Josh Sharkey: Yeah. You lose trust And that, and that, like, I'd wanna like, I don't know how long that lasts, but I, I, I think it should be just a sort of an, an evergreen thing that like, you let them know like, Hey, I'm an ai, my name is dah, dah.
Mm-hmm. Um, because even once they get to the place where it can be super conversational. At least, you [00:04:00] know, and then you end that conversation like, man, that was, that was awesome. And if you found that afterwards it wasn't ai, you're still gonna have that little bit of like, fuck, man. Like Yeah,
Alex Sambvani: exactly. No, you're, you're spot on.
Josh Sharkey: Um, well, you know, we didn't, we didn't talk about you, which obviously we, we, we will, but, um mm-hmm. Alex, is you the founder, uh, CEO of slang. Correct. Like, I think there's at least over a thousand, uh, restaurants that you guys help mm-hmm. Answer phones more effectively among other things. That's right. I didn't realize that, like it makes sense, like 60% of all calls to restaurants are mm-hmm.
For reservations, right? So this is crazy why slang is, uh, such a great service because Exactly. Um, it's such a low hanging fruit thing to solve for that I'm sure. Not only is it sort of numerical enough and so almost like, you know, binary enough where like you can just answer these things, but also the speed in which you can get back to somebody, I have to imagine is a huge part of, of the value.
Because if I call somebody from a reservation and. They have to get back to me, which happens basically every single time. [00:05:00] Mm-hmm. Um, and I wanna book a party or whatever a big, you know, usually you're calling a restaurant because you have like a more than a six stop. 'cause they don't always take those on reservations odd times, you know?
Yeah,
Alex Sambvani: that's right. I mean, we see it's definitely true that groups tend to call, um, not just like this large, large parties, but even slightly larger. Like we see the average party size over the phone is at somewhere between four and five covers. And online it's. What we see is, is typically between two and three covers.
Josh Sharkey: No, it makes total sense. And a lot of times it's actually, that's the rule when you go to book. Mm-hmm. It's like, oh, for parties are there seven or more? Mm-hmm. Call the restaurant because they need to plan for that. Right. So basically what that means is that these calls have a way higher ticket. Exactly.
An opportunity than the ones that are booking online. Exactly. And so when you can't get to the call, you know you're losing out. Right. It's a huge, I mean, look, I think it's a, it's an amazing business. Um, I love you guys are building, we can talk, we can talk more about it. Thank you. But let's, you know, I wanna maybe just step back for a minute, just because, you know, I think a lot of people [00:06:00] know about, uh, what slang does, but mm-hmm.
Your background is in data. Mm-hmm. And at Spotify, you know, there's obviously a ton of stuff you must have learned there. I'm just super like, excited to hear some of these things. So I'm just gonna begin to, to, and I don't know exactly everything that you did there, but Uhhuh, if this, so if this question is not something that you did, you can just be honest.
Okay. But did you learn, um, did you learn a lot about human behavior and psychology and the decision making process from your time, uh, at Spotify?
Alex Sambvani: Oh yeah, of course. I mean, there was so much stuff that we could learn from. The data that we had. And just like looking at when you listen to music, what types of music you listen to, where in the world you're listening to music, there's so many insights that we could glean from that.
And, and that was, it was really cool. So, I mean, just thinking back. To two. I didn't work on either of these things, but [00:07:00] um, I was kind of tangentially involved and helped leverage the insights from a couple of these research projects. But there's two really cool things that I'll share. One was we did a research project and we looked at, um, nostalgia.
So like, could we figure out if, if people had a nostalgia window, meaning like, is there a window of time in your life? Where, you know, music sticks a little bit more uhhuh, and we found that that actually is true. So we found that for most people between the ages of 16 and 21, those are really formative years.
And the music that you listen to during those years. Um, becomes music that you continue to revisit later in life. And so, um, the other thing that we found, well, why
Josh Sharkey: is that? What, what, what was the reasoning behind that? It's so crazy. 'cause I, I, it makes sense. Yeah. Most of what I love is things, you know, from.
You know, nineties hip hop from high school. Yeah.
Alex Sambvani: I don't know exactly. Like, I don't think we got to [00:08:00] the bottom of it. We just found the correlation, uhhuh. I mean, we could, we could guess, but, um, think it's like that when you think,
Josh Sharkey: it's like when your brain is developing, like,
Alex Sambvani: I don't know, it could be, it could be a brain thing.
It's like, if you think about those ages, it's, it's probably when people like start going out and social and exploring the world outside of home and, um, those might be just like, really, I. Formative experiences and music is associated with, you know, maybe it's like the first time you start going to a bar or you start going to parties or whatever.
You're in these environments for the first time when you're hearing music. Um, or maybe it's, you know, a lot of it's, those are teenage years when people are going through lots of emotions and maybe you're like, you know, making, uh, emotional connections with, um, with your favorite artists in life. Binge listening to certain albums.
Who knows? Like, I think it could be a lot of things. Um, but it, it's really cool to think about just like knowing that that correlation is there. I. The other thing that we found too was that if you look at folks later in [00:09:00] life, like forties and beyond, we found differences between men and women. So men 40 and above tend to listen to way more music from their nostalgia period.
