Link to Episode HERE.

00:00 Introduction to DYNE and Its Mission

02:43 The Role of Technology in Restaurants

05:53 AI and Marketing Automation for Restaurants

09:03 Demand Forecasting and Inventory Management

11:50 Personalization and Customer Experience

14:47 The Future of Restaurants and AI

18:09 Challenges and Opportunities in the Restaurant Industry

20:44 Conclusion and Future Outlook

“Most restaurant owners are bleeding money, and they don’t even know it. In today’s episode, I sit down with Parsa Riahi, the CTO of DYNE, to talk about the AI tool that’s quietly revolutionizing restaurants across North America. We’re talking automated marketing, demand forecasting, and competitor analysis…things that used to take a full-time team, now handled by what he calls your ‘digital employee.’ If you’ve ever scribbled prep notes on a grease-stained legal pad and thought, ‘There has to be a better way’ then this episode is for you.”

Expect to Learn:

• Why most restaurants are overspending on tech that doesn’t deliver
• How AI is quietly reshaping restaurant marketing and operations
• The real reason mom-and-pops are losing to chains—and how to flip the script
• What happens when you give a small restaurant big-league forecasting tools
• How to turn slow Tuesday mornings into Friday night rushes using demand trends
• Why “every hour is a happy hour” might be more than just a slogan

Links:

⁠Dyne.app⁠ 

Service starts now.

I talk mostly to people in and around the service industry space. I’m looking to hear from the people I wish I could have talked to when I was coming up in restaurants. Said another way: I am trying to make sense of this wild, beautiful mess of a life, and help others that are feeling similarly confused and/or lost. You can find more of my work at ⁠⁠my blog⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠, and all my social links are at the bottom of that page.

Classic Episodes You May Like:

-#3:⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠My 1st and Most Powerful Conversation with Shane Alexander⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

-#10:⁠⁠⁠⁠⁠⁠⁠⁠Nat Harry, cocktail expert!⁠⁠⁠⁠⁠⁠⁠⁠

-#14:⁠⁠ ⁠⁠Dr. Shalini Bahl, mindful marketing⁠⁠⁠⁠

As always, I’m just here taking notes, trying to figure out what it all means.

Cheers

Andrew (00:02.616)

All right. Thanks for joining me today, Parsa. Yeah, so today I’m talking to Parsa Riyahi, CTO of Dime. And I’m really excited to ask a couple questions, dig in there and find out what they’re all about. But I wanted to start this off by asking, there’s so many different industries that I’m sure that your talents could be good for. Why restaurants?

Great to be with you, Andrew.

Parsa (00:28.45)

You want to know the investor answer or the fun answer?

Do both, yeah.

The investor answer is that, you when we were in college, we were on the front line of the interns who were laid off during the pandemic. And we wanted to give back to the industry and restaurants was a place that was near and dear to us in our university. And we took our local mom and pops and pubs and bars from 200 a day to 2000 overnight with some simple AI automations that were the V1 of what became dying today. But the fun answer is that my co-founder and I,

really wanted a place where we can go and meet pretty women. And that just happened to be the bars and restaurants that were on our campus. And so when those things shut down, we said, well, how can we, you know, do what young men do? And it just happened to be through a company that now is helping restaurants automate their marketing and do demand forecasting and competitive intelligence across North America.

Sure, sure. Did you have a lot of experience at restaurants before this or was it just

Parsa (01:22.124)

I mean, my first job was in restaurants and my family owns a nice little Italian place in Seattle. So some background there, but you know, I’m a tech guy at heart and I always thought that, you know, tech lives kind of in the world of bits and not an atom. So it’s all about, you know, what are you doing in the software page and how do you actually create an impact in the real world? And my goal across this company has really been how do you create technology that helps people to better appreciate, you know,

the nuts and bolts of these things, the actual hospitality of the industry and creating meals for people and not just doing back office chores and tasks.

Yeah, for sure. And it’s such a industry that could use it. And that’s why I wanted to start by asking it’s, you know, I’ve worked at a restaurant right now and it’s a lot of like grease stained notebooks. It’s a lot of handwritten prep sheets. It’s a lot of, a lot of humans at every touch point. And it’s almost designed to be that way. But I think you envision a world where it’s a little different.

