on that menu that be four placements, okay? So you're ranked number two and number three, right? And so now that's a tangible, you can hold to every state, you can hold to every city. And so if you're talking to your whole cell, or you're like, "Hey, look, we need to be number two and number three in this market 'cause we are national." And if they're underperforming in those markets, it's not just like, "Hey, you're underperforming." Hey, this is where all of their placements are, so now you can go take their placements. And so that would be like the benchmarking piece. The other thing is too, is like if you're in Poso, but you care about all Poso, you could just filter for Poso, right? And then I'll pull up all Poso and show you the market share of what's available. And then you'll also find some value in there. Maybe there's states that you're not focusing on, or maybe you're putting too much emphasis on this one area, or maybe you need to hire your fifth person for the country and you're looking at it and you're like, you know what? I wasn't even thinking about this part of the country that put some on that, but there's actually a ton of business here for Poso, this is where we should hire. And so there's all kinds of neat things like that, but we have everything so well tied down that that's, that's all those stuff that that makes more sense when you look at it, you can really figure out those answers. And I'll also help answer those answers, so I don't need them to do the leg work on it. Any of our customers are under data analyst with it. Give me your question, I'll give you the answers. Here you go, boom. - Right, so yeah. Well, I'm sure a lot of your clients have come back after you've done these projects, or they've run it on their own and talked about how big of an impact it's had and the success they've had with it. Do you have any good stories around like people who've been able to sell a lot more wine through your system? - Yeah, obviously we've had some clients that have significantly moved up in ranks 'cause you're able to do this and we can rerun it. And also you can also look at a store for records so if you're looking at, you know, back at trends over time or whatever else as well. We also track loss placements, but that's another thing. That's another huge value we have, especially in this day and age. For your competitors as well, we can track their loss placements so you can go take them every immediately. But yeah, there's been tons. One of the other things that we help do and I don't know if I mentioned this yet or not is also brand extensions. And so right now in the age where fruits like so available and so cheap, right? If you have an opportunity to do so, we can go against your comps, like what we're seeing earlier with Paso. See if there's something in there that you're not doing currently, like the agrossos fly, what is Cab, where Bordeaux, blah, blah, blah, blah, so and so forth. All right, maybe we should buy this fruit. This is the potential for the number of accounts that they have. So the pricing aspect and understanding the viability of a brand extension has been massive because that's helped people like, and it goes in waves. It's so funny, two or three years ago, within one week I had four different clients reach out about organ shard, right? And I'm all like, Googling, I'm like, is there an article that just came out? Like why is that group talking about this right now? There was nothing. It's just like, it's just like group think. I guess they're kind of like whatever. But it's trending. Just stuff like that. Yeah. Helping people with pricing to save money has been massive. That's been one of our big ones. National account stuff. So people are landing more placements for national accounts because that's such a competitive thing right now that has been a big piece to it. And just the general targeting to help people move up in their rankings. You know, that's been a huge piece to it as well. You know, it's also getting more out of your people. Everyone's working with less. I mean, that's one of the crazy things that we'll talk pricing stuff here in just a second. But I mean, how many people are six figure people in the wine business? Two minutes. If you work for wine, who are two million a lot, right? But like, our pricing isn't six figures. Unless you're an excite client, you know what I mean? Generally, we're between about 30 to 70 K a year for most clients. And so within that, I mean, why wouldn't you have us? Just in general, you let six people go that were six figures. Like, you can't. You have now five people to do 12 people's job. Like, pretty good solution. So anyways, yeah. So are there specific ways that, you know, every client is different, but are there generally speaking certain ways to use the system or use you use their data analyst that you've seen had the most impact on their sales or their rankings? The lead generation piece is huge. You know, before this massive chaos with all the wholesalers starting to flip and tumble, which is still super valuable, but there was a day and age to where you could reach up to your wholesaler and ask for a target list. Hey, I'm coming to market, who should I go see? But even still at that, you're asking for their data, you know? At best, you're asking them to switch out a placement they already have with yours, right? Because you fit that category. It's another Russian River producer, the sales Pina or whatever, you know? And those things are treated like time shares, you know what I mean? You're in the market this week, you get your two cases, you need come back in six months, you'll get your two cases. 