The history of restaurant CX: From maître d' to AI

About the show

Hosted by Nikola Mrkšić, Co-founder and CEO of PolyAI, the Deep Learning with PolyAI podcast is the window into AI for CX leaders. We cut through hype in customer experience, support, and contact center AI — helping decision-makers understand what really matters.

Summary

Hospitality has always been about creating great experiences for guests. But behind the scenes, the way restaurants manage reservations, customer communication, and demand has changed dramatically over time.

In this episode of Deep Learning with PolyAI, Nikola Mrkšić speaks with Steven Fine, GM of Restaurants at PolyAI, about how restaurant technology has evolved and what that evolution means for customer experience today.

The conversation begins with the early days of phone reservations and moves through the rise of online platforms like OpenTable. While digital channels made booking easier, they also introduced new operational challenges for restaurants, from no-shows to fragmented systems and missed opportunities to engage with customers.

Despite the growth of online booking, the phone remains one of the most important moments in the customer journey. Many diners still call restaurants when they want a table quickly, need to coordinate a group, or want reassurance that their plans will work. Those calls often represent the highest-intent customers.

Nikola and Steven explore how voice AI is transforming that interaction and helping restaurants rethink how they manage demand and deliver service. In the episode, they discuss:

The discussion also looks ahead to the future of hospitality technology. As AI systems integrate with reservation platforms, POS systems, loyalty programs, and ordering channels, restaurants gain a more complete view of their customers and new ways to personalize the dining experience.

Key Takeaways

Transcript

[00:00:44] Nikola Mrksic: Hello, everyone, and welcome to another episode of deep learning with PolyAI, our podcast for CX leaders looking to use AI to improve customer experience and for everyone else interested in the world of AI, voice AI, and what, you know, the future world looks like. Before we start, please like, share, subscribe. Today with me, I've got Stephen Fine, who is our GM of restaurants. He is the main person looking after our hospitality vertical, which was our first, and it's fast growing today. We're here as part of a week where we launch PolyAI Host, which is our kinda like one click solution for restaurants, and we'll talk about all that. But, Steve, it's good to have you here.
[00:01:23] Steven Fine: Yeah. Good to be here.
[00:01:24] Nikola Mrksic: Yeah. So I think, you know, everyone knows I'm a history buff. So before we go into the company propaganda and just the shape of the market, when did the first restaurant take the phone reservation?
[00:01:36] Steven Fine: 18 1786. 1886.
[00:01:39] Nikola Mrksic: Mhmm. Yeah.
[00:01:40] Steven Fine: 1886?
[00:01:41] Nikola Mrksic: I don't know. We were we were looking at it.
[00:01:43] Steven Fine: We were looking this up on ChattGPT, and it was way longer than I've definitely ever been alive. I mean, that was well before me. And I think, I mean, the world's come a long way since the first phone call.
[00:01:53] Nikola Mrksic: Yeah. Yeah. I think when I saw the year, I think it was 1876. I was like, oh, two years before the Berlin Congress. That's a long time ago. And, uh, apparently, by eighteen nineties, it was common in London, New York to see kinda like newspaper ads saying, you know, like, phone in to preorder, phone in to make a reservation. Why do restaurants take reservations?
[00:02:13] Steven Fine: I think there's, like, this consistency that they want of customers coming in. I give it with all marketing, you always wanna find a way to customers to actually get to you. And without them needing to walk in pre the phone, this was the easiest way to actually say, hey, customers. We can we can book your space for you. You can come in at the time that works for you. And for them, that's like a major revenue driver. So it just kinda starts making sense as a way to to get customers to you.
[00:02:39] Nikola Mrksic: Yeah. Yeah. So on one hand, they lock in revenue. Right? I I guess what's always been interesting, especially when you see a land in New York, was the feeling of exclusivity and, like, the the difficulty of getting a reservation from a key place.
[00:02:50] Steven Fine: You can always also almost imagine at the beginning is if you had a phone, you were probably a little bit exclusive. I remember re like, actually reading this book when it was talking about people where they've got their first phones in the car. You can just imagine a chauffeur calling calling
[00:03:05] Nikola Mrksic: through Do you remember your family's first mobile phone?
[00:03:09] Steven Fine: I don't. I remember my first phone, though. I remember my first Sony a little Sony Ericsson.
[00:03:15] Nikola Mrksic: He
[00:03:16] Steven Fine: I actually it was first one was a tiny little Sony Ericsson. All we did on it was play snake. And I got
[00:03:22] Nikola Mrksic: I snake on on Sony Ericsson. I remember the snake on the legendary Nokia thirty three seven. You know the one I mean? Yeah. Of course. Break. Yeah. Yeah. We have one in the office. We tried to turn it on, and it was on a head charge from, like, whatever decade that we we founded before. Yeah. Yeah. That's amazing. I remember my dad getting his first mobile phone, and he was, you know, relatively high up in Yugoslav state company. That was a big deal. And I inherited that fault, like, many years later, and I still have it somewhere. So it was a big deal. But okay. Like, the exclusivity absolutely matters. Right? And I think that with the other side of the of the coin is it's actually quite hard to take reservations. Right? Because no shows are really bad for a restaurant.
[00:04:07] Steven Fine: Yeah. So I think it's always just, like, dynamic you have is you want and you see this. I'll I'll relate it to the current world's challenge that they have today is you want to get as many bookings as possible, make that as easy as possible. But the easier you make something, the more likely someone ends up booking and not actually coming. Yeah. And one of the big discussions I had with partners at OpenTable was the reservation on Google that click to reserve straight from Google. Yeah. Those have there's literally zero friction to getting a reservation at a at a restaurant. But with that lack of friction, it meant that they had higher no show rates than any on that channel that they'd seen before. So they actually pulled that off for a while because it was actually more hurtful to restaurants because they were stopped people walking in. When real customers were making bookings, they actually couldn't get a space, which is a huge loss.