I mean, no, think what you want is that people are feeling satisfied by the work that they’re doing. And so if you’re doing the same sort of repetitive task every day without the chance for growth in your industry, I think that’s challenging. And I think most people who work in the restaurant industry are looking always to be closer to the customer and less close to writing stuff down like a piece of paper and stuff that’s not really necessary to really get the most out of that kind of experience. And so as we look towards sort of the modernization of the POS landscape and the push towards, let’s say,

you know, every restaurant having a social brand presence in these places, we came in and said, well, what are the sort of biggest lifts that we can take off of a restaurateur’s plate as it goes from one location to 10 to a hundred as they scale up their marketing strategies, they adapt into new technology systems. You the average restaurateur is spending thousands of dollars a month on technology solutions that are all kind of fragmented. We came and said, how can we create a solution that costs just a couple hundred bucks instead of a couple thousand that gives them their money’s worth within a few days and set of weeks? And then

Parsa (03:16.16)

also drives more traffic for them so they get more butts in seats. And that is sort of the approach we took that enables them to actually grow faster into new locations and actually hire more people instead of trying to take away jobs or anything like that.

cool. Cool. Yeah, because that is kind of the fear mongering side of it. Everyone’s saying it’s taking our jobs. It’s gonna make us all obsolete. yeah, I guess probably the best way to attack is what what do you do? You’ve said a couple times that automates marketing gets butts in seats. But like, like you come in and say I hired you and you’re starting in my restaurant, what would I expect to happen?

Definitely. So if you’re a mom and pop restaurant, what we’re going to do is create the posts, the advertisements, the review responses using our system based on all of your insights about your brand, your sales trends, what nearby competitors are around you and how they’re launching different marketing campaigns, what events are around you, whether it’s a Taylor Swift concert or it’s a standup comedy show or it’s a sports match. All of these things are taken into account to create that optimal marketing strategy that creates posts and advertisements across

Facebook, Instagram, Google creates the content for your email and SMS and does that all through our system plugged into your account so that you don’t have to lift a finger. You just have to say, ahead and drive more traffic or increase my brand awareness or improve my total revenue. And because we plug into the POS system, especially for sort of larger customers who have two plus locations, they plug it into the POS systems and they see the attribution of

of those marketing stats into not just clicks and impressions, but actual dollars in their store, the actual order volume increasing and giving them that attribution from marketing to sales so that they don’t have to hire an agency and spend thousands of bucks a month on that agency to give mediocre results. can see it directly in their P &L statements that are generated by our system and actually create suggestions on the optimal campaign to drive revenue and not just clicks.

Andrew (05:16.846)

Okay. And then you also mentioned revenue forecasting. So is that like, do you take over the inventory side of things, the ordering or?

We’ll do a step above what a traditional POS system does, where they look at just what the historical sales are and try to project those out. It’s like saying if it rained last week, is it going to rain again this week? It’s sort of like a coin toss. What we’re doing is layering in lot of these, what we call future looking data sets, things like foot traffic patterns. So we get aggregated information about when people are walking around with their cell phones on the same stuff that you see in a Google Maps heat map influences things like tourism, traffic, things that are telling you how many people coming into your store.

next week, next month. We look at all the different events that are nearby to you that are upcoming, whether it’s the, you know, NFL season games, or it’s the things that aren’t showing up on the big radars that are influencing people to go nearby to your locations or drawing people away from those locations, as well as all of those things that your competitors are running, whether it’s a similar menu offering, whether it’s a campaign that’s drawing people in the same demographic profile of your customer base. And then also layering that in with things like weather and census patterns and demographic data to create this holistic

insight about not just what’s happening inside of your store, but what’s happening in all of your competitor’s store as well as in the city surrounding you. And doing that for not just your one location, but across every location of your chain to say, you know, if I want to look three months out into the future, what’s my expected demand on the Tuesday in the morning shift and how can I adjust my pricing? How can I create the right marketing strategy to turn my, you know, Tuesday hours into a Friday rush and do that all without having to actually read through all these different data points, but let the AI agents, as we call them, find those optimal campaigns, find those top

revenue opportunities that you can click on and create the right marketing strategy or pricing strategy that draws people into your store and gives you that consistent revenue across the board.

