48:42
Owner: Jeremy Julian
SUMMARY KEYWORDS
ai, restaurant, order, technology, support, customer, merchants, call, phone, ben, experience, point, voice, business, people, consumer, sale, revenue, automate, interaction
SPEAKERS
Ben (58%), Jeremy (42%), Intro (1%)
I
Intro
0:02
This is the restaurant technology guys podcast, helping you run your restaurant better
JJ
Jeremy Julian
0:13
Welcome back to the restaurant technology guys podcast. I thank you guys for joining us out on the airwaves. It’s always fun to come to you guys with new topics, new new guests on the show and today is no exception I, we have not had this company or this guy on the show before. But it’s always awesome to introduce you guys to to new guests and new products. And so today we are joined by Ben who’s the VP of sales at Soundhound. But really the restaurant division of Soundhound. Ben, why don’t you tell our audience a little bit about yourself. And then we can talk a little bit about what Soundhound is and dig into what you guys are trying to do to to make the restaurant world a better place.
BB
Ben Bellettini
0:52
Sounds like a plan. Thanks, Jeremy. Yeah, my name is Ben Bella Teenie. I’m the VP of sales at Soundhound. For our restaurants based products. I’ve been in some sort of restaurant technology sales role, either an IC role or leadership role for the better part of 14 years. And so very excited to have joined Soundhound, approximately a year back to lead our go to market strategy. As we really start to tackle the restaurant industry. You kind of alluded to, you know where what Soundhound is, and where we come from Soundhound is a voice AI technology company. So we have our own in house tech stack that focuses on housing voice AI for IoT devices. We actually started out as a music recognition app back in 2004. And we’ve migrated into supporting a number of different industries. So notably, where you’ll find Soundhound technology deployed is in every Mercedes Benz that gets produced today, when you say roll the windows down or Mercedes, what are the directions to the nearest gas station, that’s our technology that you’re interacting with? Actually, yeah, we’re with a number of different automakers embedded in vehicle today, using cloud technology. And then on the flip side, we’re also in every Vizio TV that gets produced today. So taking a Netflix or XYZ command that’s our technology or interacting with. So we have all of these deployments, all these devices, we saw keen opportunity to start entering the restaurant space, probably 18 months back, and AI is we believe soon to sweep if not already, sweeping that the restaurant industry.
JJ
Jeremy Julian
2:40
Yeah, no. And I know we’ll get into the use cases in just a little bit. I had no idea that Soundhound was that’s where you guys came from, might I I know of voice AI and kind of kind of where that’s at. Why don’t you give a little bit more about your kind of background? Where else have you been restaurant tech? Because, you know, you’ve been 14 years? It’s I’m sure you’ve seen lots of different technology changes in the last 14 years. You know, I can look up your, your profile and and all of that. But why don’t you talk a little bit about that? Because I think it’ll, it will help with some of the tech pieces that we’re going to talk about on where soundhouse At, and how how you guys have have kind of continued to migrate this product towards restaurants.
BB
Ben Bellettini
3:19
Very good. Yeah. So I started out my career in an ice sales role as one of the first 30 salespeople over at Yelp. Many, many moons ago, selling CPM advertising to restaurants and so really got to learn the restaurant industry there and moved into a leadership role over at Yelp was with the company for about seven years before going to Postmates to effectively build a massive sales organization from scratch. So this was 20 15/3 Party was still kind of having an HIE not not the the restaurant industry hadn’t quite fully adopted third party DSPs at that point. But over the course of four and a half years, we really, were able to grow a very sizable sales team and go to market strategy selling from single mom and pop restaurants, all the way up to you know, your largest enterprise QSR ours as well. And really, we were, you know, trying to create incremental revenue streams for restaurants to source revenue that they wouldn’t have otherwise had at the point where someone’s saying Who should I order from tonight? So that platform really enabled us to scale we all know that COVID really kind of fast forwarded a lot of that adoption from the restaurant and with third party DSPs but look, all that all that knowledge has kind of put me in a good place to come into sound down and support restaurants here as well. I love being at a company where you know, the technology is transforming how consumers interact with with restaurant merchants or how they transact with restaurant merchants and yeah, you know, the the two former spots where I landed, previously have that and Soundhound certainly does as well.
