← Lead With AI
Lead With AI
The AI That's Already Taking Your Drive-Thru Order (And You Don't Even Know It)
The AI That's Already Taking Your Drive-Thru Order (And You Don't Even Know It)
Loading summary
Transcript47 lines
- [00:00]
Dr. Tamara Nall
Imagine pulling up to a drive thru and being greeted by a voice so natural, so responsive, you forget it's not human. Now imagine that voice. AI taking orders faster, more accurately and more consistently than any other employee ever could. Today on Lead with AI I'm joined by Krishna Gupta, co founder of Presto Recordings, the AI voice platform quietly powering some of the biggest restaurant chains across the country. This isn't about replacing people, it's about reimagining how we serve faster, smarter and still deeply personal. So grab a seat, place your orders and get ready for a front row view into the future of food service. Let's get into it. Welcome to lead with AI I'm Dr. Tamara Nall. In each episode we we will take you behind the scenes with visionary leaders shaping the future of AI across public and private sectors. Join us as we explore groundbreaking projects and innovations that are transforming industries and making a real impact on people's lives. Let's dive in.
- [01:20]
Hi everyone, how are you? It is Dr. T, your host with lead with AI and I am so excited to have Krishna Gupta with us. He is the co founder of Presto. Welcome Krishna, how are you?
- [01:32]
Krishna Gupta
Great. Thanks for having me, Tamara.
- [01:34]
Dr. Tamara Nall
Absolutely. And you are my first guest where I'm international. I'm actually in Ghana here having a great time, business, personal. And so you're making history with Lead with AI being my first international guest. So I appreciate that. Thank you.
- [01:50]
Krishna Gupta
I love that. And I'm international too right now, so we're international duo today.
- [01:57]
Dr. Tamara Nall
I love that, I love that. Get a passport and travel. So let's start with the spark. I love to learn a little bit about the co founder, who you are at your core. At what point in life did you feel that Presto was solving a needle? And what drove you to build the company based on your values or your experiences?
- [02:22]
Krishna Gupta
Yeah, it's a great question. You know, Presto has always been in the business of helping restaurant chains be better businesses and be better for their customers and leveraging the modern, the most modern technology tools in the, in the service of that mission. And you know, at a personal level, I have always been very interested in AI and the interface between humanity and AI. I'll give you an example. My mom, she was a master's student in computer science and she was studying AI back in the 70s and 80s and she thought that by studying AI while I was in the womb, it would make me smarter. And so I literally grew up hearing about Marvin Minsky and a variety of interesting AI theories. And so I've been particularly interested in how humans communicate and voice AI, because I look at voice as one of the most authentic ways that humans communicate. If you look at primitive man or woman, right? They only had two ways to communicate, voice and touch. That's kind of it, right. And so our chemistry and biological functions work most authentically when we talk to each other. And so I've been Interested in Voice AI for 15 years and I've been watching the development of that with machine learning and then the first generation of NLP voice bots. And the spark really was in 2020 and 21 when two things happened at the same time. One is the transformer came out, right? And the Google research and the LLMs that kicked off this wave of AI revolution that became a reality in 2020. And that sort of made me think that, okay, wow, this could actually be a significant improvement in conversational AI over the clunky voice bots that people had been trying in 2017, 18, 19. So that was one. And on the other side, Covid has shut down restaurants and sit down restaurants and the drive thru was exploding. And Presto, which had we had been serving casual dining restaurants with tabletop products, we were trying to figure out how do we align ourselves to what's happening in the world. And when those two things collided, the market need and the technology progress, that's what created this spark. And we were way ahead of our time in the sense that if you look at verticalized applications of voice AI, LLM, native voice AI, this was the first application that came to the forefront and that people started talking about and then implementing.
- [04:56]
Dr. Tamara Nall
Wow. And if I'm going through a drive through, I won't know whether it's a human or an AI. Correct.
- [05:02]
Krishna Gupta
You know, it's a great question. We can make it so you don't really know. But a lot of restaurant chains actually want their customers to know. So they want the voice to be just slightly robotic, friendly, human like, but they want it. They don't want that sort of quote unquote dystopian feel that, hey, what is this? I'm being treated by. But that's changing tomorrow. Now, as. As people are getting more used to AI in their lives, more and more restaurant chain customers want the voice to be very human like. And in that case, you don't really know.
- [05:38]
Dr. Tamara Nall
Wow. And for those that don't mind, it sounded robotic. Is it because they don't want to be misleading? And I mean, I guess it wouldn't bother me, as long as I get my order, it doesn't matter. To me, and I'm sure it's faster. And I'm thinking, wow, this, this restaurant is trying to be more efficient. So, yeah, I guess it wouldn't bother me at all.