Andrew (07:04.076)

Hmm. See, when I’m hearing all this, think, isn’t this too good to be true? Like, can everyone use this?

Well, know, it’s as much a tool as anything else. You know, when you like see a surfboard, not going to immediately know how to surf when you go out into the ocean. Similar to this, you you want to have to, you sort of have to learn how to use the AI. And what we’re always trying to do with our system is make it easier and easier to use. you know, our customers love us partly because it’s, works a little bit because it takes about five clicks to get set up. And so yes, there’s a point about learning what your patterns are and learning your brand guidelines, learning your habits. And what we’ve been able to do, especially in that marketing tier is,

reduce that time down to about a week. So from the time you first get started in our system to the week that follows, you’ll have your first campaign running. It’ll be seeing how it’s working in terms of driving more dollars into your business. And it’ll be actually optimizing in the backend about what’s working and what’s not. So we can find what’s the actual optimal strategy that works best for your restaurant, for your brand. And so as it scales to more and more locations,

we have these things called agents that run in the background so that you as a user don’t have to go and actually look into each individual piece of information across all those stores, but it finds that best strategy that goes across your multi-location chain and gives you the ability to actually drive that campaign that goes and says, who are the top competitors in each one of my selected regions? What are the top types of events that my customers care most about? And that kind of create the let’s say, special menu promotion that turns every hour into a happy hour for my business.

like that. It’s like, I wonder that might be your like tagline there. Every hour is a happy hour now. so okay, for the uninitiated, you know, I’m a big Harry Potter fan. So I’ll say muggle out there. Machine learning, artificial intelligence, we got the large language models, where does this all fit in? Like, imagine I’m very simple. And just like, help me understand this.

Parsa (09:06.432)

AI is kind of a catch-all term. What we do is touch on both of the pieces you mentioned, the LLMs and the machine learning. So all of those campaign creator stuff, the stuff that spits out text from the inputs that it gets is the LLM stuff, the large language models. And all the stuff that is machine learning is like your demand forecasting. We call it time series forecasting for the technically non-muggle folks. But all that means is you’re saying, have a set of sales data from the past and I’m trying to see how that trend goes up in the future. It’s like a stock graph, but we’re doing that for restaurant sales.

and layering in those insights about not just what the past history of the stock is, but what are all the market conditions that make that stock go up and down. In this case, instead of external financial reports, it’s about what are your competitors launching in menu promotions? Where are people interested in going tourism-wise, event-wise? What are the weather patterns that influencing the trends? What are the social media hashtags that are trending in different food items that influence our expected sales for a given menu item?

OK, and so I do wonder with the machine learning thing, if you’re just projecting, how do you account for those like aberrations, those crazy events that happened like 9-11’s, like COVID? I guess no one did better predicting it not using the models and using the model.

Yeah. If you look at any model that was created during COVID, they all failed. And so there is some point of fact that global calamities are quite difficult to measure. But what is sort of unique about our system is that it’s trained on billions of data points from across the industry so that the average mom and pop operator, maybe they haven’t seen, you know, as many Friday rushes as the people who are in the more sort of, you know, city centers. So maybe they, people in the city centers haven’t seen, let’s say the rush that comes from the hometown team.

coming into like the, let’s say the NFL finals, for example, these kinds of trends that we see that build over time that are sort of gradual in some state are what we are able to actually see most that, you know, it’ll be different from year to year from season to season, but it’ll be in the right direction going up or down from day to day or from hour to hour. And it’s those places where you can see that gradual shift in the trend that we adopt for and we account for in creating those campaigns so that let’s say there’s, you know, a great, a growing interest in

Parsa (11:19.978)

know, matcha tea items over a given week. We’ll see that item take into account across different competitors that are running those new matcha offerings and say, you know, here’s the average price of the matcha offering that’s in place with your competitors. Here’s what you could offer based on your menu offerings and the demographics that you serve. And let’s see if we can drive a campaign to drive people into your store for that matcha, you know, coffee item for example.

Cool, yeah, so it is really more just like, I guess it’s probably the wrong way to look at the most extreme examples first. It’s more just those trends that are hard to tease out if you’re human.