JJ
Jeremy Julian
4:59
You So now it’s, you’ve made some good choices on the career side, I actually had the restaurant person from Yelp on recently on the show. And as far as the third party DSPs, we’ve never had any of the direct DSPs on, you know, the DoorDash and UberEATS of the world, but we’ve definitely had the people that have integrated the tech into the restaurant. So for any of those longtime listeners, you guys probably remember my interview with, with the young lady as well as, as well as some different people that are that are helping with that. So ai, ai is all the rage. You know, when I say it’s all the rage, but it is kind of one of those things that that a lot of people don’t really understand. So why don’t we take it down to kind of the the, I guess the base level? How does AI work? What is AI? What is voice AI? I’ve personally gotten educated, but I’d love for you to educate some of our audience about what is voice AI? And, you know, like, is it just the the ability to roll down my windows? And my Mercedes? I know, that’s not it. But But why don’t we talk about kind of what is AI? And what is we say I actually even mean for those that are uneducated about voice AI, because I was walking around. I saw your guys’s team at NRF in January in New York. And I’m like, you know, I think every third booth had either machine learning or AI in it, if not every other booth. So
BB
Ben Bellettini
6:16
yeah, yeah. I always like to tell folks, look, we were in AI before it was real cool. Yeah, exactly. Certainly a second time we’ve been doing this since 2004. I think I think zooming out a really good frame of reference for thinking through AI as it’s some sort of technology that enables you to automate some function in your life, whether it’s via voice, or whether it’s generating a poem, which I’ve recently seen, it’s some sort of form to have technology to automate an actual workflow or produce a product. And so when we think about AI, in our world today, a lot of our most common interaction points are devices around the home. And the way that those are typically built as in these single, at least for voice AI, these single command or query instances, hey, blank, what’s the weather today in my area, right. So that’s, that’s voice AI as a frame of reference, how most folks have had exposure to it. And in a Mercedes Benz, you know, with our technology, there’s kind of two different core functions that can produce, it can produce edge technology, which focuses on, you know, supporting functions that actually power the vehicle, put the windows down, put my windshield wipers on the X setting, but when also cloud technology as well, you know, who’s playing in the, you know, Warrior game tonight, where they playing what time, right, so we have the capacity to hit the cloud and then actually return back intelligent answers that aren’t necessarily at the fingertips of the AI. And then how we’re actually manifesting this in terms of restaurants as we really have identified an opportunity in in our go to market to create automation for consumers to place orders with merchants. And there’s many different use cases, we’ll probably touch on a few today. But that AI really helps close the loop on a customer interaction and transaction and helps both the business and the customer. Automate that experience.
JJ
Jeremy Julian
8:24
Yeah, no. And I think it’s, I mean, I appreciate you giving us an education because I think all of us have been exposed to whether it’s Siri on our on our iPhones or, you know, Google Voice on on your Android phone, or Alexa, everybody has dealt with it in some way, some shape or some form. Talk to me about how you guys are thinking about that for the restaurants like what’s the most basic use case, because I’ve seen quite a few. And I know that you guys have got quite a few that you guys are working towards. But you know, is it ordering from my Alexa? Is it ordering at the drive thru? Is it having some, you know, robot taking my order at cashier? I know the answers to these things. But I’d love to talk through kind of kind of where you guys see that evolution? And where do you think that the the most basic automation is gonna be able to help to create a better customer experience again, oftentimes on the show, we talk about technology for technology’s sake is worthless technology that makes businesses better, more efficient, better guest experience is really where you need to focus. And so talk to me a little bit about kind of where you guys started from a use case and how you’ve seen it build. And then we can talk about customer adoption and, and, you know, customer failures. I’m sure there’s been some but let’s talk through that because it’s tech, New Testaments tech. That’s exciting. And out there.
BB
Ben Bellettini
9:35
Very good. Yeah. And I am before I hit on that, I just want a quick frame of reference. You know, I failed to do it. Moments ago. But, you know, when we think about AI, I want to really stress one our AI is is completely proprietary to us. So this is our tech stack. You know, there’s maybe a handful of other folks that actually own their own tech stack when it comes to voice AI we certainly do. And when we say we have convert rotational AI, this is truly a lot of deep meaning and depth to what we can actually support from an end user in our technology perspective. So
JJ
Jeremy Julian
10:13
a huge differentiator. Yeah. And
BB
Ben Bellettini
10:15
I think I think it’s important to stress that too, because the, if you think about interactions with AI that you may have had previously, a lot of that, again, is, like I mentioned before, what’s the weather, right? But when you think about what Soundhound can provide, and then how it actually helps in these restaurant use cases, we can handle really, really complex queries are questions that are thrown at us by end users. So things like context, if somebody in a restaurant experience actually places an order, and they need to go back and edit some component of their card, we have the capacity to do that, because we keep context throughout the entire ordering experience. In some instances, when you’re dealing with AI, you throw multiple queries at once the system, it will fumble that are a separates, because we have the capacity, actually triage and handle multiple different queries or questions all at once. What does that look like? Hey, can I add cheese to that burger? And what’s the price of that? Right? Most systems can’t actually handle that. So I think the way people think about AI today, a lot of the exposures that you may have had, the tech is ready, right? It’s just about finding the right vendor.
JJ
Jeremy Julian
11:28
Well, and one of the things that we find even in the point of sale world that that continues to be a point of contention, even for restaurant operators, is the fact that the soccer mom is coming in. And point of sale oftentimes is a very linear path when you’re bringing up something, but when somebody’s talking to you, a good cashier is going to take your order and internalize it and then translate it to the point of sale. You know, oftentimes, I’m sure you’ve walked into a fast food place and you go, you know, they ask is it for her to go before you even start. So give me even ordered anything, you don’t even know what I want, I might just want to know where the bathroom is, but you already start with Are you for her to go. And then you got to order the entree, and then you’ve got to wear the side. And then you got to order the drink. And when you’re the soccer mom, and you got three kids, and they’re running around, and they want to go to the Play place, like it’s impossible to do that. Obviously, consumers have been trained to think that way. And to do that. And at the same time, my seven year old daughter has not been trained at this point. So she doesn’t order the sandwich, and then the the fries and then the drink. She says I don’t know, I want a chocolate shake, and give me the waffle fries, or the curly fries. And I want the chicken sandwich. You know, she doesn’t necessarily order it the way that the point of sale might do that. Talk to me a little bit about how AI can help, I guess, I don’t triangulate, you know, create those things. Because to your point on context, you get all of the data elements, and then you can you guys can, you know, materialize them into what the business might need to be able to act on it.