So, um, you know, which. I don't know if this resonates with folks, but it might make sense, like maybe your dads or uncles like you, they like maybe play the oldies more, like pretty frequently versus women at 40 and above tend to mix it up a little bit more. They listen to more current music, more pop. Um, but on both sides, uh, the nostalgia window is pretty similar.
Yeah, it's so funny because it's
Josh Sharkey: so, like, it's so true and you think about it. Mm-hmm. Yeah. God, I, I wanna know why that is. Yeah, me too. Why do you think it is? I don't know. I don't know. I mean, like the, the gender breakdown, I mean, yeah. Like the, you know, like, I mean, I know, you know, I'm in my forties and Uhhuh, I, I get the most joy from listening to Outkast or really [00:10:00] old.
Right. You know, most deaf or, or, or, uh, and that was from when I was in high school.
Alex Sambvani: Yeah. I mean, my wife is like,
Josh Sharkey: you know, listening to music I've never heard of because it's, you know, bands from today for
Alex Sambvani: sure. I honestly, I have no idea why that is. I mean, I think it's, I don't know if it's like radio stations, people listening to, or just like social circles.
Yeah. Or
Josh Sharkey: what did you do? What did you do? So once you, from this insight that you gained Yeah. About nostalgia and then mm-hmm. People in their forties. What, what, what, how did you. Build product around that at Spotify experience. Yeah. So
Alex Sambvani: with with that insight in particular, we ended up building, um, a playlist, a new playlist called Time Capsule.
Josh Sharkey (2): I
Alex Sambvani: heard that, that that was based based on your age. Yeah. So, and based on your listening history. So we'd go back and try and pull songs that are within that nostalgia window that you're likely to wanna listen to again. And so that, that [00:11:00] was a really fun one.
Josh Sharkey: I, I totally remember the time capsule thing.
Yeah, that was a good one. I feel like Spotify has to figure out, so, and maybe this is just because we use the same account, like my mm-hmm. You know, we have all the Alexas over my house, right. And so my kids, they know how to use Alexa. Now they're like four or six. Mm-hmm. And they're like, play, so they're always playing like, you know, these.
Kid songs. Mm-hmm. And, um, and my wife will be playing, you know, whenever she's playing. And so then I, I get my Spotify wrapped at the end of the year and like mm-hmm. You are most listened to was Beyonce and Yeah. Coco Melon music. And I'm like, what the fuck? And then I start getting music fed to me because of that.
Mm-hmm.
Alex Sambvani: Coco Melon. Yeah. You get away from it. It's so legendary. I mean, I used to, I don't remem I don't know how it works anymore today. Uh, or I don't know how it works today, but I, when we, when I was there, I. We would try and strip out kids music and try and strip out Christmas music from the rap. Um, but it's hard.
I think it's hard to do it like super
Josh Sharkey: accurately. What was the [00:12:00] genesis of, I mean, it, of rap, it was such a genius idea. We, we do something like that now and at media. Uh, obviously we just stole the idea. I think a lot of people stole the idea. Oh yeah. It's funny, it's, there are rarely times when something happens that create an evergreen new, uh, you know, trend.
I don't, I don't even wanna call it trend because it just, now it's just a thing every year everybody puts out their rapt of the data of their platform. Spotify, unless I'm wrong, came up with it.
Alex Sambvani: I think they did. And um, so I joined Spotify in 2016. I first started working there in 2016, um, and the rap campaign was, it was already a thing, so I'm not sure exactly how it got started, but, um, I did work on that campaign one year in 2017 and it was a ton of fun.
It was cr a crazy experience and I learned a lot about. How it works and, and how to do that type of, you know, marketing at scale. Um, like datadriven, um, insights. What was the impact that it had on [00:13:00] revenue for Spotify? Great. I mean, it was the biggest campaign that we would do every year. Um, and I'm, I'm sure it, it's still, uh, I'm not sure, but I would guess that it still is this the biggest campaign that they do every year?
Big numbers. I mean, millions of subscribers. I don't know, I can't remember the exact, but it, it was definitely successful every single year.
Josh Sharkey: Yeah. Mm-hmm. Okay. Let's go back. We'll get, we'll get back to slang. I just, uh, yeah, for
Alex Sambvani: sure.
Josh Sharkey: Here's a question for you, 'cause I heard you, I heard you talking about it.
What is. Conversational design. There's a woman named Rebecca. Rebecca Hoe. Yeah. That works for you. Mm-hmm. Correct. And, um, I heard her job and I'm like, what the fuck is that? Makes sense for what you do. Yeah. But maybe tell us like, what is conversational design and then how do you all use it, um, in slang to make what, what you make?
Alex Sambvani: Yeah. It's cra it's, it's, it's a niche term, but it's, it's growing. So conversation design, you know, just like there's visual design, like a, like web design or design and designing the visual experience of a, of a mobile app. You know those, those designers are thinking [00:14:00] about. How it looks. Is it intuitive? Um, do all the buttons make sense?