If you had a thousand hours a day, you could definitely do what the AI does. The tricky part is what do spend your time on, right? And so if we can synthesize the most important, I did some information down into one key system, that’s exactly what we’re going to be focused on so that we can actually drive more traffic for the store on every average case as well as most of the edge cases. But of course, adjusting for a COVID, it’s quite difficult, but basically everything but the COVID you’ll be ready for.

Yeah, yeah, the home game, those sort of events that are more reasonable, that actually repeat probably. I did want to ask, so this is something I hear a lot of, and I’ve never really understood the whole like, what’s the point? I’ve heard a lot of people talking about how training the systems and like personal identifying information is such a big thing. And I’ve never had someone who’s in the AI space that I could ask like, why is that such a big deal?

Is it that you would actually leak that information through a system you trained inadvertently?

Parsa (12:56.75)

First off, we as a company never train our models with customer information that gets sent to other customers. So you will have your own model for each one of your stores and each one of your use cases so that it’s fine tuned to the use case. Now, to your point about why do we do this, imagine I have a thousand customers and they all have the same AI running for them. We’re launching a campaign about coffee. Well, the wording is going be almost the same across all those customers. And then it’s kind of like a race to the bottom. It doesn’t really work that well. But if I know that one person sells

coffee with a particular barista that creates the art in a way that looks really nice. Or one person that has a cafe that is situated with a nice view of the water. These kinds of nuances that are distinct to your business that aren’t going to be appearing in a general coffee campaign or what we fine tune for and actually train the model on to create it in a way that is following your brand guidelines, following the brand voice of your customers. We’ll pull things like the reviews that every restaurant gets and we’ll see what are the keywords that people are mentioning most about.

a given stores items, know, are there more positive sentiment about pizza or are there some issues about the service? And then we’ll use that to actually create the campaigns to tune on what customers are caring most about so that it is not just fine tune for the particular brand, but down to the individual location by location level, what are the differences demographic wise, interest wise and sales wise that we can actually create the optimal strategy for? And we run that for all of our customers, you know, in clicks instead of in weeks and time and research.

Okay, so I guess I was thinking of it as something that was more of a security issue where it sounds like what you’re saying it’s more of a personalization issue.

There’s definitely both pieces. So we saw security by having dedicated models for each customer and we saw personalization by having dedicated customers for you, dedicated models for each customer. So it’s sort of like a two birds one stone.

Andrew (14:42.348)

Got it, got it. Okay. Yeah, I just, I’d always wondered, cause I figured it had to do something with that. I just, you never have someone you can ask that question to. And then, yeah. So if I have a coffee shop and at my coffee shop, Fred’s really good at foam, then my ads are going to be like that or something like that. Whereas, you know, you don’t want everyone else advertising the foam if no one else is doing it or something like that.

Of course, happy to answer it.

Parsa (15:10.21)

Be able to stand out from your competitors and not have them stand out from you. And so if you have a business that has a really strong, you know, strategy, whether it’s, you know, relating to a particular item you sell, or it’s some sort of key thing that you guys do that no one else does, that should be brought to the forefront in the way that you use AI and doing that through the regular chat. You just won’t happen because it doesn’t have that insight about the last five, 10 years of your, of your business information. They can only process a little tiny context from.

Okay, got it. In your opinion, what what is it like? Imagine it’s 2035? What does the restaurant look like? Like is every restaurant doing some version of what you’re doing here? Like me you go to restaurant? What do we see? What do we experience?

I think in the past restaurants were one type of thing that everyone sort of struggled. And now it seems like there’s a lot of different fragmenting paths. So if you see, let’s say, know, Travis Kalanick’s company, Cloud Kitchens, is trying to make more of a delivery based only offering. Or if you see a lot of folks who’ve heading into the luxury restaurant space, a lot of players are trying to sort of you know, cannibalize QSR. And then there’s of course the mom and pops that will always be riding the different ways of the industry. What you kind of want as a

industry strategies, what is a solution that can align all those different incentives towards what is not just best for the PE firms to buy these QSRs or for the big company CEOs who don’t pay the Uber drivers and now will probably not pay the restaurant operators in these cases, but how do you actually make sure that the business owner, the small business owner is benefiting from these systems? so as opposed to seeing like, does the industry look like? I sort of think in the, do you give the small business owner the tools to build up the industry that they want to see?

rather than let the big companies define the path for them. And so what I find AI useful as is it’s kind of that great leveler where it allows for mom and pop to compete with the large Starbucks chain or for the average non-clinical person to quickly get set up in those places where they’re not necessarily as experienced so they can spend more time with the customer doing things they actually love and less time on the boring back office stuff. And so in that, you sort of build like a grounds up solution, grassroots solution.