BB
Ben Bellettini
12:52
Yeah, I think one, one example of this, or one thing that’s quite interesting in terms of how we handle some of these quite variable instances, I think, which is what you’re referring to today. Each region has kind of a different iteration of a dialect, right. And of course, there are many, many different languages that that we all speak. But Soundhound is actually deployed with over 30 languages that we support today. And a lot of that was driven by a lot of our auto manufacturing deployments. So we’ve actually assess, and we have, to your point, you said that the hot phrase, machine learning that’s assessing all those billions of queries that we’ve accumulated over the course of the years of our IoT deployments. And so when people speak with strong accents, when they speak, with strong Southern accents, or Northeastern accents, our system has an exceptional capacity to be able to deal with all of these different iterations of dialect, right, and I think many of these are regionalised, and many other AI’s may struggle with this, but we can help with those quite variable instances, that that may hit. And one thing that’s really interesting, too, I mean, US business is a melting pot, a lot of individuals that were supporting these businesses, maybe all of their staff is ESL, or the people that are calling in, maybe speaking in different dialects. And so when you plug in AI to support that customer experience, it can be a massive lift for an interaction point and ordering instance, so that I mean, that’s just just one example of how AI can help.
JJ
Jeremy Julian
14:38
Yeah, it’s funny, I when you I chuckled when you talked about the Dialects we we have a sales rep that works for a company that lives in Rhode Island, and he’s got a very, very thick accent and we were at a different customer doing a presentation. And we were talking afterwards and and the customer would say, Oh, where are you from? Where you from? Where are you from? And he said Rhode Island, and he said Rhode Island and the way he said it the cost summers like, oh, where’s that? Because he didn’t understand that he actually meant that. He said, Rhode Island, he couldn’t even understand what he’s saying. And we weren’t like, you know, but but the guy’s got such a thick accent. We’re like, oh, yeah, we have to translate for him from time to time. Just because his his accent, I guess, talk to me a little bit about how how the how the computer is, is, is learning these things, because some of these self learning models and again, in tech, a lot of people don’t comprehend, that the self learning models continue to get better over time, the more and more data they get, the better. They’re going to understand that my coworker from Rhode Island, he said, Rhode Island this way, and I might say, you know, a different way, and you’re gonna say it a third way. But the AI, you know, this, this artificial intelligence is learning the different dialects and the different ways that people say it, but they all mean the same thing.
BB
Ben Bellettini
15:47
Yeah, yeah, that’s a really good point. So, I mean, let’s think of a really practical example of this today, right? I may say Coke, someone else may say Coca Cola, someone may say, soda pop, there. And that’s just one, you know, item that someone could order, right. And so a lot of the backend work that we do to map different expressions, to the way we think of them as our like, core ordering options in a restaurant, domain, or experience, oh, there’s a ton of back end work that, that we do, and much of it is driven by machine learning, to best associate all the different intuitive ways that someone might actually try to order a product or item. I think, one one other thing that’s really interesting about our technology, especially applied to a restaurant instances, you know, I’m thinking of someone going up to the register and trying to place an order, think of all the non ordering lingo that’s flying around, yeah, when you’re actually trying to place that order. So there’s this whole stream of consciousness where AI system is listening in, but just trying to parse out, specifically, when someone’s saying words or phrases that intend to reflect menu items, or modifiers, or, you know, whatever it might be related to the menu. And so a lot of a lot of times, we can actually, you know, listen to 74 different words, and kind of parse all of those out. And then the second someone says, cheeseburger, our system immediately recognizes that and so you can’t get thrown off. If you can’t, excuse me throw off our system, by throwing a lot of non order, you know, commands or statements at it. What’s interesting, too, is our technology even has the ability to support non worrying questions, and an end user and AI interaction. So if you say, you know, what are the hours of the business? Or where are you located? Or what’s parking look like? Do you have any gluten free options, there are a lot of non ordering instances that we support as well, which is pretty
JJ
Jeremy Julian
17:59
well, and again, a lot of times, you know, and I want to get into the use cases, because I do think it’s important for people to comprehend that it’s not the, you know, it’s not the call on American Express. And you’re sitting there listening to this, this auto attendant, where you’re like zero, attendant, attendant, attendant, attendant, again, I’m picking on American Express, it could be visa, it could be American Airlines, it could be any of them, but a lot of them that have this this auto attendant, a lot of times when I talk to restaurants about AI and or voice, AI, they think of it as this as this nameless, faceless and not that it isn’t nameless and faceless, but it, it’s not linear. It grows and learns and understand what you’re doing contextually. It’s not there. So what is the most common use case? Like is it? Is it really just that phone order? Is it you know, is a drive thru as a kiosk as a cabin? You know, I think there’s a lot of different applications, but talk to me about really where most people are starting? And what are they finding that it’s helping? Because I think it’s I think it’s really cool, what some of the stuff you guys are already doing?