Are the words that we're using in every area mm-hmm. Like, are, do they make sense? Are they self-explanatory? So like the experience and the, not just the visual aspect of it, but the actual experience side of it and making sure that it's intuitive and pleasant. Pleasant and you know, in some cases even delightful.
Like the designers are the ones thinking through that. Yeah. And so when we're thinking about AI conversation experiences. We think about them as an, we think about conversation as an interface, just like we think about a website as an interface. And so that conversation needs to be designed like someone needs to think through all of the words and the pathways and what happens if something goes wrong, how do you get someone back on track?
You know? So we actually design out all of our conversations and. Think through, you know, what is the optimal, most intuitive way to say this? How do we make it [00:15:00] super? How do we make it friendly? How do we make it warm? Like what, or whatever the goal is. Like there's a lot of thought to like the actual words that are being spoken and then also the structure of the conversation.
Like what? What can the system handle? What can it not, what are all of the different paths that, that of conversation that you can go down? Do you want it to tell jokes or not? You know what I mean? Like there's so many of these decisions that you have to make that are actually design decisions. Yeah. Um, this is a design domain that has existed for a few decades, but it's been very niche.
Mm-hmm. Um, and now with conversational AI becoming so much more common, it's, it's growing as a, as a, um, as a field. But even today, there are very few people in the world that are experts in this stuff like. You know, definitely less than 500 people in the world, maybe even less than 200 people in the world that are truly good at this stuff.
Josh Sharkey: What are the skills that you need in [00:16:00] order to become a conversational design expert?
Alex Sambvani: So you need to be, I mean, it depends. Some of these people, they come from like linguistic backgrounds. Maybe they started out as writers. Maybe they, maybe they studied linguistics in school. Some of them come from visual design backgrounds because really what you need is you, well, you need to have strong command of whatever of the language.
Uh, and you need to have a strong command of like, and, and a strong attention to detail of language and like, 'cause, like even just like little words make a big difference. Um, and then you, you need to think about it from, um, the user experience perspective. So you have to be, you have to understand. A lot of the standard like design principles that you would apply to visual design Makes sense With conversation design too, which is just, yeah, like is this thing usable?
Is this thing, um, delightful. Like what's the emotional experience? [00:17:00] Um, like there's so many, like, there, there are a lot of parallels, but a lot of it is also, you know, net new stuff. Yeah. Um, yeah. And there there's like lots of little things too, like, you know. Outside of the, um, like the words and the pathways, there's like other design design decisions that really matter.
So like if you're talking to an ai, um, can you interrupt the AI or not? Like do you let the AI finish its sentence before the user can talk back, or do you want the user to be able to talk over the AI interrupted? And that makes it even more complicated.
Josh Sharkey: I find that, that for some reason with most. Uh, like, uh, chat interface language models.
You can't mm-hmm. Like you have to, yeah. You have to press a stop button or something. I don't know why that is.
Alex Sambvani: Yeah, it's because if you could interrupt it, um, it's just, I. More likely to fail so you can, uh, it might be more, it's, it's, it might be more frustrating sometimes for folks to not be [00:18:00] able to interrupt it, but when you do make it super easy to interrupt, it's just like easier for the system to go off the rails.
And so I think that's why most folks opt not to. Yeah. Is a solve for that. There's several ways you can solve it. Yeah. I mean, and, and the technical term in conversation design is called barge. Like you're barging into the conversation and there's lots of different ways you, you can solve for it. I mean, in voice, the main, like one of the main, um, risks with allowing the user to interrupt the system is, uh, like background noise.
So like if you're out, if you're at a bar and you're like. Trying to make a reservation using our system. Like we're gonna pick up a lot of background noise. And so the AI might think that a background noise is you trying to talk to it. Mm-hmm. And so it's gonna like keep pausing to let you speak if it hear, picks up that background noise.
So you can go really crazy with like noise cancellation and sentence speaker isolation and like all that stuff. [00:19:00] Um, so there's definitely technical ways to solve it, but it's still, it's still a challenging problem.
Josh Sharkey: Yeah. Yeah. Mm-hmm. You know, I remember listening to Mark Zuckerberg talk about the, the meta glasses.
Mm-hmm. And sort of how to make it feel more real when, when, mm-hmm. It was virtual reality and the, the notion that the hands were really important. Arms actually didn't matter so much, and so like, they didn't actually design arms, but the hands and the fingers were really important to see. Right. Is there any, is there any sort of, uh, anything analogous there with voice in terms of like, things that you saw interesting insights about?
About how the, how there's authenticity with the, with the sound or the feel of the voice, like what makes something feel authentic when you're, you know, when you're speaking.
Alex Sambvani: Yeah. It, it is a great question. So I think for voice, the, in my opinion, the most important thing is the inflection of the voice and the cadence.
Like, does it sound natural? Does it sound like I'm talking to [00:20:00] a. Real human, are they en enunciating the words the way that a human would, and if that is not there, it becomes very distracting and it feels robotic. You know? I think like a lot of people get caught up in like the, does the voice sound like a robot or does it sound like a human?
Really, I think the most important thing there is like, is the inflection and cadence. Mm-hmm. Like natural then. So that's the most important thing to me.
Josh Sharkey: So just not having a monotone
Alex Sambvani: speech?