[00:04:59] Nikola Mrksic: No. Yeah. No. No. No. I mean, it's a huge deal. I know that even for the NHS, I remember a leadership debate between Rishi Sunak and Liz Truss where Rishi, who I think most would agree is more technically competent, tried to really impress upon her and the audience how important no shows were to substantially improving the capacity of the NHS. And I think they fell on deaf ears because people don't really understand just the downstream effects of turning people away and then the restaurant's empty. Then people think something's wrong. They don't book. When we look at, like, back to the history. Right? We went from the phone and then, like, the Internet happened. Yeah. So when did the first online reservation?
[00:05:39] Steven Fine: So OpenTable were the pioneers on this. So I think there was a couple people before this, but when
[00:05:45] Nikola Mrksic: They're the ones that made it.
[00:05:46] Steven Fine: They're the ones that made us a reality. So OpenTable started 1998 when they launched that. And then, obviously, like, there's the progression of more and more people getting online. Yeah. And then as more people go online, there's more places to then make a reservation. Right? Restaurants start having their websites. So and with that, you start going, uh, well, this is an unbelievable way for to get people where they are. So when people are searching the web, whatever they're doing, you've now made it so easy with OpenTable and later other providers to actually just make a booking directly into your restaurant. With the same challenge of friction, though, you also do end up with no shows. And this is Yeah. Slight side side thing, but one of the things we always saw with the phone is people often call to cancel reservations. Yeah. And, I mean, I know personally, I feel really bad if I'm not going to make it, and I just quickly hop on the phone and cancel. Often those don't get answered, and that is then a no show. Alright? So even a cancellation is better than a no show
[00:06:47] Nikola Mrksic: as 100%. Yeah.
[00:06:48] Steven Fine: Cancellations you can fill.
[00:06:49] Nikola Mrksic: Yeah. No shows. Either with walk ins or with others. Exactly. Right? Yeah. Yeah.
[00:06:53] Steven Fine: Exactly. And then I think that that whole move to online was, like, amazing for the industry. It started creating a lot more structure in their table management as well because you start pulling from online, and you pull that into a system that then people your staff can track. Yep. Yep. Much more efficient. You can move tables around easier instead of this, like, paper note Yep. On the way down.
[00:07:17] Nikola Mrksic: Yep. We'll talk about that in the public occupancy management because I think that's been really fascinating. But I remember when we started the company, our first enterprise client client to this day for five years now, a company called Whitbread, the largest FTSE one hundred hospitality company. And we started working with them in their beefeaters pubs. And those are really iconic for PolyAI. They're, like, you know, been there a lot. We have the logo hanging in our office. But what I find really interesting is just, well, two stats. One is that for London and New York, which you think of probably, like, the most, like, urban fine dining and dining and, like, populous metropolis. Right? You think of online reservations, and you would think they would completely take over. And the split is still fifty fifty. Like, it just it reached that level and never really turned to fully online despite it being objectively. If you wanna do it in the shortest amount of time, it's probably easier to do it online if you have a phone in your hand. Why is that?
[00:08:14] Steven Fine: Well, I think Chicago would challenge you on on a third metropolis.
[00:08:18] Nikola Mrksic: We've all seen the bear. We've all seen the bear great show.
[00:08:21] Steven Fine: I think that, like, when you look at human behavior, we've always like, the phone is especially on bookings that you want kind of immediately, it's always the channel that people turn to. Right? They still turn to this day to pick up the phone to make a booking to check if they can get it in the next forty five minutes in the next hour. The ones. Yep. So, like, very, very high intent bookers often book, like, immediately on the phone.
[00:08:47] Nikola Mrksic: Okay. So it's a higher intent job.
[00:08:49] Steven Fine: High end
[00:08:49] Nikola Mrksic: They're much more likely to show up.
[00:08:50] Steven Fine: 100%. And I think, actually, we worked with OpenTable. There was amazing data to show that the phone no shows were about 25% lower than people who were booking online because it's high intent people. Right? When you're calling, you're you're like, this is my decision. Yep. When you're doing it over online I mean, I know I've booked, like, three places at the same time just going, I'll let people know which one I wanna cancel, which one we wanna go to on the night. Yeah. I mean, I guess I'm the the worst person. Yeah. Or for a
[00:09:20] Nikola Mrksic: Def definitely. Definitely.
[00:09:22] Steven Fine: I don't do it anymore because now I now I know how their lives look work.
[00:09:25] Nikola Mrksic: Yeah. You've seen you've seen inside the machine. Yeah. Yeah. Okay. Okay. So so wait. Then kinda, like, when you look at the whole industry, like, where did it go from there? Like, like, what happens with these different systems running in restaurants? And, like, why is it valuable to have, like, an open table and or or or one of their competitors?
[00:09:42] Steven Fine: So if we go back to, like, all the systems that restaurants brought in so, I mean, historically, if you go back, they would have had, you know, next to nothing. I remember this the the story of a group that a tech company in in the in The UK who brought in, like, a POS system. And it was literally them in
[00:10:01] Nikola Mrksic: a pub in POS, point of sale. Of sale. Sale.