BB
Ben Bellettini
19:02
Yeah, very good question. So in terms of application and use cases that our technology actually supports, many of those use cases are ordering experiences between a customer and the merchant and obviously these these since we are voicing company need to be voice ordering experiences, you nailed the kind of core go to market that we started with which is phone. If you look at the phone today, you know, and you’re going in any business phone is almost always going to be second tier to the customer in house. And so many times customers are calling in, they’re not getting through, they’re getting put on hold multiple customers are calling at the same time. And either those are are you know, getting put on hold or just never getting picked up. And so when we think about this customer experience, you know, we’re constantly thinking of how do I serve the customer in house so that they’re leaving here with a really you no pun intended good to So their mouth about our business. And what what our experience does is enables you to kind of forget about the less palpable customer that still needs that, that soft touch that white glove approach. And our AI makes sure that they don’t wait takes accurate orders, speeds up their process and to kind of peek ahead. You know, we have the capacity to increase check sizes and do a lot. So phones kind of a really, really obvious way where we can help. What we’re finding, especially as we move up market supporting a lot of different QSR. Ours is there’s an appetite to really speed up and automate drive thru lanes. And so
JJ
Jeremy Julian
20:38
well, I think, but I think even in the phone, let me let me dig in the phone real quick. And just because I even have a, I personally have an itch that I’d love for you guys to scratch. As it relates to that, I happen to have a family of six. So you know, going out to dinner is not an easy experience. We go out a lot because I’m in the restaurant business, and I like to eat out. But oftentimes I’m calling to figure out what is the wait time? And can they see us, because I don’t want to drive 20 minutes to go have dinner, and then show up and have them not be able to see this. I’ve got four kids. So I’ve got to always figure out the no veto rule. And if I show up to a place and I can’t eat there, and I gotta go find another place, it’s a 10 minute conversation to figure out where the heck we’re going next. And it might be 20 minutes the opposite way. And so oftentimes, my engagement might not be even wandering, but to go figure out what how long is the waitlist? And can I get in for a party six? If the restaurant doesn’t have a digital experience? Is that a use case that you guys can solve? At this point? Is that something that you guys are in a place where, you know, hours of operation, again, is another one that I call often? Are you guys open? How late are you guys open? You know how much longer there’s a barbecue place here in town where I live? They close early? Hey, how much do you guys still have brisket in Do you guys still have, you know, whatever it is that I want, I will call over there. But I don’t know, probably five times out of 10, the phone just rings and rings and rings or like at the you know, pick up Hold, please. And I’ve already driven there and I’m already in the personal house, they pick back up the queue. And so I’d love to have you have you riff on that for just a couple of minutes if you don’t mind. And then we can talk through drive. Sir. I don’t mean to cut you off. No, no. But I think it’s huge for people because you don’t understand. And so many people even that are busy with phone orders. I know there’s industry for call centers that you’re paying 10s if not hundreds of 1000s of dollars to have somebody sit off site, doing the same work that no AI can automate some of the stuff. You know, I
BB
Ben Bellettini
22:25
look it’s I’m glad you dug in. Because yeah, it it can be a big friction point to for customers to capture this information in it’s ready state, like how is it active right now. And sometimes there’s a lot of different platforms that won’t support some form of active information that you can source. The good news is if the way that we look at it is if there’s a digital experience online, or if there is your point of sale system that’s housing some sort of data that the consumer may need. So as long as the merchant is managing that data effectively, then our AI system in many cases can pick that up and feed that back to the customer, right. One interesting example of this right now is thinking about a business where you’d have to actually potentially place a reservation, you call it front, you know, that host may be sitting somebody right now, they might not be able to actually pick up the phone, right? So our system can actually handle that consumer and enable them to place a reservation. And there’s a number of different kind of revenue actions, where that’s what we’re hearing from merchants is like, how do we recover revenue, margins are super tight. It may not be even about creating those revenue streams, but creating safety nets behind the revenue streams, you already have to make sure you’re recovering potentially lost revenue. So
JJ
Jeremy Julian
23:46
when you talked about even automating some of the common tasks that happen once you understand what those things are, you can build these models that can help with that, whether that’s a reservation, or that’s going on the waitlist, or that’s telling somebody what time they’re open till they’re, you know, do they have handicapped accessible parking? Or do they have a gluten free option? Just again, people don’t comprehend how often that you have to do that. And with where the labor pool is today, you know, in 2023, it’s hard enough to hire people, nonetheless, to get them trained up to be able to do these things. Yeah,
BB
Ben Bellettini
24:15
I mean, I mean, speaking from my just personal experience, I cannot tell you the last time I walked through the restaurant door, a restaurant stores and I didn’t see a help wanted sign, I mean, yep. Look, I’m sure those that are listening on this call. It’s not news that labor shortages are critically impacting our ability to effectively serve the customer right now or at least maintain, let’s say the same level of customer service that we typically had. There’s a unique kind of storm or climate right now with increasing wages, thinning margins with food costs rising. And so as if, you know, restaurant margins weren’t already razor thin and During that we’re capitalizing on revenue opportunities with consumers interacting with businesses. Now it’s like, ever important that we’re capitalizing on those and and giving the guests a really good experience. And so we are seeing a massive uptick in demand. I mean, we’ve had literally 1000s of inquiries from restaurants, who are saying, I need help, I can’t staff, these typical channels. Well, I in our, you know, we had kind of a soft touch interaction, how can you help me triage some of these, these consumer interaction points, and that’s what really what the AI does to solve, you know, phone as well as one good use case there.