Josh Sharkey: Yeah.
Alex Sambvani: Like just pronouncing words norm, like, like a native speaker of English. And, um, that, you know what,
Josh Sharkey: that's, that is such a hard, I live in a town called Croton on Hudson Uhhuh.
Um, and our, uh, we've tried to, uh, uh, fix like correct Alexa a hundred times. Yeah. Every time she says croton on Hudson. Yes. And that sounds like, I guess that's just a difficult challenge to, to solve for.
Alex Sambvani: Yeah. Uh, yeah. It's, it's difficult. Um. But we're definitely getting there now where it's [00:21:00] like, we're in like good shape and, and, and as technology continues to evolve, like I don't think that I, I, I think we're, we're just, we're within like a year or two years of these systems being completely indistinguishable.
Yeah. Um, from humans. And then, then there's a actual like, uh, like sonic quality of the voice. So what does that mean? We're talking about like the, the way that it actually sounds, the, you know. Um, to the ear, uh, like, does it sound like a robot, like tonally or does it sound like a human? Right? And I think that's actually separate from the inflection, right?
Like, you could think of like a, a cartoon voice, right? Like a cartoon voice could sound natural, but you know, it's a cartoon, right? Mm-hmm. So, so I think the most important thing is that the cadence and the inflection sounds natural. Yeah. But then from there, you can start to think about. The sonic side of things.
So like how do, what does the voice itself actually sound like? Um, is it [00:22:00] breathy, is it not? Does it sound like a news anchor? Is it or not? Like, or a regular person? Is it like high energy, low energy?
Josh Sharkey: Um, what, what makes, what are some of the things that make a human voice sound, human versus robotic?
Alex Sambvani: That's a hard one.
I think it's kind of, that's kind of like an intangible quality. Let's, it's like just, does it sound like. A, does it sound, I, I think you, I'm sure that there's like an audio footprint you could analyze and like you could, there's like a scientific element to it, but I think it's, it's more of like a gut intuitive feeling that we have, which is like to the ear, does it sound like a human is talking or does it sound like a non-human?
You know what I mean? Yeah,
Josh Sharkey (2): yeah.
Alex Sambvani: Um. So think about like, I don't know, like in movies there's like alien voices, right? Like you, it's like clear. Like you could probably, like if you heard a clip of, in like an alien speaking English from a movie, like you'd know it's an alien. You know what I mean? Um, versus a human talking.
Yeah,
Josh Sharkey: yeah. I just, mm-hmm. I'm curious what those, someone much [00:23:00] smarter than me must know. Like, what, what are those qualities? Because I imagine you have to program the thing to, with those specific qualities in order to Yeah. To have it sound more and more like a human. Yeah. I have to imagine there's something with the.
With, um, the esophagus and the way this sort of mm-hmm. Things vibrate in our throat before they come over our mouth. That make that Totally. And then just are, you know, however we're thinking be as we're, as we're speaking, but mm-hmm. It seems like a really hard problem to solve. It does sound like, you know, I've seen some ai that sounds pretty damn real, right?
Like, there must, there must be something there unless they're actually just recording. And, um, emulating the exact voice of somebody else.
Alex Sambvani: Well, that's how many of these models are trained? They're actually trained on actual human audio, like actual spoken word audio. And so that is how like the many of these foundational models are built.
Josh Sharkey: Yeah. Yeah. Mm-hmm. Well, I mean. There's obviously so many details of how this works, but what do you think are like the biggest misconceptions that dumb people like me [00:24:00] have about, uh, voice ai and, and what we, what you perceive it to be versus what it actually is?
Alex Sambvani: Hmm. That's a good one. And I bet definitely beg to differ that you're, you're a dumb person, but I, I doubt it.
Um, but, um, in terms of, yeah, let's talk about that. So what are, what's the common misconception? I think. So when people think of an automated voice experience, the first thing that most folks think about is like that horrible call center bot that you get when you call an airline, right? Or a bank or whatever.
And a lot of folks think that that is still where we're at in terms of technology, but we're in a completely new wave where AI can actually be. Effective. It can be good, it can be fast, it can be fun even, and the technology just keeps getting better and better every day. So I think this, I think the, [00:25:00] the, the, the common misconception is that, um, AI is often a, a negative experience, right?
If it's an automated experience. I think that's a misconception because we're, we're, you know, at least at slang we're proving that that's not true. You know, every single day we have, we work with, as you mentioned, Josh, over a thousand restaurants across US and Canada, and we have extremely high guest satisfaction, uh, which means that the guests that are interacting with our AI are actually finding it useful, helpful, and we hope even delightful, like we have some guests.
Say that they're surprised or they thank the ai. They're like, oh my God, thank you so much. You were amazing. And like, we didn't expect that to happen. Actually. We didn't think, you know, we, we love seeing that, but that actually wasn't our expectation. But, uh, what we found is that if one, the technology's just gotten better.
Mm-hmm. Um, and two. People are starting to trust the technology more and finding it and, and using it successfully. [00:26:00] Um, and then three, if you design it well, that's why this, that's why conversation design is important. And that's why, you know, being thoughtful about what you're having the AI do and, and being thoughtful about what you're even, maybe even more importantly, you're being thoughtful about what you don't wanna have the AI do.