[00:10:04] Steven Fine: You can pay with a card. And What? And they it was a guy at who went into he ran a pub, and he said, this is just obscene that I'm writing every single thing down on piece piece of paper. Then the pricing changed, and he needed to distribute that information to all of his team. He's like, can't we just, like, put this behind, like, literally a calculator, and it pulls it pulls it in? And that was in, like, the foundation of the POS. It was literally I remember the story. He says, he has, like, little stickers on numbers that each sticker represents something behind that. So it represents something in inventory, represents the pricing. And when they clicked they clicked, you know, a pint of Guinness, that pulled through to say, okay. Here's the pints of Guinness. This is what And, like, all this stuff historically with tech is it's to make everyone's life easier. Right? So it's less manual work. It's more reliable. It gives you better data on your customers so that you have used that in the past. And then I think, like, you that is, like, a foundational part to it. Then you move to, like, open tables. That is now it gives you data on exactly when customers are booking, how far in advance they're booking, how many people they're booking for, and that allows you to actually realign your table settings to increase occupancy. So I think there's, like, all these massive benefits that everyone's gotten from technology. And, I mean, it's always the point of technology. Yeah. Make people's life more efficient and make businesses more Yeah. Like, immediate to manage.
[00:11:34] Nikola Mrksic: The occupancy stats are really interesting. Right? Because when you've created what you need to facilitate these reservations automatically, you also create that, like, ledger of what's available. Right? And then you're basically filling up inventory. I think the one that I like the most was we noticed that a lot of the restaurant table configurations well, we look at when when we say no. Right? Kind of because it's interesting. Right? And it turned out that large groups calling an hour before they were gonna show up on Saturday night, mostly Friday, Saturday night, mostly groups of men looking for the next place to have drinks. And they were being turned away from restaurants which had enough capacity, but just not the right table setup. And I think digging through it after many, many months, I think we realized that the restaurant managers don't love these groups because they're loud, they're noisy. And then, you know, decision is really at the top level. Do you want to be the kind of place that we'll have, you know, 12 loud men, you know, getting lying around each? The answer is for the most part, yes, if it's a liquor left behind. Yeah. Okay. It's because these are the money well, I mean, the high margin groups. Right? But I just thought it was really fascinating in that this channel could tell you that this is happening downstream. Right?
[00:12:42] Steven Fine: Exactly. And I actually, it's not the only time I mean, we did it with with Whitbread. We did that, and we opt to help optimize their seating through the excess demand. We've actually seen, especially as you move from single site groups, where then the owner manager is literally there, they can really assess everything. It's a lot easier.
[00:13:02] Nikola Mrksic: Yeah.
[00:13:03] Steven Fine: But with a lot of the multi site groups that we've worked with is we present where customers are being told no at times. And we've actually had because it's hard to manage across so many sites, we've had them actually go into their table management system, look at where the times that we've been saying no, and then literally reoptimize their whole seating within their TMS. Yeah. And, ultimately, like, reoptimization is a way to drive more covers, a way to drive more seats. And it's always hard because I've always like, I think about this often when it comes to the phone channel is no other channel do you get to see when there's no when a customer tries to book but doesn't there's no availability. So if someone goes online and there's no availability, they just don't click that time.
[00:13:53] Nikola Mrksic: That's right.
[00:13:53] Steven Fine: They're only ever going to click times that are available. Whereas on the phone channel, they're asking for a time.
[00:13:58] Nikola Mrksic: Yep. Yep.
[00:13:59] Steven Fine: They're being presented that, no, you can't book at that time. Here's the closest one, and then booking the closest one. So now you actually have data on what people are trying to do, which previously used to be, like, a total black box. Yep. Yep. Like, total black box.
[00:14:15] Nikola Mrksic: Yeah. No. It's really fascinating. I mean, I think that the other stat that I'm remembering as we speak is uncannily across many different restaurant groups in in America, in in Europe. The missed calls rate tends to be around thirty percent. Sometimes a bit up, sometimes a bit down, but, like, 30 seems to be, like, the golden rule. And if you just think of, like, would people show up and the theory you know, we talk about ROI software, like, whether Gen AI pilots are converting or not and why not and why yes in this case. And you can make up any number. If you were genuinely recuperating 30% of the business, that would be, like, unbelievable ROI. Right? And we know it's not that because some people go online, some people walk in, but you never know how many. And I think the most precise kinda calculation for this came from Fogo, the child, one of our clients that recently had our kickoff. Do you wanna maybe give us a bit about that?
[00:15:12] Steven Fine: Yeah. So I think, first, Fogo were the most data robust client I think we've ever worked with. I remember when we started the journey with Fogo de Chao, they came to us with a business case breaking down everything that was happening. So it is 25% of calls from us, 25 to 30%. Yep. It's 500,000 calls a year Yep. Which is insane. And they knew that with each each reservation is worth about $200. Right? And what they came to us with that challenge was like, we can't do this. And it's it's not just about the in hours calls that people are missing, because you miss calls when staff are busy at your peak time.
[00:15:53] Nikola Mrksic: Yep.
[00:15:53] Steven Fine: Yep. A huge portion of this is after hours. Yep. Yep. Yep. No one's there to answer the phone. Right? And it's it's never going to be a problem you solve with human capacity. Yeah. So what they came that exact issue. And after implementing it, we ran a pilot with them to actually see the impact of us. It was supposed to be three months and only with, like, three sites three to five sites. And what they ended up doing after a month is cutting it. They were like, no. Cutting it in a good way. Yep. They're saying, look. We can let's let's do this and scale it to 86. They had and they've got private equity backers, so data is everything to them Yep. Yep. Is they saw 7,200,000.0 uplift in incremental revenue. Right. That's about $7,000 per site that they have. So, I mean, if you could add $7,000 to Per month. Right? Month. Sorry. $7,000 Yep. To every site every month. Yep. Like, why would you not? Why would you not do that? Yep.
[00:16:56] Nikola Mrksic: Yep. Yep.
[00:16:56] Steven Fine: It's like the most obvious ROI that I've seen, and it's why I was desperate to take on restaurants as a Yeah.