JJ
Jeremy Julian
25:40
Yeah, well, and I just, I was just walking around to a very large mall, in, in the Midwest, yesterday, actually. And it was amazing how many stores were closed, because like closed with signs on the door that says, we don’t have staff to open today. Like, I probably saw 15 of them in this mall, which was insane to me, I’m thinking to myself, Oh, my goodness, the fact that 15 stores weren’t able to open because they could not staff them. I know all of our restaurants that are sitting on our listening, you know, that are listening have very, very similar challenges where they can’t open or they can’t open mean. And again, last week, I was in, in Houston, and I went into a nationally recognized chain. And the weight was 30 minutes for our party, not because they didn’t have seats in the dining room, they didn’t have staff to serve us. And they said that they said, you know, we’ve got one server on the floor, we can’t overstate them, because you’ll have a bad experience, you get to wait outside until that server frees up a little bit, which was a shame, because ultimately, they lost probably 150 to $200 order from our group, because they couldn’t staff it. And you know, if they didn’t have to, and again, not to say that we want to automate everything. But if they didn’t have to have a host be there full time. If the host was there to answer the phone and 80% of the time, how much could How much more could they have potentially served those guests from a revenue opportunity? Perspective?
BB
Ben Bellettini
26:58
Yeah, it’s a great point, too. And, look, I think for those listening right now, too, I think there is maybe, you know, we’re at this phase where adoption is really increasing with AI. But I think I’m sure there’s still folks that are out there that are thinking right now, you know, well, you know, is the automation less personalize? And is this? You know, is this a good enough experience for the customer? Well, one, as long as the AI is there, which it is with with sound downed? It is but to I think the other thing to look at is, once again, what does that guest experience look like right now? You just said to yourself? Yeah, if we could redeploy, saw the staff that is currently front of house to help with back a house and what if we could process more tickets and more guests in a really, really speedy fashion. And right now customers care about convenience, and speed, right, and so, waiting seven minutes at a table for my waters to hit the table and to get my first initial greeting. That’s not a better customer experience experience than speedy, rapid triaging there. And so we’re not, you know, look, we CRI not at all in labor and employee replacement. All we want to do is actually create fewer headaches for the staff. That’s, that’s at house right now that we can support some of these channels and augments, you know, some of those typical interactions.
JJ
Jeremy Julian
28:21
Well, and I think creating revenue opportunities that you’re probably missing today, again, we’ve got a chain that we work with that said that they go we know we’re missing, probably a million dollars worth of sales across the 70 stores that we have, because we don’t answer the phone. We don’t answer the phone fast enough. The kitchen have capacity, our staff has capacity to make the food, we don’t have capacity to pick up the phone and take that order. And you know, we’ve driven more and more people to the online ordering experience. And that helps. But at the end of the day if I picked up that phone, because again, I have four kids are longtime listeners know that oftentimes when I’m driving home from soccer practice, I don’t have time to get on my phone and go order the food. I want to call the restaurant while I’m driving there, order the food, go pick it up and leave. And if I can’t get that, ultimately, I won’t go back.
BB
Ben Bellettini
29:11
Yeah. Oh, sorry. I just wanted the one last thing to add on there. I think. You know, it’s so interesting when we do talk to third party, or excuse me when we to talk to QSR is in restaurants. I think there’s also this kind of hidden unknown with the phone, because maybe you don’t pick it up that much. You don’t realize how much revenue is still trying to penetrate through there. I don’t think third party DSPs have have gotten us to a point where there’s kind of this perception that the revenue is now flowing online. And look, there has certainly been a lot of
JJ
Jeremy Julian
29:44
guests experience was off. Sorry, I don’t mean it’s awful. It’s an awful guest experience and so they have to find some way if I want that product that’s the only way I’m gonna get it that night. If is to go on a third party delivery or go on the app to don’t get that, because calling is useless. Yeah. And
BB
Ben Bellettini
30:03
look 39% of off premise orders are still getting placed through the phone today. I mean, that’s
JJ
Jeremy Julian
30:10
an incredibly nice. That’s an interesting statistic.
BB
Ben Bellettini
30:14
Yeah. And so I mean that think about, you know, you don’t think of that as a revenue source today. But a lot of that’s just because you’re not looking at your phone provider data to see how many missed calls you actually got, right?
JJ
Jeremy Julian
30:27
Or how many calls did you have dropped? Because they answered it, and then they were on hold for 12 minutes, and they just hung up.
BB
Ben Bellettini
30:31
It’s a hidden hole, you’re not really seeing that the true, you know, revenue potential, because it’s, it’s, again, it’s not the customer that’s right in front of you.