'Cause next week there's another guy coming in, he needs this two cases. The lead generation piece is quite, quite important. And then also if you need to beat up your wholesaler and they're still receptive to getting beat up. (laughing) The benchmarking piece is nice and helpful as well. Yeah. To show them that they actually are underperforming. Yeah. Another neat thing that has happened with some of our wholesale clients, and this has been a really kind of neat one, is if they want to go talk to ex per, you know, winery, I'm not gonna list wineries, but like if they want to go talk to ex winery, they can go and say, hey, for this market, you know, the whole market for your category is up 2%, you're down 3%, and I have the tangibles to prove it. And so it takes that, even though they're not a client of ours, that winery, for them to have that conversation with tangibles, can really help nudge them from leaving one wholesaler to another wholesaler. Is that mostly working against smaller wholesalers? 'Cause like, you know, Southern has tons of data, all the comps and they'll provide you with some of the comps how you're doing versus other things. Is that I guess, maybe a different manner? Yeah, well, once again, it's their own data. Yeah. It's only what's in their house, and so it's not general market, and so that's where we really come in to understand what the market is actually doing. Although, they're big enough where it's close to. No, no, I'm sure it is, but once again, if you're asking for that target list, or whatever else from that one on, then it's like, once again, trading horses, and where do you sit in that hierarchy? Right, you know what I mean? Are you food for them, right, for the other supplier, right? Well, they need that to hit the goal for the year, so you know, it's all the games. I wonder if that's interesting for someone like us, other to say, like, who are my salespeople?
people not touching better big accounts, right? They've got big lists. But we do that. We do have reports with our also clients. We do it with everybody. And when we run less, it's literally, these are the spots you're at where your competitors are. Literally, there's your blanks, and you can see where you're at now. And so, yeah. - So it's sort of a bit about pricing. You mentioned that the average customer's gonna pay in 30 to 70K a year. But what's the pricing structure? Like what's the entry point? - It's determined by really the size of the client. - The other thing is too, is that we priced accordingly, which was very, very fair. We do have some clients that pay under that, but they were early adapters. And so, like, we reward them with holding pricing. We might nudge here and there up, but for the most part, if you were an early adapter, we hold your pricing. You believed in us when this thing wasn't a thing and so on and so forth. It goes on typically the size of the company. That's the best way to kind of break your nails. - Me news. - Revenue, bottles. - Revenue, typically, typically. And the other thing is too that we have to count for is, how much revenue are they doing actually on premise? Are they a jug or not that, I mean, does it matter to them on premise? And so, we have to evaluate those kinds of things as well. But we have clients that pay well in the six figures multiple times over a year. - Interesting. Yeah, so, but basically, the average is 30 to 70. - Yeah. - Yeah. - Okay. - You said that the original people who bought it, and you're kind of holding their pricing, but I'm assuming you're reevaluating as over time pricing for those newer clients that kind of come in. - Yeah, it's a delicate balance, 'cause right now we want to be as helpful as possible, and as pathetic as possible, with what's going on in the market. The year and I mean, but at the same point, time data is expensive, right? And once it's out, it's out, and that's the reason that we do subscription models. And so, we don't do reporting, we don't sell reports. - And that pricing is similar-ish to, like, Nielsen or IRI or VIP even, right? It's in that range. - I think we're way cheaper, but it's around there. We're usually always cheaper than all of those. And so- - And you don't care if there's everybody in the company using your data. - We do not care. Unlimited training and limited seats, as long as they have the same email address in the back. I mean, that's kind of very little, one little check mark that we look for. And obviously, there's like one or two exceptions, but they're all kind of original clients. X-Wirenery is still part of this, so we'll make that happen. - Have you ever done any case studies or calculation on what kind of ROI your customers usually see? - I haven't done a case study. I probably should do it. One of the things