If you're thoughtful about those things, you actually can deliver something that is, that is a good experience. And um, I would say right now, you know, we can deliver good, very good experiences and in the future, as AI gets better and better, AI's gonna be like blowing our socks off. It's going to be doing things that we would never expect humans to do, right?
Like having a super deep understanding of who you are as a person, your history with the restaurant. Delivering a level of personalization and convenience that we've never even seen before. And so I, I, I'm super excited about that. And I think AI and, and, and many and, and the whole, you [00:27:00] know, technology landscape in the hospitality space that's, you know, continuing to evolve.
I think all these tools are gonna help operators do what they're, what they love to do, which is deliver amazing service in person and leave a lot of this, you know. A lot of these like customer communication, guest communications tasks that frankly, many restaurants struggle to do well. Um, I think AI's going to, going to help and, and do a lot of it.
And it's gonna be win-win. It's gonna be win win for the guests and win for the restaurant.
Josh Sharkey: This show is brought to you by, you guessed it, MES Mees helps the thousands of restaurants and food service businesses all of the world build profitable menus and scale their business successfully. If you're looking to organize your recipe IP and train your team to put out consistent product every day.
In less time than ever before. Then me is just for you. And you can transform all those old Google docs and Word docs and PDFs and spreadsheets and Google sheets [00:28:00] into dynamic actionable recipes and mees in lightning speed plus stop all that manual work of processing invoices. 'cause me will digitize all your purchases automatically.
And there's a built in database of. For every ingredient, which means you're gonna get laser accurate food costs in a fraction of the time. Visit www.getmeez.com, that's G-E-T-M-E-E z.com to learn more and check out the show notes moving forward, because we're gonna be adding promotions and discount codes so that all of you lovely and brilliant me podcast listeners, get a sweet deal on me.
A hundred percent. Mm-hmm. I mean, and I think the most important thing is just building trust and what, what, I'm not saying this just 'cause you're on the show, but, but what, mm-hmm. What slang is. And companies like slang are doing are a big part of what, uh, we need to get to the next level, right? Mm-hmm. Um, most important thing is to solve the problem.
If you're solving a problem for somebody, they're [00:29:00] gonna trust it, right? If like, totally, Hey, I need a reservation and I couldn't get it on the computer, I called and this thing helped me get a reservation, then the customers gonna be like, well, that fucking worked. That's great. And the restaurant's gonna be like, well, hell yeah, that worked.
So who cares whether it was AI or my uncle, you know, like Right. The, the problem got solved. Exactly. And, and that becomes more normalized. Over time. And the more we can do that, the, you know, the better. I, I am curious. I want, I mean, I wanna talk about like, sort of some of the future of ai, but, but, um, this, there might not be any of these, but is there any like crazy way in which people have interacted with slang where you're like, where did that conversation come from?
And then you're either the ai like totally bonked out or, or just really un unexpected conversations.
Alex Sambvani: We've seen, we've seen some guests like realize that it's an ai and like. Try and flirt with it. We've seen, which is funny, nothing too explicit. We've seen some, we've seen folks try and like ask, like joke around with it.
We, and as I mentioned, we also see folks like try and, uh, [00:30:00] sorry. Hey,
Josh Sharkey: that's all right.
Alex Sambvani: I've got this. Uh. Just got a puppy.
Josh Sharkey: Oh, nice. What kind of dog is that?
Alex Sambvani: She here, she's a bernedoodle. Oh my goodness. And she's like, she's been sitting here and now as a, she could be part of the show. Yeah. She's getting a little lazy.
She's not ai, right? She's real, she's not ai. Some people say she looks like AI though. Um, in her photos. So I, we've we've seen some crazy stuff. Yeah. It, it, and it's really, it's really fun to, to see that stuff. But I think overall what we're seeing is that, you know, folks, folks will say things like, oh, is this chat BT?
Like what is this? You know? Um, 'cause guests are just, they're getting familiar with ai. Yeah. And they're, they're trusting it more and, and giving it a try, which we love to see.
Josh Sharkey: Yeah. Yeah. Okay, so now I I, I just will ask, and this is probably everybody's asking this all the time, but you are working in AI all the time.
You're thinking about this all the time. Uh, and maybe we can, we can either just stick to, uh, the vertical of voice ai. Mm-hmm. But feel free to expound on it. What, in 10 years, what does this look like? What does slang look like in 10 [00:31:00] years with the, as a, as the technology curve, you know, keeps getting deeper and deeper?
Like, what, what does 10 years from now look like?
Alex Sambvani: Yeah. Um, I think voice AI is definitely going to continue to grow as a category. Uh, there's so many companies working on this problem and lots of great stuff happening. Like, you know, in, in many verticals, healthcare, financial services, even outside of hospitality, that, that's really exciting.
I think the technology is gonna continue to get more and more human-like, I think we're going to, um, do, we're going to reach levels of personalization that we haven't seen before. And so what I mean by that is like, let's say you call a restaurant to make a reservation. Today, if a host even answers, um, they're, you know, you're gonna be treated somewhat generically.