- [06:00]
Krishna Gupta
Yeah. So I think, I think the reason is, you know, the demographics are at play here. And so if you look at some of the restaurant customers and some of the restaurant, you know, for QSR chains, some of the older customers are not as interested in being served by an AI. You know, they still quote, unquote value, at least in their heads, they value the human touch. Now, like I said, as they're seeing the AI filter through the rest of their lives and as they're Starting to use ChatGPT or their children or grandchildren are coming to them and saying, hey, look at this amazing translation tool, or let me, you know, I can practice my Spanish with this thing. You know, they're, they're starting to realize that, hey, this is actually not a bad thing to be able to use AI in such a fashion.
- [06:50]
Dr. Tamara Nall
Awesome. Awesome. Now let's move to the holy smokes moment. What is that moment when people first experience Presto and that they realize that this changes everything for their operations and their business?
- [07:03]
Krishna Gupta
Yeah. So look, I think we have two types of customers and we always think this is a B2B2C business. So at the restaurant chain customer, for me, the holy smokes moment, or for them, I should say, it's when they realize they don't have to spend as much time or money or energy at the order window and can dedicate that towards a better customer experience, towards a more accurate order, a more well functioning kitchen. And the greatest example of this for me was when the CEO of one of our customers, Taco John's Heather Neary, who's a dynamic, very modern and forward thinking CEO, she actually went and worked the drive through herself and she used the AI and she saw it in practice and she came out of that being like, wow, this made my job so much easier and so much more pleasant because I could spend my time doing better things and just sort of being repetitive with customers. And so that I think. And it extends from there. Right. And so these restaurant employees then start treating the AI like a team member. They start relying upon it. And so you know that that's a, I think, an energizing and a holy smokes moment.
- [08:16]
Dr. Tamara Nall
Yeah, no, it actually does. And I'm going to tell you, our listeners are very curious people. So talk to us about the brain of Presto and if we open up the hood, what's making it magically work.
- [08:30]
Krishna Gupta
Yeah. So look, we've always taken. So let's take a step back. There's a lot of different approaches to conversational AI and voice AI. And if you look at the spectrum of things people have used in the past, on the one end of the spectrum you have a. And I'm talking about the last, whatever, let's say 10 years, right, or five years, you have a purely LLM large language model approach where it's like, oh, I have no human anywhere and I'm just going to completely automate everything. And on the other side you have what people have used in the past, which is purely a, hey, I'm going to use a call center and you can have fully humans. And I think we sort of believe in this hybrid approach we call a neurosymbolic approach, which is the way basically immerse the LLMs with a brain, an NLP, a natural language processing brain. And so the LLM is free. I mean, LLMs have their limits. They hallucinate, they make mistakes. Right. I mean, the other day, it's a great example, I was coming up with a flyer for the agenda of one of my summits and I was just saying, hey, why would I have anybody design this? I'm just going to Type this into ChatGPT or whatever. And the damn thing couldn't get it right. Right. I would switch one thing and it messes up something else. And I switch one thing, it's right. And it's just because the edge cases for LLMs are still problematic. And so, you know, you. But in a drive thru application you can't have 10% or 15% of the customers getting some crazy hallucination. Oh, I ordered a burger without onions. Turns out I get a burger only with onions. Right? It's like, you know, it's like, you know, so, so, so I think, I think we've always believed when I started this, you know, we had two teams, LLM team and an NLP team. We've merged those two. And so we have an LLM product and then we have the brains, which is an NLP thing which uses human in the loop as well to ensure the context that is so restaurant specific. It's so even individual restaurant specific, you need to be able to feed that in. Right? So that's how it works. And if you want this application to work not only accurately but consistently across hundreds of restaurants in a brand, which means you're talking about many franchisees with different ways of setting it up, you have a drive through in Minnesota, and you have a drive through in Florida. You know, the outdoor environment is different for each one, so you have to figure out how to filter out background noise. You know, people might order a tea, and they mean a hot tea in the north and an iced tea in the South. So all of these contextual things, right, you need some way to ensure that that intelligence is kept up to date. So we merge those and that's called a neurosymbolic approach. And that's a very complex interplay between the AI and the restaurant stack. And I believe that's the only way to build applied AI products for verticalized use cases using voice AI today, whether that's restaurants, home services, healthcare, financial services, whatever. And by the way, I'm involved in all these different verticals with the various companies. So that, that's what's happening under the hood.