Parsa (17:20.013)

enables the small business owners to actually create the industry they’re looking for rather than waiting for the initiative and where they are going to be. And so the folks who jump onto these tools and get ahead of the industry trends are the ones who are going to be able to survive and thrive in this space and get those double digit profit margins and everything that everyone’s looking for. But the ones who sort of stay back and wait for the industry to catch up to them, it’s going to be too little too late, I think.

Okay. Yeah, because I like what you said before, and I think it applies here, where if you spent those thousand hours, you could do what the AI does. It’s just, do you want to spend those thousand hours? And in the past, if you had a mom and pop store, you were just spending all of your time working. And the spare time you were marketing, and the spare time you were ordering and hopefully projecting correctly, and now you can take some of that load off. Okay. I like that. Yeah. Because, you know, right now I work at a restaurant that

is it’s a group of about 100 altogether. And yeah, I can only imagine if I was also running the marketing in store and also I had to set up all the inventory systems and you like you add all these different things, eventually it’s like, there’s only 24 hours of the day. So

Well, I’d definitely be curious to learn more about the work you’re doing in your business. And, you know, if there’s places that we can support, be more than happy to offer my services.

Sure, yeah. So I’m guessing that this is a service that’s so useful. Does anyone ever start and then leave? Or is it pretty much like you start and you realize, we keep this forever?

Parsa (18:51.852)

It’s one of those things that it gets better over time. And as part of our sort of product philosophy is I want to make sure that you can spend less time on my product and not more. We’re not trying to get a social network that’s like Facebook, you get stuck doom scrolling forever. The goal is how can I get my system set up so that I have this digital employee that’s running my back office for me. And that’s sort of the way that we framed it. So why would you fire an employee who costs one 10th of minimum wage that’s creating thousands of bucks a month in revenue for you?

it doesn’t really happen. And of course, if there’s a reason that it does happen, and maybe that we have, let’s say built into the right POS integration for them so far, there’s some new system that they want to plug into that we have to build for. And of course, we always build those things rather quickly. But generally our customers love us and it sort of shows up in our internal metrics too, where we see it’s super easy to use, it’s really sticky, it’s growing astronomically fast. And we just signed a…

a major deal across the US with Integra, the big GPO that services over a hundred thousand restaurants in the States. And we’re doing pilots for a few thousand of those in the year to come. you we’re seeing a lot of industry traction with it across, you know, not just mom and pops, but across mid markets, across enterprises. And so we’re quite excited about that and quite flattered by that. And we’re going to do our best to make sure that the trust, the industry places in us is well-founded.

Yeah, but is your goal to sell it one day or what’s the end goal there?

I mean, you know, I think every, every founder of every company says, I want to get a big check. my, um, it’s hard to say no to a big check, but I think my, my, guiding philosophy is how can I spend my hours to create products and solutions for, for, um, the largest group of customers who are going to create the most amount of impact for, um, know, call society, call it the economy. So I think like when you think of a small business, it’s that most underserved part of the, U S and Canadian economies that

Andrew (20:26.254)

I wouldn’t say no to a big check, yeah.

Parsa (20:48.718)

really isn’t touched by a lot of solutions in a great way. And you often have to subsidize the way to grow those places. so creating tools that can enable that type of market to grow independently of any foreign intervention seems like the way to ensure that we give back sort of agency to sort of mom and pops to business owners to drive their outcomes. And so if I can best do that through, you know, a sale of the company to a bigger player like a

call it a toaster or square, can scale that up to the whole industry. That’s great. And if we can scale it on our own to a place where it’s moving faster and faster to create that outcome, then that’s great too. But sort of what’s best for the customer first and let me come second.

is there anything that AI can’t do? Is there something that you’ve run up as a problem? It’s like, we’ve tried this, but that HR and restaurants just, it doesn’t work or, you know, something like that or.