[00:17:03] Nikola Mrksic: Yeah. Yeah. Of course. As a division. Of course. Yeah. Yeah. No. No. Totally. And I think that, like, when you look at those examples, it's especially for an enterprise group at that scale with expensive covers, it's a real moneymaker. Right? Because I think, like, half of that is profit. Exactly. Really, like, flows into the bottom line in a spectacular way.
[00:17:21] Steven Fine: But I think what's important is, obviously, that's we're we're talking enterprise here, and it's why I bring it down to per site level is because we also serve individual locations.
[00:17:30] Nikola Mrksic: Oh, totally. I mean, I think the point is that, like, an enterprise sometimes, in Fogo's case, spectacularly, that they're really good at capturing the data, so we know. But, you know, we I think we're deployed at well over a thousand different sites around the world before we started working with them. And only when we had their the the data were we able to actually put some numbers behind it in a way that we trusted. Right?
[00:17:53] Steven Fine: Yeah. So I think I mean, we I work we I did some work with a little individual location in Sarasota recently, and we worked out on on their data that we were booking $15,000 in reservations every single every single month. I mean, he was blown away because prior to this, you don't really know how much money is coming through the photo. You kinda guess. You're like, yeah. I think I get a lot of reservations.
[00:18:20] Nikola Mrksic: I mean, man, like, you meant I mentioned you mentioned Chicago. Have you seen the bear? Yes. Yeah. I mean, like, you know, I think that, like, most startup people can actually me a lot of anxiety. That's No. No. No. No. No. I think that, like, a lot of startups see, I'm, like, getting animated here. Like, a lot of startup people can really identify with that show because every day is kinda like that. But, uh, they're just fighting for, like, bear no no pun intended, bear survival. Yeah. It's really hard business. So to think that they're, like, sitting behind there, like, quantifying it all is insane. But, also, like, it's an example. You know? It's one of our top two verticals, the other one being health care. Both have examples of these jobs that involve picking up the phone to serve as demand, but they're not primarily a job description. Right? It's not a contact center dealing with it. It's actually almost more sales job than it is a service job, but it's one of your jobs. And it suffers at peak when you can actually do the most damage. Right?
[00:19:11] Steven Fine: Yeah. No. I mean, I it's I always think it was like distributed sites. Yes. Right? Like, it never makes it's, like, very hard to justify having a contact center, and we actually do have some customers who
[00:19:23] Nikola Mrksic: Oh, yeah.
[00:19:24] Steven Fine: Set up contact centers to solve this exact issue. And I always think, like, that's there's this idea I remember reading a product book ages ago. Like, there's always an existing alternative to solve a problem. Like, there's not only one way to solve a problem. And we saw, like, quite a few, like, large groups tried to solve this problem with contact centers, and they did to for the most part of in hours. They solved the in hours challenge, but it's excessively expensive
[00:19:49] Nikola Mrksic: Yes.
[00:19:49] Steven Fine: To solve this within and then you've still got all the after hours times that you can't ask for.
[00:19:55] Nikola Mrksic: Super peak times, you'll still get because, like, the microcosm of, say, 80 sites at scale during lunchtime or dinner time when they get the most calls, like, sure, they're only picking up phone calls, but the spikes are such that you need a very atypically shaped contact center with staffing to service the full capacity. So even if they centralize it, they still miss a lot of the calls during peak times.
[00:20:20] Steven Fine: Exactly.
[00:20:21] Nikola Mrksic: It doesn't even solve the problem. And then they're sitting there doing nothing for most of the day.
[00:20:24] Steven Fine: And and the I mean, this the cold thing to me is, like, it's high intent customers booking for the same like, very often same day or tomorrow. Right? High intent. It's when they're shopping around. So they are the people that you know when they book, they are going to come, and you also know that if they don't get through to you, they're looking to book somewhere. Huge you see this huge at some of our clients who are largely, like, corporate dining Yep.
[00:20:49] Nikola Mrksic: Yep. Yep.
[00:20:50] Steven Fine: Is their PAs often call. They have Yeah. Yeah. And what you often see is that they are shopping around. They're like, I'm just gonna find the place that I can book
[00:20:57] Nikola Mrksic: That would pick up.
[00:20:58] Steven Fine: Yeah. That would pick up. And they and they calling in because they wanna they wanna add notes to it. They wanna make sure it all works. And that's where I think there's just this huge opportunity that was not available to restaurants before. And I think it's, like, a good way to think about, like, it some of this tech started only being able to be available to the enterprise.
[00:21:20] Nikola Mrksic: Yep. Yep.
[00:21:21] Steven Fine: You do long deployments. You have, like, intense professional services behind it. But as you start, we've done this enough, you start going, well, we can actually build this at scale Yeah. Individual Yeah. Locations for smaller sites and give them the same power that once was only available to the enterprise Yeah. Now becomes available. I
[00:21:41] Nikola Mrksic: mean, I remember our first deployments. You know? These were, like, multi month builds, testing, rollouts, tuning. Like, different restaurant groups had different behaviors, different back ends. It was still quite complicated. So what changed?
[00:21:52] Steven Fine: So I think there's a few things around, like, the integrations that you have. Ultimately, integrations are what drive past completion. Right? To make a reservation, you need to be integrated into an open table. To make a large party booking, you need to be integrated into a triple seat. You have to be integrated into these systems to complete the task.
[00:22:13] Nikola Mrksic: Yep. Yep.