- [11:36]
Dr. Tamara Nall
No, I love that and I love the way that you're able to easily explain it for those listeners that, you know, are a little bit afraid or think it's too technical and put it in very, very simple terms for people. And then at the same time, if people want to learn more, they can, can do that as well. So I really, really appreciate that. Thank you. Now, you talked about Taco. Was it Taco John's before? Is that the example?
- [12:00]
Krishna Gupta
It was, yeah. Yeah.
- [12:02]
Dr. Tamara Nall
Okay, so talk to us about another like draw dropping example where even you got chills and, you know, did your little Victor dance. If you want to show us what that is, that's fine too. But you know, where you were like, wow, we are really like killing it. This is amazing. I am amazed at what we've been.
- [12:22]
Krishna Gupta
Able to do to me, you know, again, and you'll hear this across our chat today, tomorrow, I'm very human centric, which is odd thing to say for somebody who's so immersed in AI. But I believe the best AI applications are the ones. And the best AI companies and teams are the ones that remain close to the humans involved, whether that's the customers or the consumers. So to me, the idea that groups like Taco John's and many of our other customers view these AI applications or products as an extension of their team is very, very powerful. And when I hear that coming in, I, that is truly what makes me really excited. There's other things too, right? You know, for example, when, when the AI is able to quickly tell that there's a combo offering available, hey, I ordered a, you know, Big Mac or whatever, right? Some kind of burger. And, and I can. And then I ordered something else and quickly I can tell it that, okay, these two things means that I'm, I'm basically getting a combo. So I can offer you that combo and actually save you a consumer a couple of dollars. Those are things that are really cool because they're not necessarily things that, you know, the employee, the human employee wouldn't necessarily always remember. Right. The upsell is also really, you know, really can be really magical. And we're not quite there all the way yet, but I look at the perfect upsell and I think about all the different things that we could do to drive personalization. And the final thing I would say, which we can do but I think will be the most jaw dropping thing when I talk about it, it is the most jaw dropping thing is that I can finally get demographic information. The QSR restaurants have no clue. If I ask them a 30 year old man or a 60 year old woman came through your drive through, what would you offer them differently? They literally have no clue because there's no other way for them to get the data. But now using the voice, I can tell if it's a male or a female. I can tell the rough age range. I don't need to know their name, I don't need to know anything about them, but I can kind of match that to their preferences. So if I say, you know, Carl Jr. Which is one of our customers, you have a lot of franchisees who are customers there. You know, they ran a special for Alex Earl for the Super Bowl. Right. And Alex Earl is this very famous TikTok influencer. You know, she's going to appeal to a certain demographic. But, but you know, you don't know if that burger special was taken by just 20 year old women or did 30 year old men also really like that? Right. I mean, so there, there is data like that that totally changes and over time. What I want to be able to do is perfect the AB test so a restaurant fan can say, hey, I want, I'm thinking about this special or this new product. Can I quickly ab test this and get really good data within a week as to whether I should roll this out or not? Right. So, so those are some of the things that really get me excited and get our customers excited.
- [15:16]
Dr. Tamara Nall
Yeah, no, that's powerful because on the one hand, going back to the combo, there are times I order stuff and I'm like, I think they probably is a combo but there's all these screens, you know, there are people behind me don't want the food to get cold, etcetera so that.
- [15:32]
Krishna Gupta
And you just said, sorry, interrupt you, but you said something really important. All these screens, right? It's what I said earlier. Primitive man did not look at all these screens, did not look to, you know, identify 50 different items that they wanted. They just talked, right? They just spoke. And so that, that element of bringing back the, this actually brings back the conversation versus hey, look at our 100 items and tell us what you want. That's a very suboptimal way of driving an ordering process if the point is to delight the customer, right? And look, our name is Presto. I want people to create magic with our product. And so that's how we think about it.
- [16:13]
Dr. Tamara Nall
Well, that's good because we're about talking about those magical stories. That's amazing. Now let's talk about the ethics, because there obviously are some ethical boundaries that you have to keep in mind. What are your thoughts about that?