JJ
Jeremy Julian
30:41
Yeah. So Ben, can be a little bit about you started to talk about it, and I cut you off, and I apologize, but you, you were talking about what does that look like to order through this AI? Is it the robot that, um, you know, how many times have we all yelled at, you know, Alexa or Siri? And we’re like, No, you don’t understand what the heck I’m talking about. I’ve seen your product, I wouldn’t have you guys on if I didn’t believe that it really can be as good. And quite frankly, oftentimes better than a human being might be to take that order. You know, whether you say that, you know, cheeseburger with extra pickles, and no, no mustard on it, you know, that that experience? Not only is it consistent, because it’s a computer that’s doing it, but also, you know, it’s going to always show up, it’s not gonna ever call in late second ever call in sick, but also it, you know, these learning models? So talk to me a little bit about what, whether it be phone or through the drive thru or or, you know, yeah, talk to me a little bit about that.
BB
Ben Bellettini
31:35
Yeah, that’s a really good question. So there’s kind of two different consumer ordering experiences, the way that we think about it. So to kind of separate your products, we talked a little bit about phone already, I won’t drill too much in on that. I think the the key there is like when you call in, you’re greeted by our voice AI that prompts you to ask questions or order. And then we go back and through via text to speech technology, where we’re prompting the consumer to order, we’re confirming what they’re ordering, enabling them to make modifications. So as long as your point of sale system allows it,
JJ
Jeremy Julian
32:09
but it sounds like a regular human, it doesn’t sound like a robot. Yeah. And then again, I don’t want to like this whole, you know, it’s not going to sound like Rosie the robot from the Jetsons. Like, it’s gonna actually sound like a human beings talking to you. And if you weren’t wise enough, you oftentimes wouldn’t. But this is the thing that I hear. And I’m trying to, I know, I keep teeing you up for these types of things, but I’m sure you talk to restaurant operators that think, oh, it’s gonna be Rosie the robot, and everybody’s gonna be like, Oh, I hate this, you know, give me the real person. But at the end of the day, if you’re not wise to understand that this AI is truly AI, it um, it very well, you know, you, I’m sure fools people all day long every day.
BB
Ben Bellettini
32:44
Yeah, it’s true. We’ve got, funnily enough, we’ve gotten requests for different celebrity voices and whatnot. And I think that’s, that’s a testament to, to how well the team has done producing a very lifelike AI conversational experience. I said conversational AI earlier in the conversation. And I think that’s really key, because that’s a, that’s a distinct separator for us, we have the ability to really synthesize and assess exactly what people mean by a patent, deep meaning understanding. So a lot of the way I don’t want to go too deep into the tech, but a lot of the way it works through other AIS is it does kind of this transcription and then tries to kind of auto map to a different, you know, answer that the system may provide. But what we do is we we truly try and synthesize each word, each phrase, the intent behind it and have much better responses, because of how we’ve actually innately built our technology from the ground up. And you’re right. I mean, it is relatively lifelike. We don’t want it to be too lifelike. And actually, yeah, fully consumer, we want to be transparent that you’re you’re communicating with AI, but it’s conversational.
JJ
Jeremy Julian
33:55
Yeah. Yep. Well, and talk to me a little bit about those, those consumers that want to opt out. So they’re in the middle of a conversation with an AI, they’re struggling with it, because they wanted this item that’s off menu, and it can’t translate that and the model doesn’t work for them. What does that look like for the customer? What does that look like for the restaurant brand, that whether it’s phone orders, or drive thru orders or whatnot,
BB
Ben Bellettini
34:17
there’s a there’s a ton of optionality about how we can handle that customer. That’s a really, really good point. There’s a number of ways where we can actually make sure dependent on use case that that consumer is able to connect with an employee at location, whether that’s passed through language, let me speak to a human or, you know, I want to talk to a person we enable, depending on the merchant settings, enable those consumers tend to get back to the store. We also have to, you know, depending on the use case, call center backup or support, if that business is just like we talked about earlier, unable to actually support those calls or those interactions that are coming through. And so we do have kind of restaurant’s specific call center support that can back in the AI, as well to make sure that we’re capturing 100% of the interactions, either through AI, or support. The good news is a fractional percentage of our interactions never end up there. Beyond nine out of 10 times AI is handling itself.
JJ
Jeremy Julian
35:19
Yeah, and that’s nine out of 10 odors that no human being needs to interact with, as it’s going through it. And so now you get 90% of your time back. Talk to me a little bit about you, you’ve alluded to it a couple of times is getting the water into the system, you know, and obviously, not every point of sale system is going to have integration. So I get that, but we can start with the ones that do. Does it look just like a human being or a web order coming into the system? Is that kind of kind of what happens? Whether it be a phone order, drive thru order, or you know, seminar or some other interaction from voice? Talking a little bit about what does it? What does it look like not to the end user but to the to the restaurant to or to the merchant? As you talked
BB
Ben Bellettini
35:54
about? Yeah, you really good question. First of all, hats off to our integrations team, if they are indeed listening to this call,
JJ
Jeremy Julian
36:02
integrations are off. Okay, I’ll let you keep going. It’s
BB
Ben Bellettini
36:05
not easy, though. I mean, you’re saying that though, because you’ve seen so many instances where 1000 flops, it doesn’t go right. But I really have to, I have to kind of stress the experience from a merchants perspective, that our integrations team internally and set up, we have tremendous integration into the POS systems that we support today. And so if a customer calls in, they place an order, they pay and finish up, then that order fires through into the point of sale system into the kts ticket prints out to your points, just like it’s an online order. So when you look at the operation of supporting on the back end, an order that’s placed through AI, you can except for a couple of quick data points that print on a ticket, and maybe a dining option, that set you wouldn’t know otherwise, that it wasn’t an an online order. So it’s super, super seamless in terms of processing, and picking up these tickets on the back end.