They'll, they'll book a reservation for you. Um, but it's, it's not gonna be like a tailored experience. Like you, you know, you don't have a relationship with that, [00:32:00] with that host. They don't, they don't know you. Um, but in the future we, we imagine AI becoming like a, almost like, um, it's like you're talking to the same person every time, right?
Yeah. And it, it knows you and gets to know you better. So when you reach out to make a reservation, it already knows what table type you like. It already knows, you know what time you like to eat dinner. Are you a, are you a six 30 person or are you an eight 30 person? Um, it probably, it knows when your birthday is.
It knows if you have a spouse. It knows like all these things about you. And it can bring that information into the conversation in a way that. Makes you feel seen and heard, but is, but you know, ultimately makes that interaction more convenient and faster for you, which people obviously love everyone, you know, times our most precious resource.
So I think personalization's gonna be a thing. And then I also think as the technology gets better, we're gonna, and this is where we're going as a business at slang, is, you know, we're just, we're interested in just having the AI take on more and [00:33:00] more and more of the guest lifecycle for you as a, as a business.
So. How do we, which, which ultimately delivers more value to operators and guests because, you know, there's so many things that today are broken. Just think about the experience of private dining, right? I booked a private dining reservation, like a, like a 10 person thing for my birthday, a few, a few months ago.
I live, live in New York, right? So, uh, reached out to six restaurants. Um, only two of them emailed me back. And the two that emailed me back, it took me like three days of back and forth to actually figure out like what menus they have and like. Pick one. Yeah. And like make a decision. Like, I was like this, there's gotta be a better way.
Right. So like, I think a, there's lots of ways where AI can speed up that workflow, that back and forth. Um, and I think that there's a lot, there's so many use cases like that. So that's what I mean when I talk about like, getting into more areas of the lifecycle outside of reservations where we [00:34:00] can just like help create efficiencies, um, and, and find win-win opportunities for operators and for guests.
Josh Sharkey: Yeah. No, that makes sense. Is there a time when this AI gets just completely agentic, uh, for, for slang? Ooh.
Alex Sambvani: You think about the difference between agentic AI and like more traditional, let's call it conversational ai. I think about the distinction being agentic AI as something that can actually, um, achieve a task or a workflow for you from end to end.
Yeah. Right? Yeah. So if it needs to log in to other systems and Yep. You know, do API connections or whatever to like actually take care of something for you from end to end. That to me is kind of like one of the key distinctions between agent AI and other types of ai. And certainly, yes, I would say we're, our product is already there.
We handle many things from end to end. Um, there's, there are some things that um, I [00:35:00] think will, over time we're just gonna be doing more and more of age genetically. So, for example, you know, I already mentioned private dining, but like, let's talk about like catering, right? So a catering order, um, is more complicated than a reservation, right?
There's so many, like questions and things that need to be sorted out. You might need to, um, you know, there might need to be several. There might need to, there might need, um, there are several subtasks, for example. So let's say like there's a, let's say. Your restaurant and you wanna facilitate a catering order.
It's not just about like making the catering order, if, if the catering job required like someone to go, um, deliver the food and like set it up, right? Like I. One, just one example could be like, okay, well, like they're going to an office building, they need to like, get security, they need Q clearance right?
To like go upstairs and like deliver the food. Right? That's maybe something that an agent [00:36:00] could do for you, right? Like you could have the agent like coordinate with the client and or like reach out to the building security or whatever. Like, um, so in that instance, the agent is kind of like the connective tissue between the restaurant, the guests, and like the operational tasks that that need to get done.
So I definitely think like that's the future.
Josh Sharkey: Yeah. Yeah. Are you guys using any sort of MCP just to connect with, uh, the customers, uh, other systems today or,
Alex Sambvani: yeah. So right now we do, uh, we do integrations and, uh, through, through partnerships, uh, with, you know, with, with the leading folks in the restaurant space, and we're always adding more.
So, um, our approach has been to, you know, partner directly with, you know, the best in class vendors, um, in, in each category and, and be able to. Kind of, yeah. Yeah. Bring be the connective tissue between voice and these different systems.
Josh Sharkey (2): Mm-hmm. Yeah.
Alex Sambvani: Yeah. So one of those examples is our, our OpenTable integration.
So we launched an integration with [00:37:00] OpenTable in September. I. We learned recently that we've actually become their top integration that they've launched over the past 12 months, which we're really excited about. Wow. That's awesome, man. Yeah. Super excited about that. And so
Josh Sharkey: basically you, the customer, uh, when they're talking to a slang AI rep mm-hmm.
I'll call them, they can actually book, um, a reservation via open table from the, from the call. Yeah, that's right. That's awesome.
Alex Sambvani: So, so if a guest calls a restaurant that's on open table. The AI can pull the availability, it can book a reservation. Mm-hmm. It can modify or cancel a reservation and it can even add special notes.
So if the guest is celebrating something or if they have dietary restrictions and they want the restaurant to know about that so they can have a better experience when they're there in the four walls of the restaurant, um, the AI can can handle that. So it's performing extremely well. We're booking like anywhere between 50 and 500.