- [16:26]
Krishna Gupta
Look, ethics are very important and everyone has a different definition of ethics. So I'll start by saying, first of all, there's certain things, so we don't touch certain things on our end, like allergens. If somebody says, hey, you know, does this product have peanuts in it? You know, I'm not having our product touch that, right? I'm sending that to a restaurant employee in the store and they're dealing with that. So we're very careful about not having any risk to the end consumer. Second thing people talk about a lot is, you know, the humans and what happens to the humans in a world of AI. And my philosophical view on this, Tamara, is that in general, every, every technology revolution has created net more wealth and more jobs for people. And I believe the same is going to be true here. I don't think that, you know, restaurant chains are going to have as many team members taking orders. That's true. But I do think that they will have plenty of opportunity for those team members to, you know, help customers create a better and more accurate ordering experience. And we're very, very focused on augmenting those team members and basically extending their abilities so that they can focus on what they like doing and what they can do really well without having to do a bunch of the repetitive stuff. And I think that's the way AI will filter through for the time being. I'm not a big believer that, you know, everyone's job is just going to be automated in the next five years. So that's how, you know, I think the way I talk about it is the human AI, right? There's a dance that needs to be practiced and learned of how do you get the AI and the humans to waltz together? And in every application, that waltz is a little bit different. It's set to different music and a different time. And you have to figure out how to perfect that. And so in this application, it's about how do you get the humans interjecting in the right time in the right places between our humans in the loop, the restaurants humans, the AI, the brains behind the AI That's a waltz. That's a pretty damn complicated thing. Right? And so that's how I think about the ethics.
- [18:39]
Dr. Tamara Nall
And you mentioned something, you said that way, with AI as team members, others can focus on things that they like to do. What does that look like? What roles would that illuminate towards?
- [18:53]
Krishna Gupta
For example, I think there's food preparation, right? And so right now, what happens in the environment is that everyone's just rushing between 10 tasks. And that leads to a. They're not happy. The employees aren't happy because they don't feel like, you know, that they're actually doing anything to perfection or doing it really well. And then the end result is, you know, you get a bunch of stuff wrong. Right. And that doesn't help the end consumer. The other part that I find interesting is, you know, right now there's two windows. First one is order taking, and the second one is, you know, you're picking up your food, right? And right now you don't have time to do anything other than pick up the food. But imagine if you had 2 seconds, 3 seconds more, just to have a really to leave the customer, the end customer, with a little friendly, hey, you know, have a great day and hope you enjoy the burger and let us, or let us know how it tastes. Right? You know, those things make a really, a real impact on the end consumer. But right now, you have no time to be. It's like, okay, here's the food. Boom, boom, boom. Like, move to the next thing.
- [19:59]
Dr. Tamara Nall
Right, right. No, I like that. That's beautiful. Thanks for that. I really appreciate that. Now, what does the future look like with Presto?
- [20:06]
Krishna Gupta
Yeah, the future. You know, I. I believe in a. So I believe in vertical AI a lot. And by what I mean by vertical AI is that AI applications and platforms that are tailored for specific industries. Right? So Presto's only ever served restaurants, and they only intend to serve restaurants. And our vision is to build a connected end to end AI platform for restaurants. And so, you know, we start with voice AI. We've actually done a bunch of work on computer vision. In the past as well, which we're now looking to integrate as well. Robotics. And so I believe that over time we're going to have this end to end connected vision and platform that we can work with our customers to be their AI part. We want to be their enterprise AI partners. No reason they should. You know, they don't want to talk to seven different AI companies. They want one AI partner who can help them navigate all these different things. And at the root of all this is what I was teasing a little bit earlier in this interview was around data and personalization. Being able to capture real time, really perfect information about your consumers in an environment that was not possible before unlocks a lot more interesting AI applications over time. So that's how I think about the next three to four years. And from an end customer perspective, I think that means you have a really smooth personalized experience where you're, you're going through and not only getting an accurate order that is hot and friendly and whatever, but also maybe getting items that you didn't know you wanted but actually wanted. Right. You know, there's lots of things you can sell. You know, I'm in Sicily right now and if you at any point in the day upsold me a lemon granita, you're going to get a cha ching and you're going to get a big smile on my face, but I might forget to order it. Right, so.
- [21:52]
Dr. Tamara Nall
Right, right. Yeah, I get it. Okay. That's amazing. Now make some of our pessimistic listeners believers. What's something they can do this week to explore, build, try to get a taste of Presto?
- [22:07]
Krishna Gupta
Well, they can go through one of our drive throughs. I think that's the best way to experience it. And then, you know, there's a bunch of people that are posted on Tick Tock about, about our drive throughs. And so, you know, you can, you can see Carl's Juniors and Taco John's. And so if you live around these several other brands that we, you know, we, you know, Wiener Schnitzels and and so on so forth that, you know, you can try out. And there's more that I can't talk about on, on this podcast, but, but we will be, you know, we're, we're nationwide, we're expanding with several different restaurant groups in every region of this country. And so as that is growing, I would love for your listeners to try it out and then sort of start dreaming about what is possible from there. This is a very tangible thing. It's a very tangible Way for someone to touch what AI can do, because it's just. It's a local drive through.