JJ
Jeremy Julian
37:04
I love it. And for those that you do have the integration with I’m assuming product mix and what’s available on the menu, you know, it’s like what happens in the fate in the case of an LTO, when you’ve got an LTO that’s coming out and your AI doesn’t know the model for the, you know, we’re in the middle, we’re recording in the middle right now. And Arby’s now has fish like Garvey’s never has fish, you know, 10 months out of the year, 11 months out of the year, they don’t have fish, but for the six weeks, they they realize that their sales are gonna be really struggling on a Friday, if they’re on fish. So they put out an LTO to put out this fish. And again, I’m picking on a really large brand. But what happens in that case where the model doesn’t understand the cod sandwich, or whatever it is that Arby’s is selling nowadays, how do you guys deal with that modeling and the otiose? And then the flip side of that answer, if you could bend is when you go out of stock on something, when we’ve run out of something, do you guys be able to you know, can you guys manage your way through both of those scenarios? Yeah, so
BB
Ben Bellettini
37:56
let’s let’s start with the latter of those two, first of our 86 items, we support that today. As long as that setting is updated within the point of sale system, we basically check on that via the API for every item and frankly modifier that gets associated with any order. So if it’s out of stock, we see it, it doesn’t have to be a core item. Either. It can be a modifier, we can do five, six different levels of nested modifiers, we can do half and half pizza. If you have the configuration, within your point of sale system, we can support it. And in terms of what’s really interesting, I’m glad you mentioned the LT O’s. We’re were actually recently with a QSR and demoing in an experience. And they said hey, can you know what happens if we wanted to add this, you know, X item to the menu? How fast could you get it up and live and and that’s actually one of the key separators to Soundhound. Because of all these different domains that we support the billions of different queries we processed already. We have a very, very mature AI and so we don’t require a lot of machine learning to get up and running. And with how frequently menu items change for merchants, we have the capacity to effectively I shouldn’t say immediately, but imminently change items, update menus, and have seamless consumer ordering experiences off of those as well.
JJ
Jeremy Julian
39:23
I love it. I love it. Thanks for Thanks for you know men again, I’m asking you questions that I get asked when I’m talking about products like yours to our you know, our customers because everybody’s everybody’s different. They have a different worldview and, and trying to kind of comprehend the last, I guess, line of thought I want to have on this as is we’re in the hospitality industry. And a lot of people you know, and we talked about it a little bit but what you know, I guess we’re looking to enhance the guest experience and you hear people that you know, Olo just got done with their their conference this week and the Have Noah, the founder of a low put put out a video that that kind of went through what he sees the restaurant of the future looking like we’re, I walk up to Ben’s table, and I realize Ben and Jeremy are having a business dinner. And it relates to Ben like, you know, red wines from Napa Valley. And so I offer Ben a red wine from Napa Valley, because the eye has told me that Ben is here because there’s phone checked in or some way to do that. And you know, and one of the comments this morning, first thing, you know, when he posted those videos, I’ll you take all of the hospitality out of hospitality. But at the end of the day, I don’t agree. I think that enabling computers, you know, 20 years ago, I remember I was 20 phone numbers. Now, I don’t memorize any phone numbers, because my phone has it. Why would I need to memorize phone numbers when my phone has it in there, and I can look it up faster than I can memorize it. No different than AI. I used to remember, you know, when Ben came in every time he liked red wines from Napa Valley. And every time I was there, and I served Ben, I got that. But Jeremy is not there anymore. Because he left because of COVID. Or he’s moved on to another store. He’s He’s left the brand. Now what happens to Ben Ben’s experience the cheers experience? AI is going to help with that. Talk to me a little bit about what? Where do you see the hospitality being enhanced? And for those that are fearful that you’re taking all of the hospitality out of the hospitality industry? You know, why is AI something that everybody needs to be looking at? Because I do genuinely believe that there’s a use case for every single business to implement AI and machine learning?