[00:22:14] Steven Fine: So when we've now that we've built out we've built out the core integrations that we know most restaurants use, it means that you can just reuse those very quickly. And in an agentic way, make sure it's tailored to each of those restaurants. Yep. That's right. So the integration portion is basically taken down to zero. Yep. And, I mean, this I mean, we are we launched PolyAI host, which is takes under five minutes to get an agent set up with your knowledge base, knows about your restaurant, and is connected
[00:22:46] Nikola Mrksic: to OpenTable. So just to kinda, like, re reconfirm. So PolyAI host a solution for OpenTable users where they go into the partners tab. Yeah. So click a button, and, what, three minutes later, an agentic workflow has taken all their information from their website, from their OpenTable instance, and created their preservation history.
[00:23:08] Steven Fine: Yeah. Like, I probably shouldn't I'm not gonna swear on the on the podcast because
[00:23:12] Nikola Mrksic: Go ahead.
[00:23:13] Steven Fine: But, like, this shit is amazing. It is it is, like, unbelievably powerful to be able to sit on a call with a restaurant owner, go through and we don't even need to be there, but we do as part of our sales sales. They get they we sit there. They go through a three minute flow that pulls all the information and connects to their open table, and they are ready to start taking calls. It makes something that feels scary that was the scary lift
[00:23:45] Nikola Mrksic: Mhmm.
[00:23:46] Steven Fine: To be incredibly easy so they can get these results that used to only be available to enterprise. Right? So imagine the capability to, in ten minutes, start adding $7,000 to your business next month.
[00:24:02] Nikola Mrksic: It's incredible. Right? It's it's like real ROI. And it just works. Right? I think the beauty is, like, you know, when you say enterprise, really, in the in a world that is still true for complex workflows and com you know, custom integrations, it is still this world of deployment strategists and forward deployed engineers. There's still, like, this CapEx investment and effort on both sides, including the customer's IT teams to build an application that works for them. But this this is almost like a SaaS like promise. So, like, click. Here you go. Right?
[00:24:30] Steven Fine: Exactly. I look. I we I mean, we serve both markets.
[00:24:34] Nikola Mrksic: Yeah. Yeah.
[00:24:34] Steven Fine: The enterprise needs that extra, like, love and care and help. Right? The reason for that is just because when you're looking at 20 sites, there's a lot of complexity in put pushing that onto someone who's frankly never built out these systems before. It takes time. But when you've built up a more templatized solution that is templatized in terms of integrations, but not templatized in not giving personalization to that restaurant Yep. We just wanna get this in the hands of everyone that that can do this.
[00:25:06] Nikola Mrksic: No. No. We have big plans this year for this. Yeah. It's
[00:25:09] Steven Fine: how do you think it will go? I mean, based on what we've seen, like, last year, we we we saw about a 10 x increase in in phone calls through our systems Yep. Yep. Which is incredibly, I mean, amazing. It gives us more data. It allows us to better, like, train our models. And with that, I'd expect I mean, I wanna I wanna do 20 x. Yep. And we like, this is that moment where we we've made it so easy for customers. We're offering them to get on for free at the moment so customers OpenTable customers can come on for free and sign up. Yep. And that's our whole point is we wanna put this power in the hands of every single restaurant owner.
[00:25:51] Nikola Mrksic: Yeah. No. It makes sense. And I look. I think it's it's a far cry from where the tech used to be. And now with it, it's just everyone should have it because those that don't will be missing the calls from those PAs that are trying to make that that reservation. Right?
[00:26:05] Steven Fine: When you think of this whole kinda, like, workflow, what's next? So with this is then we we start looking at, well, what do you tag on to that foundation? Right? If OpenTable and your TMS is your foundation, then you start tagging on everything that drives revenue. Right? So you've got large party bookings, huge revenue driver for businesses. Yep. Yep. And what you're seeing in an industry where, frankly, like, margins are often tight Yeah. That's an extra, like, few covers makes a big difference to them. A large party is just starting to account for 20 to 30% of restaurants revenue. Really?
[00:26:40] Nikola Mrksic: Yeah. Like, it is a key only, like, big metro areas or, like, in general? Or
[00:26:45] Steven Fine: Largely focused on metro areas, but you're seeing as you move
[00:26:48] Nikola Mrksic: These are corporate events, stuff like that, where people really kinda, like As where we really open up the checkbook. Right? Exactly. Yeah. Yeah. Makes sense.
[00:26:54] Steven Fine: Like, I mean, I've we've gone to company events. Spoke with us. We spend more money.
[00:26:59] Nikola Mrksic: Absolutely. Absolutely. We do.
[00:27:00] Steven Fine: We spend more in a company event.
[00:27:02] Nikola Mrksic: There is a client who I can probably name. Andrew Landry's. Yeah. I'm confident that across different Landry's side, the events we've hosted, that we pay Landry's more than they pay us. And we're talking about serious 6 figure a month
[00:27:13] Steven Fine: both ways. Yeah? Yeah. Yeah. So so it's like a then your your next thing is you're okay. Well, now I'm going to I'm gonna do large large party bookings. And you're integrated, leads are captured, and it goes through a CRM. Now you're okay. Well, now I've taken, like, a significant amount of work, but also I'm cap making sure no revenue call gets missed. And and then your last your last kinda lever to that is food ordering.
[00:27:37] Nikola Mrksic: Mhmm.
[00:27:38] Steven Fine: And when you start going into food ordering, that's that collection is the bigger thing. I think, like, on the whole, Uber Uber Yeah. Deliveroo, DoorDash, all of those, like, that is the mechanism that most people are ordering through. But there's still a huge portion of collections. Yeah. And curbside is huge. Yeah. Curbside pickup is huge. And that's, like, the full breadth of where restaurants capture revenue, right, especially over the phone. And as you tick off each of those, like, you live in this world of promise where you a restaurant never misses a revenue call again.