Reservations per month, per location. Wow. Um, and we've already crazy. I know we've already booked since September. We've booked like [00:38:00] 250,000 reservations for, for our restaurant partners. So, dude, dude, that's nuts. Super.
Josh Sharkey: Congratulations. That's, I, I mean, I remember you talking about this integration mm-hmm.
Or the partnership, um, a while ago, and, um, seeing it sort of come to fruition is pretty amazing. Yeah. We're pumped. This is a tough question, like did, did you have to pick a lane? I think we might have talked about this in terms of like open table Rey talk.
Alex Sambvani: Yeah, that's a, that's a, it's definitely like I use, is it
Josh Sharkey: exclusive or
Alex Sambvani: Our partnership with OpenTable, uh, is, is not exclusive.
Um, but we're super, we're in super deep partnership with them. They're, they've been amazing to work with and, you know, obviously there's many players in the market, but OpenTable is, is the, the og, they're the leaders and they're, um, and what they've done on the product side just over the past few years has been.
Very impressive.
Josh Sharkey: Um, and I feel like they've, they've made a big comeback, like mm-hmm. As a brand as well, uh, with new leadership. Yeah. Just doing a really good [00:39:00] job of, of, um, be becoming part of the community again.
Alex Sambvani: For sure. I mean, there was just an article that came out a few weeks ago, maybe it was in February, actually, I.
Around. So it was more than a few weeks ago. Uh, but it was a, I think it was New York magazine. They were highlighting how OpenTable just won back. Don Angie from rei and it Oh, yeah. That super big deal. Yeah.
Josh Sharkey: Yeah. That's right. Mm-hmm. I saw that. Yeah. Yeah. So I, you know, I'm curious, um, and maybe you're not using it all, but, but, um, are there ways that you're using ai, uh, for yourself today?
Independent of, oh my God, using chat Bt to ask question?
Alex Sambvani: Well, yes. I mean, I. We use AI for a lot of things internally. I mean, yeah, we use AI on, on the sales side to like analyze calls and, and do account research and all sorts of stuff. But I mean, in my personal workflow I am, you know, I'm just obsessed with Chat gt.
I use it like 50 times a day at least. [00:40:00] Um, my partner jokes that I'm like in a open relationship with chat GPT, um, because like I'm using it so much, um, and it has so much of my time and attention, but I, I just love it. And, um, the, I mean, for me, what I've been using it for recently is, um, I use it for email, obviously.
I use it for. You know, researching things. You mean just like refining emails
Josh Sharkey: that you're gonna write or?
Alex Sambvani: Yeah. What I'll do is like, I'll actually just like talk to it. So I use the, the mobile app. Mm-hmm. They also have a good iPad app and desktop app, and I'll just talk to it. I'll be like, Hey, draft this email for me, and I'll just like dump stream of conscious to it and it'll structure it and it's like learned my style of writing.
Yeah. So like, I don't have to edit it at all. Um, so that saves me a ton of time. I'll use it for, I. I love use, so I, I actually started paying for the chat GPT Pro subscription, and I, I use deep research for a [00:41:00] lot of stuff. Yeah, me too. Yeah. So useful. Um, I mean, you could do it for anything, you know, I, you know, I haven't seen this person in a while and I wanna like, get them a gift.
Like what do I like, come up with some ideas or I'm about to meet with this. Brands, like, tell me their story. Uh, I am, I don't know, like even for personal stuff, like I'm gonna Mexico City, like, help me create an itinerary. What restaurants do I need to hit? Like, it's such, it's so good at those things. It's crazy.
It is absolutely crazy. And, um, it's been a lot of fun to, to kind of push the boundaries of what it can do.
Josh Sharkey: Yeah. I, I, I, I totally agree, man. I'm using it for. So much. I, I mean, I have chat, GPT, GR, and Gemini kinda like use 'em for, for different things. But I've also been like using a lot of these cps. Mm-hmm.
Like I use this thing called, this tool called Lindy. Okay, cool. I haven't mentioned that yet. And you know, I just create agents and you can give all these skills and it's all, so there are, there's an agent that I [00:42:00] have that, um, has my calendar plus um, uh, a database that I have of basically every contact that I have.
I, all of my contacts, their birthdays, their anniversaries, if my employees, when's their work anniversary, when's their birthday? And um, you know, it does everything from any meeting that I have. It's there at the meeting. Um, other than these, 'cause it's, I feel like it's face of and, um. It, it creates action items for me.
It creates reminders. Mm. It updates my coda with, um, with tasks that I need. It will set new, um, calendar invite. It's ent. So will Clint also just like, um, start to like, um, ask people to, to set other meetings? Whoa. And then it reminds me every week, um, hey, uh, tomorrow is Alex's birthday. Here's some information on Alex.
The last time you guys met this, do you wanna send a gift card? And then I also have connected to these, I don't wanna get too much in the weeds, but I can have it send a gift card for me. I can have it, I can have it, uh, send, I can tell it what I wanna say. Um, and so I'm getting [00:43:00] reminders. Hey, next week is your mom's, uh, wedding anniversary.