I I think we’ve tried to have AI find us love in these things as sort of the edges of the extremes. In the business world, think of it like, it’s not that it can’t do things, it ruins something. So when you have an AI take your order as a customer, I don’t get to talk to a person, I talk to a robot and it has like that kind of lack of hospitality. So I try to find places where, where can you bring the AI in that doesn’t ruin the experience but actually augments it. And so although it’s a tool that could do anything if it really wanted to,

it’s not that we should let it do that and should actually promote that kind of idea. And so we always think about what are those lifts that are things that people don’t enjoy doing that are important to do that are the critical drivers of business for the industry, but that also are sort of designed by programmers and can be automated by programmers. Whereas the rest of our experience, know, the of the world’s oldest industry has been around for thousands and thousands of years. Why would an AI be the thing to solve the customer experience? Whereas marketing social media is like, it’s been around for 20 years.

Parsa (22:42.626)

better than an AI tool, you can actually solve that so that can stay back to building in your community, having local experiences, evolved to actually be best at.

Yeah. And also, you know, anyone who’s done any serious social media, not for fun, but for business knows it’s kind of a pain. It’s in that category of like, like, I know I should be doing this. And, know, I kind of put it off till Tuesday afternoon. And then you’re like, I guess I have to do it finally. fair enough. Okay. Well, cool. You know, I think that I hit actually all the questions I had, is there anything major that I may have left out or?

Well, you know, I’m always happy to chat with your listeners if they have any questions so they can reach out to me and we can set up a timeout. just got to go to dynap.ca, D-Y-N-A-A-P-P.ca and we can set up a call. But you know, what I’m most curious to learn about is sort of where you guys are in finding AI in your business. Maybe sort of give like a tactile sending off note for the listeners here and how…

you are sort of thinking about your answer as you’re from a business owner and restaurateur perspective, as opposed to for me telling you what to do. I’d to learn about why you’re actually interested in where you’re seeing this go.

sure. I mean, the largest bulk of the time that I spend, because I’m the general manager, direct with other managers, it’s staffing issues, organization and ordering. You know, those are the big ones and restaurant industry ordering will kill you because you have a product that is perishable and you order too much. It goes bad. If you order too little, you can’t sell it. So you kind of always have to have this constant pressure to get it just right and no more, no less.

Andrew (24:24.851)

So yeah, probably ordering, I think, would be the one that would fall into.

Yeah, we’ve solved a bit of that with Integra. So they do procurement and what they want to know most is like, how can I use the demand trends at DYNE to better suggest what items that my customer should order so that they are actually staying ahead of that. And there’s this idea called the bullwhip effect. So every 5 % change in your demand causes a 10 % change in your restaurant level supply you have to have to be ready for those ups and downs. It’s a 20 % level for the integrals of the world who providing that to the restaurants.

when you can reduce that 5 % error margin to a four or 3%, that’s a three or 4 % decrease in the inventory costs that a restaurant has to pay. And that’s the difference between, as you probably know, profitability and non-profitability, especially when it comes to perishable items. And so that’s one of the places where we’re always really keen about is how can you ensure that you have that expected stock forecasting and demand forecasting that allows for you to create the right pricing strategy to let’s say keep your

top selling items selling in the right way, or maybe bring attention to items that aren’t gonna be running out as soon and get those off the shelves and all those places where you have those sort of things that you do as a GM every day, but maybe not in a way that is automated and streamlined to optimize for profit margins.

Oh yeah, we are pencil, paper, and Excel. No joke. Yeah, which most of the restaurant industry is, which is probably why you are doing so well, honestly.

Parsa (25:48.885)

I hear you.

Parsa (25:54.648)

Well, I’d love to learn more about that after this call too and see where we can.

Awesome. Yeah. And for everyone listening, I will have the link to DYNE in the show notes. Make sure that everyone can book a call. Yeah, thank you for taking a moment. Talk to

Thank you so much.