[00:28:14] Nikola Mrksic: Yep. Yep. I think the other one that I found really interesting was with one of our clients, we ran these, like, mother's day, father's day campaigns where I think once you made a reservation for, I think, two weeks before, it will go, like, have you booked for mother's day? And most people go, no. Well, would you like to? Because if you come in between us I think it was between two and 6PM, we'll give you a round of drinks for free. Now the reason that is done is that they staff up, and they're, like, full to the brim. Even if they don't run this campaign, for lunchtime and dinner, it's, like, one of the these are, like, the heaviest days outside of, like, the Christmas period. Yeah. But I think it was what like, between two and six, they end up empty. Yeah. So this, like, round of drinks is basically very little extra CAC for getting, like, a whole party in, and it drives I think what it was like a do remember the
[00:29:03] Steven Fine: like a 5% uplift on revenue on the day.
[00:29:06] Nikola Mrksic: Okay.
[00:29:06] Steven Fine: And I think I I mean, there's a there's a part to this, like, upselling and upselling motion is with if you have people in an in an industry where attrition is very high Yep. I mean, if you go to one one front of desk that's there is probably not going to be there in six months' time
[00:29:27] Nikola Mrksic: Mhmm.
[00:29:27] Steven Fine: Like, they these people move on this very, uh, like, it's a, you know, university, college job Yeah. Yeah. Yeah. Is you don't get consistency. Right? And what the promise that, you know, you can make with AI is it's going to be consistent on the upsells that drive the behaviors that you want. Yeah. So a really nice example with the customer is loyalty is a huge thing for their business. They drive a significant amount of business through loyal their loyalty program. And what we're we're able to do is we have, like, agent memory where we remember customers who have called in, and we offer them to sign up for the rewards program. Right? And we know that 40% of the people who we offer it to end up signing up for the rewards program. And that's like you go, okay. Well, great that you can take all of our reservations. Great that you can, you know, you can do our large parties. You can do food ordering. And I was like, well, you're actually driving business well beyond Yeah. Just Yeah. The the time of the call. You're driving it off of the call because now we've got them in our database. They're signed up for our reward. Yeah. And it actually becomes this revenue engine that didn't used to exist because they weren't being upsold on on rewards.
[00:30:41] Nikola Mrksic: Right. So rather than becoming just the kinda, like, system of record in the vanilla operational sense, it becomes like, I guess, you know, kinda like a Salesforce who goes from, like, a service cloud and sales cloud to marketing cloud. Right?
[00:30:52] Steven Fine: Yeah. I mean, it's it's, like, starts to become a a whole revenue engine to the business. And it's what's exciting is it's just like a revenue engine that was like an old revenue engine parked in the garage that never existed. Yeah. And you, like, just put it there and
[00:31:07] Nikola Mrksic: Like, did you know that it was a v 12? 600 horsepower. It's beast.
[00:31:10] Steven Fine: Exactly. So I think, like, this is where this world becomes very exciting, and it's almost a lot of this stuff gets put a little bit on autopilot. Mhmm. Like, you you tweak and adjust chair. You make sure it stays up to date in certain things. But it like, you said it earlier, phone AI and what we do and how we do it with our tech, it just works.
[00:31:32] Nikola Mrksic: Yeah. Yeah. It works in, like, a very human way. It brings the technology out to everyone in a way that is not different to them speaking to humans, and it drives all these benefits. Because, you know, I always you look at, like, the primordial state of a system at the very highest end. Right? And that's often the best it can be if well provisioned. And, you know, I think of, like, the better restaurants in in Belgrade where I'm from. Right? And they've got that like, the the waiters, the staff, the general feel is very kinda, like, French inspired, like many things in in kinda, like, nineteenth century Serbia onward. And, you know, waiters who know what you should have and, you know, service that's, like, very professional, but not overly in your face and kinda, like, a more American style where they'll they're they're constantly checking in and, you know, each has pros and cons. Uh, but one thing that I always found incredible was you call up and you say, hey. Have you got a table for, like, three people? Yeah. One. 5PM. Yeah. Okay. Name? Nicola. Okay. Cool. Bye. Yeah. That's a call that took fifteen seconds. It works because the person there is trusted to jot down a line that says 3PM group of three, Nicola. Yeah. Underline, whether they added to a CRM or not, I don't know. Right? But it is very different to when I think of the first versions of our experience where it would then be like, oh, great. How many people? Like, that many. You go through that. Oh, great. It's available. Yeah. What's your name? Nikola, last name? Marczic, last name, Marczic, last name. And on. Right? And let's say that you somehow survived my my last name. Right? Then, you know, there's some, like, GDPR inspired thing. You know? Is it okay to send you a text confirmation to the number? You're just saying, you know, like, oh my god. That's why everyone assumed that the phone would be dead. Yeah. Right? And by and large, you know, we still have places where we ask for those bits of information. But I think that memory piece is exactly the important bit because if someone agrees that you can, like, use their data, next time they call in, it's like, I wanna book a table. Table four. But today, 5PM, three people. Cool. Okay? In COVID number four? Yeah. Cool.
[00:33:37] Steven Fine: Yeah. All good. It's amazing.
[00:33:39] Nikola Mrksic: Like, we're back to that experience. Right? Yeah. And not even that, but then, you know, if you see a pattern, like, cool, and it's like, they see that I come in on Saturdays. Nicole, like, how how are you doing? Like, table for four again?
[00:33:49] Steven Fine: Yes. Exactly. And do you go, like, no.
[00:33:51] Nikola Mrksic: No. No. It's six people this time, like, tonight at 6PM. Cool. And then, you know, you can remember, like, shellfish allergy? Yep. Shellfish allergy. Exactly. And pro you probably don't have to confirm it. Right? But then it becomes, like, about that next layer of hospitality where they're, like, AI meeting, like, high end Exactly. Like, upmarket behavior builds a better version of that.