that we hear a lot is the money saved with travel, because they're not going there and they're not at some reps hand on where they should be going. Right? They might still ride with a rep, but they can also, they might spend an extra day in the market and go hit the right accounts. You know, or where they're gonna spend money for dinner that night, that they can look at this and be like, this absolutely does not make sense. And then maybe have a conversation with the old seller. And then also, which market should she go to? Maybe they're gonna go cover whatever place and whatever state and they're like, "Oh, you should go do a ride bill and ride with a Jacksonville." You know, or maybe Jacksonville's the market, they should be hitting. So it's like whatever, but you see what I'm saying. So it's a better spend. - I'm assuming the use case on staff as well, 'cause we've mentioned that if they drop down from 12 to five staff, that's an apples to apples, 30 to 70 K versus seven employees cost. Like there's a clear easy use case there that we've been trying to buy. - Right, I only one staff if you could cut one position and you have a positive ROI. - That's usually the thing, and we're first trying to get people to kind of believe in AI, 'cause it wasn't really no one was talking about when we started doing this. You have to think of this as manpower, not as a platform, you have to. All right, you're saying this is expensive. - Yeah, all right, so you have 4,000 employees. Okay, can you tell me one of them that can do enough work for? You know, like you really have to think of it like that. And so that help people get over that little hurdle initially, but I think people are starting to really think of it more and especially as people talk about assistance, you know, with AI and all of us other staff, so it's interesting. - Yeah, so what's next on the product roadmap for some AI? - What I would like to do is just to continue to explain coverage. I want us to do all of Asia, I want us to do Australia. I want us to be 100% global on-premise, and then I want to see all of that data, you know? And so we're happy to do that, any customer that wants to get wild and crazy and help us commercialize markets that they want to be in. We're happy to do it, and we'll make arrangements to do it on the cheap, you know? But we want people to help us kind of commercialize that, you know? - And are you picking up clients from those other places or is it mostly your domestic American clients that want to sell internationally? - It's not, it's actually your paying clients will want to, you know, some more in Europe. The agency model over there, this is a dream come true, because you know, you might have worked with 14 different people across one country instead of one big wholesaler, you know? So it's, this is really, really helpful with doing the work over there. It's mostly been that, but with now the state of the industry in the US, I think we're going to start seeing a lot more of both, but more for international request, you know? - So if I'm like a major conglomerate, then I want to see how my wine is being distributed into all these different countries. That'd just be one client for you guys, and you guys wouldn't charge more for the expansion? - So like currently, it depends on the client, really. If we were to go commercialize more areas, we would expect them that they would want to believe in that, and where there would be a separate charge for that, because it also would guarantee like exclusive access for a year or whatever before, you know, off-roading to other people. And then the advantage of being first in the market with this is insane. That's one of the big selling points now that's been really funny is that your competitors are on it. Like if you don't want to use it, that's one thing. That's a negative to you to things you won't have, but now it's a double because your competitors are on it. So it's like, I don't know, just been kind of a funny thing, whatever. - A lot of interesting takes in understanding how summer is working and what it can do for companies. I do think I'm surprised some of those big hotel chains are not going to reach out to you guys. - Yeah, they want to talk. I'm happy to talk. Yeah, I'll build a short-- - I'm sure they want it done in there. - I get for like a small like one or two store, small location, but like any of those major kind of players, it seems like they could do some interesting stuff with that data, and it would be a drop of the bucket for them. - Yeah, I mean, send them my way, please. (laughing) - So we like to wrap up each episode out of personal dough, and so we're curious on what has been the most memorable winery visit that you've experienced and why? - I had a really, really cool visit this year. So I flew in for a PNV. I went to PNV one time, but it was totally different. I was helping broker some stuff as well. This was right at the beginning of the pandemic. - For those who don't know that, the premiering app of Alley. - Yeah, premiering app of Alley. So this was my first full tilt premiering app of Alley this year. And it was also, it was my birthday as well. And so I came in on my birthday February 17th, and