BB
Ben Bellettini
41:26
I have a lot of thoughts here being in this industry for almost almost 15 years. You talked a little bit about personalization, that’s of keen interest to me. Let me get there in a second. I think the number one thing practically for our product is consistency. I mean, how many times have you had an employee call out sick? How many times have you tried to get an employee to really hone in on upselling, that brownie at the end of the conversation is just forgot to do it, and processed 70 Guest orders in a two hour window and didn’t do it once right? Our system can be programmatically configured on behalf of the merchant to do that every single time without fail. And so consistency across a number of different cohorts is something that you can expect from AI. And so what that does is not only that, that improves potential revenue lines, for restaurants. But that also provides opportunities for like, really, really strong and uniform branding for these businesses. You know, we didn’t really talk about all of our use cases today. Of course, we support phone support, drive thru customer and action points. We support kiosk we support in app. And one thing that’s unique to Soundhound is we can provide a really, really unique voice interaction experience across all of these different platforms, branded specifically to your liking where you’re using one vendor for all these different interaction points. And so I think, I think uniformity and consistency is key, I’m really interested about how customer data and customer transaction history can improve the consumer experience as well. How easy was it think of the last time I’m thinking of a time where I had to call to your point for a large family and place 17 different items. While I’m unfortunately pretty predictable person, when I call into my favorite place, I’m ordering the same thing over and over. So how great well Can it be when instead of as a customer calling and placing a five and a half minute order in a in a grueling background noise and you can’t kind of understand what the other individual is saying. And the system prompts you to say, Hey, Ben, thanks for calling in again, are you interested in placing the same order as you did last time, and all of a sudden, that’s a that’s a 14 second experience, maybe my cards on file. So there’s so much opportunity to just make people’s lives easier, both on the merchant side. And on the consumer side with the adoption of this technology?
JJ
Jeremy Julian
44:02
Well, and it’s funny, because you get all the all of the people that are all about, you know, personalized data and security. It’s like, but when’s the last time you logged into your Google account and logged into Chrome and tried to browse the internet? When’s the last time you went into Twitter or Facebook or, you know, in incognito mode? It’s really a difficult experience when it’s not customized to you. And so understanding that, that people do want the cheers experience as often as they can. And if AI can help with that, whether it’s at the drive thru or it’s on the phone order, or it’s in the dining room, they want that and so what does engagement look like with with engaging with your team? Ben? You know, you’re the VP of sales, you know, you’re looking at get word out about this, because there’s so few, you know, if there’s 700,000 restaurants in the US, you know, you guys have a very small market penetration at this point. But I do truly believe that 10 years from now, we’re not going to know what we did without it. And so what does engagement look like with you guys? Who’s the ideal client? How do they engage with your team. What does that what does that look like? Since you get to? You know, head up the the restaurant segment?
BB
Ben Bellettini
45:05
Yeah, excellent question. So I think when you look at all the different storefronts that exist, you can be a mom and pop shop that’s been in, you know the family business for a century. Or you can be a franchisee that’s part of a large QSR. Or you can be you know, the corporate CIO or CTO, or CTO of a QSR, it doesn’t really matter, your business type, there is a voice AI use case to support it. And so we have voice AI experts really at every different level that can help guide you through the implementation of this technology. I think the other thing too, is, look, this isn’t, you know, we don’t want just a simple cookie cutter experience for everybody too. So when you do speak to somebody on the team, they’re going to help guide you through a lot of the different types of settings that we can create for you customizing an experience that enables it to feel very, very similar to the customer experience that optimally they they get with you today that they may not get with the kind of consistency they want. And so we certainly welcome to the conversation, you know, we come to through the aspect of, you know, how can we support whatever channel it may be with your voice AI needs, and customize a solution for you.
JJ
Jeremy Julian
46:21
I love it. I love it. So when do I get to order food from My Vizio TV? Is that coming soon?
BB
Ben Bellettini
46:27
I am so glad you asked that. Because that is the dream state look, we’re gonna create a dense, a very, very dense populace of merchants that are using our technology and all those IoT platforms I was talking about earlier, whether they be auto your in vehicle driving home, and you order on the way home, but you don’t you want to be hands free. If your front of your TV you’re popping on Netflix at 630. On Friday, you want to order a pizza. That’s something that you’re going to see very, very soon.
JJ
Jeremy Julian
46:54
I love it. I love it. So soundhound.com or what? What’s the website that they need to go check out to learn more about, about your guys’s stuff and engage with somebody on your team?
BB
Ben Bellettini
47:03
Absolutely. Just go to soundhound.com navigate to our restaurants site, and feel free to submit a form and we’ll we’ll be happy to help you.
JJ
Jeremy Julian
47:11
Awesome. Well, Ben, I thank you for the education. I’m genuinely excited about you know, I I clearly have opinions about this. And I’m a restaurant tech guy, but but I do see it making the world a better place. And, you know, while I can understand that the nefarious nature of you know, self learning AI, I do think that that is going to make our lives easier. And I do think that even Alexa, as much as you know, we might we might tease about Alexa being difficult, I promise you it’s a heck of a lot easier when I tell it to turn off my bedroom lights in my bedroom and rather than having to get out of bed or you know, my wife and I teased about the clapper, you know, from 20 years ago, but you know, it makes my life easier and I’m willing to pay for that convenience. And so the fact that that you guys are doing this is not only going to help the Restaurant Brands but it’s gonna help the consumer create a better guest experience. So to our listeners, guys, we know that you guys have got tons of choices as I say on the onset while you guys are online checking out Soundhound please go subscribe to the podcast if you guys haven’t already done so. Lastly, go give it a rating on your favorite podcast player because it helps others find find it if you found this episode interesting. Share it with with other colleagues. Ben, thank you so much for your time and to our listeners make it a great day. Thanks Jeremy.
I
Intro
48:24
Thanks for listening to the restaurant technology guys podcast. Visit restaurant technology guys.com For tips, Industry Insights and more to help you run your restaurant better