[00:34:10] Steven Fine: I mean, we spoke about this, like, just before was, you know, why were people and hospitality people hesitant to adopt this kind of techno technology?
[00:34:21] Nikola Mrksic: Care about the experience.
[00:34:22] Steven Fine: Exactly. Yeah. Right? They they they care about their customers. Like, the more like, sometimes people who aren't in the industry, like, you you get a bit of bad service, and you're like, these people don't truly care. When you're in the industry, these people and operators, they care so much. They, like, they care so much about people and the experience, which is why, like, the idea of putting an AI in front was was scary for them. And I think what we've seen I mean, we've been doing this for Yeah. Seven seven years now. I think what we've seen in the advancement of technology and integrations is it, like, feels real now. Yeah. It starts knowing who your customers your customer is, who's calling in, so you have that level of personalization. Yeah. I don't that's, like, to the groundwork of, you know, TMS table management systems, so open tables, is that you can you you know who you Yeah. Yeah.
[00:35:16] Nikola Mrksic: So they build down the info. You know? And Internet terms, they lay down the fiber. Right? Exactly. They build the data centers Exactly. That you need for AI.
[00:35:23] Steven Fine: And now and now it's just it's it's I mean, the quality of voice now nowadays is it sounds natural. We make things
[00:35:30] Nikola Mrksic: better than I do in English. Right?
[00:35:31] Steven Fine: Like Yeah. My accents my accents sometimes like, I I mumble over some words. I try enunciate.
[00:35:37] Nikola Mrksic: The the only podcast with a mumbling, sir, but I saw it after. Like, uh, who swear during the podcast, which is not really a surprise. They have the ethnic stereotypes. Yeah.
[00:35:46] Steven Fine: So I I just think, like, what you're able to do to do now is, like, this is this is the moment where it's all really coming together. Yeah. And, honestly, like, it's so good now. It's gonna be
[00:35:59] Nikola Mrksic: better and better every year.
[00:36:00] Steven Fine: It's gonna be better and better with with all the next levels of technology that that come out.
[00:36:06] Nikola Mrksic: Yeah. Yeah. I mean, look, the next release of our LLM Raven v four, exact naming to be determined. There are some drafts probably, you know, fully live in in about a month. My last name just works every time because the reasoning model within goes, okay. Weird. Nikola. Uh-huh. Serbian name. Let me let me try to decode the thing that came after, uses a different decoder, figures it out. It's I couldn't have imagined this happening in this time scale. It's it's incredible.
[00:36:32] Steven Fine: I mean, I mean, you've been through the whole Yeah. Journey of it, and, I mean, you can see it.
[00:36:38] Nikola Mrksic: I mean, you know, like, not through the whole
[00:36:40] Steven Fine: Yeah. History.
[00:36:42] Nikola Mrksic: You know, DARPA started finding speech recognition so Americans could spy on the Russians in the sixties, but it's really taken off in a very, like, singularity kinda way. When you think of, like, these high end properties, because I think these were always, like, the the things to win. Right? Like, you know, I think we work with, like, Boca Raton. We work with Caesars. We work with Gordon Ramsay restaurants. Right? Like, what sets them apart from, like, the wider kind of market?
[00:37:07] Steven Fine: I think it's the the quality that they require. Right? It's not a it's not a we'd like to be nice. It's if we're if we're not good, it's not good enough. Right? It's unacceptable to not be the best. Right? That's when you you that it's not acceptable to not be the best.
[00:37:25] Nikola Mrksic: Yep. Yep. Yep.
[00:37:26] Steven Fine: And I think that's why, you know, it's to to them, like, some people were willing to adopt, you know, two years ago. Yeah. To them, they've started to adopt en masse Yeah. Because of the personalization, the quality of this. And, look, frankly, that we're lucky that they're choosing us. We're grateful for all of the Yeah. The work that they do. But I think it it goes to like, testament to when you get the, like Well,
[00:37:50] Nikola Mrksic: they're choosing us because we're the best, but it's as simple as that. And we've worked really hard to win their business when technology was at the cusp. Right? And now, like, that has, like, downstream effects.
[00:38:02] Steven Fine: Exactly.
[00:38:02] Nikola Mrksic: When you when you think of kind of, like, older school verticals, you know, I think that, like, even before the new wave of AI, you you know, you've had, like, IVR systems by the likes of Comcast, AT and T, Verizon. And that's where you kinda see that, you know, the trade off of cutting cost to experience is heavily skewed towards cost. Whereas for these guys, it means lost revenue. It means I think Caesars were you know, they swore that they would not use this when we first met them. Yeah. Here we are.
[00:38:29] Steven Fine: I I literally remember going through some of the customers. I would name the customers. They were going through their IVR menu and restaurant customers, and it would like, to make a booking, and you'd go click 3. You'd hit number 3, and it would go, please go online to this link. And you were just like, come on. There's like, we are there's so much that you could be doing to capture
[00:38:53] Nikola Mrksic: Yeah. Yeah. Yeah. They said deflection. These guys are successfully deflecting revenue.
[00:38:56] Steven Fine: And and, I mean, the best thing is the data that comes off the back of that is 100% Yep. Automation containment. Like Yep. Yep. Like, it's not about it's not like, with restaurants, you're not saving, like, dollars and cents because the person is there working. Yep. Right? Yep. Yep. The person is there working, but you're missing revenue. And that's always been, like, the big point behind this is, like, if you take if you automate all these reservation calls, and we'll typically see 60 to 70% of all calls Yep. Automated, we book at 90% Yep. 90 plus percent.