went over to Arna Roberts and spent the day in Sonoma. And so it was like the whole full tilt, full circle kind of thing. So Arna Roberts was fantastic, we had pizzas there. Tasted it to like most of the lineup, love the wines. It was just a really, really fun day. And then right after that, we stopped by over at Ridge, the Geyserville, right? - Yep. - And so one of my original wines at first bought when I got in the wine was Geyserville. And so I was just like, this is fantastic. You know what I hadn't been there before, I was like, this is so great. And look at the vineyards right there. - Yeah, I prefer that location versus the Santa Cruz Mountains one. - Yeah, it was, it's so pretty. Yeah, I love Sonoma. And then we went to Devola for dinner, and it was just like it was an awesome day. And it was birthday, you know what I mean? So it can't be today. Yeah. - Yeah, pizza for lunch and dinner. - Yeah. (laughing) - Well, thanks Peter. - You're really good to see you. - Yeah. - Yeah. - Maybe that was part of what made it so memorable. - Right. - It was really, really fun. It was a good day. - Awesome. Well, thanks for showing so much about your business. And we hope our audience members get to learn about it. Hopefully, you know, this is a business to business podcast. Hopefully there is also some new customers for you. - I hope so. Thank you guys. All right, thank you for the time. This has been fun. - Hey listeners, if you love the show, support it by buying a show notes book. They not only compile two years of episodes, but also organizes them into themes for better learning. They can be an inspiration to listen to or relisten to an episode or provide a quick reference of the key learnings from a show. Go to xchatto.com and click on the store page for easy links to buy. Thanks for listening. Thanks for joining us. If you loved this episode of xchatto, we'd love for you to subscribe, rate and give a review on iTunes or wherever you get your podcast. Until next time, cheers.
Podcast Summary
Key Points:
Jeremy Hart, co-founder of SOM AI, has a background in wine service, wholesale, and importing, and created the platform to address inefficiencies in selling wine to restaurants.
SOM AI was initially conceived as an app for travelers but pivoted during the pandemic to help restaurants turn wine lists into retail shops, eventually commercializing as an AI-driven data platform.
The platform uses AI to tag and sort wine SKUs from on-premise menus (restaurants, bars, hotels), allowing users to filter by producer, region, price, and account type.
SOM AI covers about 100,000 accounts in the U.S., with additional coverage in Canada, Europe (Germany, UK, Austria, Switzerland), and Singapore, updating data every two weeks.
The tool helps suppliers identify where their wines are placed, analyze market share, and target accounts based on location (e.g., near stadiums or retail chains) for strategic sales.
Summary:
Jeremy Hart, co-founder of SOM AI, shares his journey from working in wine service and wholesale to creating an AI-powered platform that transforms how wine is sold to on-premise accounts. Inspired by his experience mentally cataloging accounts for wine placements, Hart partnered with a coder to build an app that originally targeted wealthy travelers but pivoted during the pandemic to help restaurants sell wine retail. , Canada, Europe, and Singapore, updating data every two weeks.
, steakhouse, sushi bar), enabling suppliers to identify market share, track competitor placements, and target accounts based on geography—such as near NBA venues or retail chains like Total Wine. Hart emphasizes the platform’s ability to provide “air superiority” by eliminating legwork and allowing sales teams to focus on direct, data-driven outreach. Long-term clients like Jackson Family Wines have used SOM AI to reinforce partnerships and boost placements.
Hart notes that the pandemic accelerated the need for digital tools, and SOM AI has become a leader in wine data and AI, helping an archaic industry adapt to modern sales strategies.
FAQs
SOM AI is a data and AI platform that helps wine suppliers sell to restaurants and other on-premise accounts. It leverages AI to track wine lists, analyze market data, and generate sales leads.
SOM AI targets about 100,000 on-premise accounts in the U.S. and resurveys them every two weeks to track changes. It also covers Canada and parts of Europe.
AI identifies specific wine SKUs from menus and tags them with attributes like region, producer, and style. This allows users to sort and analyze data, such as market share or competitor placements.
It eliminates the manual legwork of identifying sales opportunities by providing a list of leads tailored to a supplier's wines. This helps suppliers target accounts more effectively.
It started during the pandemic by turning restaurant wine lists into retail shops, generating about $700K in sales. It then pivoted to its current commercial platform for on-premise sales intelligence.
It covers restaurants, bars, hotels, and any place where wine is consumed on property. Accounts are also categorized by type, such as sushi bar, steakhouse, or cocktail bar.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.