- [23:01]
Dr. Tamara Nall
Yeah, no, I love that. So as Krishna said, go on social media. He's being followed. Their technology is being followed. Look up Presto and find a local restaurant near you to experience a drive thru. Awesome. So we do have a segment here that's called From One Genius to Another. So our last guest has a question for you, and I'm going to read it so I make sure I get it right. What is the biggest problem with AI that you see and what others might be missing?
- [23:32]
Krishna Gupta
I'm going to link it back to the theme of this podcast for me, which is the human element. I think there's so much talk about AI replacing humans, making humans redundant. And look, some of that is absolutely fact, actual, true, relevant. But we can't lose sight of who we are and what our society functions on and why our species exist still. And I think a big part of that is we are very good at communicating with each other and leveraging each other as groups to sort of advance our society and our race and our species. Right. And so I think, like, there's a. At the root of any AI application, I believe there needs to be a human touch. And a lot of AI companies forget that. They think they can just automate every step. And what they end up with is an application that isn't as accurate, hallucinates, feels impersonal, starts feeling robotic. And it might sound cool initially and it might work initially, but over time, I think the companies that will win this wave of AI for the next 2, 3, 4 years are the ones that are keeping the human condition and the human experience at the center of their philosophy.
- [24:53]
Dr. Tamara Nall
Great. I love that. No, and you're absolutely right. People talk to me about that all the time and ask me that. I'm like, humans will not go away. You will always, always need humans in the loop. You'll always need that.
- [25:06]
Krishna Gupta
We should talk about this tomorrow. Imagine if you were an AI podcast interview. This podcast would be not nearly as engaging. It wouldn't. It just wouldn't. Right. And so, so, so, so, you know, we need tomorrow and your personality and your incisive questions on the fly to make this as interesting and compelling as it is.
- [25:25]
Dr. Tamara Nall
Absolutely, absolutely. And I'm not sure if you knew this, but we actually hit number one in technology on Apple Podcast. We were there for a month, so I'm really excited. It's because of my guest and the wonderful products that y' all are developing. So I Feel very blessed. Blessed to be on this planet.
- [25:39]
Krishna Gupta
Congratulations.
- [25:41]
Dr. Tamara Nall
Thank you, thank you, thank you, thank you. So let's do our bonus rapid fire most overrated take trend humanoid robotics. Okay. Most under hyped AI Trend Voice AI. Okay. Love that one book. Everyone should read on the future or AI or just in general.
- [26:01]
Krishna Gupta
Go read philosophy, go read Nietzsche. If you, if you understand the human condition, you understand everything.
- [26:07]
Dr. Tamara Nall
Awesome. And the boldest AI prediction you believe.
- [26:11]
Krishna Gupta
I believe that many more people than we think are going to end up with AI boyfriends and girlfriends.
- [26:16]
Dr. Tamara Nall
Yay. I'm so glad you said that, Krishna. That's actually my area of focus. I too believe in AI humanoid human relationships and that people will form relationships, get married, have families and so we're definitely going to talk about that later. I love that.
- [26:36]
Krishna Gupta
I think it's kind of scary, but unfortunately it is true. It's happening because of lots of reasons that we can talk about at another time.
- [26:45]
Dr. Tamara Nall
Right? Right. No, absolutely. Okay, that's good. I'm writing down all those folks that on this podcast say that so that we can form some allegiance, et cetera. So thank you so much for this time, Krishna. How can we get in contact with you? How can we be tapped into Presto social media? Just let us know here.
- [27:04]
Krishna Gupta
Yeah, look, I'm on Twitter. Krishna K. Gupta, easy to find on LinkedIn. Presto has a LinkedIn page and we're putting out some. You know, I'm gonna be putting out a three minute video every week around AI for executive decision makers and sort of making some, some light of the latest movement in AI and how that applies to executives in applying AI to their businesses. So please do follow along. That's gonna be on spot and those are the areas you can find me.
- [27:35]
Dr. Tamara Nall
Awesome. Well, thank you so much and again, thank you for making history as I'm recording here in beautiful Ghana and we look forward to seeing the future of Presto and are making drive thru and rest our restaurant experience being much more amazing. Thank you so much.
- [27:52]
Krishna Gupta
Thank you. Have a great time in Accra.
- [27:54]
Dr. Tamara Nall
Thanks. Take care.
- [27:55]
Thanks for tuning in to Lee with AI. I'll see you next time as we continue exploring the cutting edge innovations shaping AI across the public and private sectors. Until then, keep leading with AI.