I. And so I get an email every week with like, here's the upcoming like, important dates. Um, yeah. And then I can kind of like text it things and say, Hey, remind me, like, like add this important date. Um, but it does the same thing with, uh, I mean, I have so many of them, so many of these workflows, um, that are, dude, I need
Alex Sambvani: to check this out.
Looks so cool. Yeah. I mean, it
Josh Sharkey: does all my email triage too, by the way. Like, I, I never touch my email anymore because I just programmed it and I said, you know, if it comes from. Here's my investors, my internal investors. If it comes from one of them, put it here. And if it requires a response, draft one, um, you know, ask for approval.
Um, if it comes from my family, it's here. It's from school, from here, and, and it does it all for me. Um, so there's only like four emails a week that I ever even usually need to address. And it's because it's also like an agent. If I bring it into an email for calendar, like I'll just be like, Hey Lindy, can you find out some time?
And it knows my calendar. I gave it parameters of. Of, um, you know, what it can override and it just [00:44:00] does it all for me. Um, and there's so many things like this. It's just crazy. The amount of, I connected this tool called Repli, which is, um, so you can build your own applications. I mean, it's wild dude. It is wild what you can do now, and there's the amount of like productivity you can have, um, with one person is just insane.
Mm-hmm. And, um, uh, every week I get like a little bit more excited, like, I believe you know how much.
Alex Sambvani: I love it. I mean, the techno, it just keeps getting better and better, it seems like every day. So. It's, it's a wild, it's wild times we're living in, but I'm gonna check Lindy out. Uh, that Oh, you should, yeah.
That email triage use case seems super valuable. Oh, it's, it's dope. And
Josh Sharkey: you can, you can connect to like every, almost any application you can think of. Um, wow. Zapier has a similar, um, MCP. And, um, and obviously has way more integrations, but you, it's a little bit more like you have to use cursor, um mm-hmm.
To, to, to build it all, to have a little bit of an understanding of, of, of, of code to, to use that. Mm-hmm. But Lindy is a, is a, is way more [00:45:00] like no code. Yeah. Like no code, just explain, which, you have to be very explicit. That's the thing I learned with any sort of AI that mm-hmm. I find super interesting in terms of.
It's funny, we, we deal with recipes, and recipes can be implicit or explicit. Mm-hmm. And it's very, uh, it's very easy for someone to screw something up. And most of the time it's your fault because you didn't explain well enough. Right. And these, you know, these, these AI agents, um, or these prompts that you want to give it, um, you realize how clear you have to be.
And if you're not clear, uh, I mean obviously it's, I love that it's iterative. So when you're not clear and something messes up, you are like, oh, got it. I gotta be more clear about this. I have to clarify this, I have to add this exception. Mm-hmm. And over time, it gets better. But, uh, the importance of, um, how clear you have to be with instructions.
Mm-hmm. And we think about like our assistant or a team member, and then you give them an instruction that's not super clear and expect result X and you get y And it, it's actually helped me realize like, oh, I need to be way better at the instructions that I give my team, [00:46:00] um, because. You know, it's the same as when you think about what you, you know, what you're gonna explain to an ai.
You have to give it context. You have to give the reasons why you have to, you know? Mm-hmm. Help it understand what skill sets you want. So it's pretty cool.
Alex Sambvani: Yes. Uh, that I, I'm gonna nerd out on it this weekend. I already, I can already tell. Thanks for the tip. What are you most excited about next for, for slang?
I mean, this, we're out of fun, um, fun time in our business where. We've kind of got critical, you know, scale. We've, we've got over a thousand restaurants are, you know, we're speeding up. We're, I mean, through that. We now have, like, you know, I'm excited to leverage all the data we have and the scale we have to partner even closer, more closely with our restaurant clients and become even more of a, of, of a consultant partner to them.
Like using all the data we have, right? Because we know, you know, we're pro we're probably working with someone else in your [00:47:00] neighborhood now. We, you know, we can bring you benchmark data, we can bring you best practices. We're already starting to do this stuff. But, um, I think that's, that's a really exciting evolution of our business now because we, you know, we're still in, you know, the early days still feels like day zero in many ways, but we do have a lot of data and a lot of relationships that just make.
Our value that we bring to the table, even even bigger. And so that's kind of, that's been top of my mind recently, so I'm excited about that.
Josh Sharkey: That's awesome. Well, congrats man. Um, it's amazing what you built and obviously you're just getting started and, uh, it's been cool to see from the sidelines what you guys are Thank you are doing.
Alex Sambvani: Thank you. Thank you. Thanks for being a champion and, and friend throughout, uh, throughout it. So, uh, I appreciate you having me on and um, and this
Josh Sharkey: was fun. Absolutely, man. Thanks for tuning. The music from the show is a remix of the Song Art Mirror by an old friend, hip hop artist, fresh Daily. For show notes and more, visit get [00:48:00] mes.com/podcast.
That's GETM eble z.com/podcast. If you enjoyed the show, I'd love it if you can share it with the fellow entrepreneurs and culinary pros and give us a five star rating wherever you listen to your podcast. Keep innovating, don't settle. Make today a little bit better than yesterday. And remember, it's impossible for us to learn what we think we already know.
See you next time.
.png){kind=icon}