[00:39:31] Nikola Mrksic: Why would you not? Yeah.
[00:39:32] Steven Fine: Would you not? I mean,
[00:39:34] Nikola Mrksic: to me, like, it was always I thought we were, you know, in banal terms, just selling people the convenience of not picking up the phone. And I think that's a much loved feature of it because if you're in a restaurant, if you're pouring pints or seating people, and then, you know, the feeling especially in, like, these large subverticals we have, restaurants, hotels, you come to check-in to a hotel, and the person's there on the phone talking, and you're standing there after, you know, an eleven hour flight, jet lagged the cell, and you're like, please give me give me give me that the room card so I can be out of your life.
[00:40:06] Steven Fine: But I think I got you know, there's always this like, the one side, yeah, you you do the revenue. But it was an amazing story that a customer said told yesterday that at our at the off-site was when we implemented Poly, it was the first time that our general managers at our restaurants could sit down with their team and have a meeting.
[00:40:28] Nikola Mrksic: Oh, like the one with the bear.
[00:40:30] Steven Fine: Yeah. Exactly. Like, without being interrupted. Right? They were always it was, like, it's a it's a quality of life.
[00:40:37] Nikola Mrksic: Noise. Yeah. I just like, you've spent more time in restaurants than I have, but, like, that sound, it's just always there in the background. It's so grating. It was like, I swear to God, we can't quantify this one. Maybe one day we will. But, like, the impact on employee churn of that phone line just ringing.
[00:40:53] Steven Fine: They don't want to they don't want to deal with it. Like, they don't want to. It's not like it's not So it's not in the in their job description. But
[00:41:01] Nikola Mrksic: Yeah. And it's nagging. Like, they're like, you know you should, but then, like, you should do something else. And yeah. Yeah. Okay. How do you feel with the Hogan Ramsay campaign?
[00:41:09] Steven Fine: Grateful that he trusts us to to represent his brand. That's a relief.
[00:41:15] Nikola Mrksic: Thank you, Gordon.
[00:41:16] Steven Fine: Yeah. Like, I mean, we've done amazing work with with his restaurants. And I think, like, the fact that he's willing to put someone with such style standards, he's willing to put his name behind us, I mean I mean, you were with him Yep. On the day. Yep. I mean,
[00:41:30] Nikola Mrksic: any Yeah. It was, like, one of the more well, it might have been the most fun day of PolyAI to date for me, where, you know, as we recorded the whole thing, and I spoke with him. We made some tweaks, like, worked with him, his team. Like, he is the most genuine celebrity I've ever met. He is exactly what you see on the screen, in and out, like, lovely guy, but also great banter. I think, you know, the, like, the perplexity of his language model is phenomenal. Right? Where I think, like, one point of the the whole thing, he looks at, like, he calls this restaurant that has an IVR system, gets angry, and then he we'll link it in the description. It it's the most viewed ads of the Super Bowl. It's got, I think, well over double of the open and anthropic ads views put together even though the anthropic one was really good. More people might have seen it on live TV because they would have paid for more, but we don't have data on this, so we'll claim victory. And what's really interesting is, you know, he's recording. He calls this cafe called Cafe Eco. It doesn't work, and he you just see him turn this. He goes like, cafe cafe shithole. Hey. That's really this thing in. And we're just like, it's phenomenal. Like, you know, I don't think we've produced the whole, like, kind of, like, uncut version of, like, all the phenomenal things, but I swear to god, he, like, produced double the the memes. Yeah. And kinda, like, off script moments. And that's it was a real privilege. I find what like, you know, my wife and I have watched probably every season of Hell's Kitchen. That's our, like, guilty pleasure. And, you know, I find it very therapeutic to just, like, watch him yell at people. Because it's it's for good reason every time, and he just does what I think anyone who wants to manage really high standards and doesn't care I don't think that he doesn't care about people's feelings, but he's willing to put that aside for the sake of quality, which is really kinda like what we what we try to do as well.
[00:43:20] Steven Fine: Yeah. I I mean, I think he represents the brand. I mean, I also I remember my uncle telling me a very nice story that relates to this is I mean, he's so harsh in the, like Yep. In the pursuits of perfection, But he also like, the way he pushes that on people Yep. Is actually it's it's it's a privilege to the people that he does it to. It might come off harsh, but it gets the best
[00:43:44] Nikola Mrksic: Oh, no. And he cares, and he's, like, a really good sport, and he, like, yeah. Yeah. No. I mean, you can he cares about the people around and people he works for. He's, like, really funny. So, like, you know, like, he is, I think, the complete package. He's the real deal. Yeah. Cool. Where else does does kinda, like, AI in hospitality go? Anything else that we can kinda wrap up with?
[00:44:03] Steven Fine: Yeah. So I think, look, where this is is going, we've spoken about all the integrations and how that all fits together, and I think that is really where you like, I meant say the the words again because they keep kind of coming. It becomes the revenue engine, and it's the engine that makes sure you never miss another revenue call. That's what we, like, prioritize, simplicity and making sure you never miss another revenue call. I think where you start getting it becomes super, super interesting is when you start finding a true way to connect all the systems that restaurants work with. Right? Like, today, they they're fragmented, and we do a lot of work to bring those each together as much as possible. But when you start having a when we you're connected with all of them from your loyalty system to your POS to your table management, you start, like, having a full full view picture of the customer and what exactly what they want, and that's when you start going, like, I'd rather have people call in than do anything else.
[00:45:07] Nikola Mrksic: Yeah. Yep. Yep. Yep. Awesome. Well, Steve, thank you for joining me today. Please like, share, subscribe, and we'll see you in the next one.
[00:45:15] Steven Fine: Thank you.