Guardrails, Costs, and Real ROI: Practical AI in Restaurant Operations

Tal Clark: Welcome back to the Instant Payments Podcast. I’m your host, Tal Clark, CEO of Instant Financial, and we’re back for part two of our conversation with Marcus Wasdin, a leading voice in restaurant technology leadership. Marcus, let’s jump right back in, and I wanna circle back to what we were sorta wrapping up with the last time.

Talk more about what you developed when you were at Par in regards to AI. Coach AI, I believe it was

Marcus Wasdin: Yeah, no, yeah, absolutely. So, it was the team that actually… I happened to be there leading the group when they launched it. There’s a whole team that, that did so much work and effort on this particular product, and were trying to drive it forward, and I happened to be the lucky guy that was there when we took it to market and trying to figure out, what’s working, what’s not working, and how do we make those tunes and tweaks.

And so, what I would say is it was a super exciting product, right? being able to give the power of natural [00:01:00] language to people that are businesspeople, to ask the questions that they want to have answered, without having to know how to build a report or, do all the super, technical, stuff that you have to do typically to bring a lot of this data together.

‘Cause sometimes these, to get the answer to the questions that the operators want to know, it’s gotta touch multiple things, right? And so having that power at their fingertips is just great. And so, I’d already mentioned that one use case, I think, on the last episode that we had, which was, analyzing operating hours and optimizing for profit.

We had some folks that did that, and they actually had some real returns out of that, which is, could they have done that without that product? Sure. But could that operator have gotten the attention of someone in their FP&A group to do all that analysis for them? Probably not, right?

‘Cause the FP&A group’s probably pretty busy. So, so pushing that ability down into the, operational organization, I think is a pretty key thing. And so that was one of the big learnings that we had. [00:02:00] One of the other things I would say is a learning that we had as well is that we w- you know, one thing that most people, I think, you know that, that mess around with AI at all, whether you’re using, ChatGPT or Claude or whatever, we all know that it will sometimes tell you the exact wrong thing very confidently, right?

It has got all the confidence in the world, and you read it, and you’re like, “That is not right.” 

Tal Clark: That’s right. 

Marcus Wasdin: And so that’s commonly called a hallucination, right? they put those hallucination- hallucinations out there. But when you’re delivering a, commercial product to someone that’s gonna make business decisions based on it, then you have to minimize those hallucinations.

You can’t have it confidently telling a restaurant operator to make some changes. They make the changes, and it’s actually detrimental to the business, right? And so there was a lot of work that we had to put around that part of the product. And sometimes we over-rotated on those guardrails. So I talk about putting up guardrails, and if you over-rotate on the guardrails, it won’t [00:03:00] hallucinate, but then it becomes less user-friendly, right?

‘Cause you and I are used to just typing stuff in the ChatGPT or Claude, and it’ll just make all kinds of assumptions and then give you an answer. Well, ours wouldn’t make the assumptions, right? We wanted you to give it the specific details of your situation so that it could give you a right answer as opposed to making assumptions which could lead it to giving you bad data.

Tal Clark: Yeah. Yep.

Marcus Wasdin: And so but you– that’s a fine line, right? And then the other thing that I would say is, AI’s still a little squishy. And what I mean by that is most of the other systems like, if you did a query on a database or you did whatever, you did the same query over and over again, you’re gonna get the exact same result, right?

Go to AI. You can try it yourself. Ask it the same question two or three times in a row, and you will get a slightly different answer every time, right? And so if, you try to deploy AI into something that is very precise and specific, like restaurant operations or HR operations or something like that, where consistency is key, that’s another problem that we’ve [00:04:00] got to figure out how to overcome and overcome it in a way that’s still very user-friendly, right?

So that was one of those lessons. It’s “Hey, we’ve put up the guardrails, but they can’t be so high that now this thing becomes unusable because it’s like pinging me for every single response.” So, it’s– it was a cool, process to go through, and I think they did a good job of landing on a product and a solution that is gonna be super, super useful to the operators.

Tal Clark: Well, that’s really good. And then, I– the, other, one I wanted to go back to, because I think it’s one that people think about, as it relates to both AI and employees, and that is the, that’s the voice AI at the drive-through example you gave earlier, and you mentioned the cost of that.

So I wanna, I, s-tell that, let’s talk about that a little bit more in regards. My question is, I guess, if I’m understanding you correctly, is that the cost of the voice AI technology is, too high for an operator to actually adopt and [00:05:00] use it in a way that allows them to maintain the margins they need to.

Is that right?

Marcus Wasdin: Yeah. So what I would say is, if you look out there, there’s a couple problems with voice AI when you’re interacting with a guest. And so problem number one is, it’s a great use case, but the cost of delivering it, if you just leverage a commercial AI engine to answer all of those queries, is through the roof.

And a lot of the, a lot of the players that are out there too have a human in the loop. So what I mean by that is they’re a voice AI company, it is an AI solution, but if it stumbles or falters, they also have a call center that they’re having to pay for.

Tal Clark: Oh, wow.

Marcus Wasdin: That can step in and take over that, conversation so that the guest has a decent experience.

Tal Clark: That call center, is that call center on the quick service restaurant side or is it on the technology provider side?

Marcus Wasdin: Typically on the technology provider side, right? So they’ll offer you a solution that says, “Hey, we are 90% [00:06:00] effective,” which sounds great, but is nowhere near good enough. 

Tal Clark: Yeah. 

Marcus Wasdin: And because that still leaves 10% to, for people to stumble, and then they have that human in the loop on the provider side that will then step in, right?

So it’s an extra cost of goods that you’re trying to recover when you’re selling this solution. And so I think the, the people that are gonna win in that particular area are the ones that are gonna figure out how to lower that cost of goods. So ’cause if you think about, dude, think about a 14-hour day, unless the restaurant’s open 24 hours, and every single utterance that happens at a drive-through day in and day out, right?

That’s a lot of data going up and down.

Tal Clark: Yeah, no doubt.

Marcus Wasdin: And so if you’re shipping all of that out to some, commercial AI model to get a response, your token burn is gonna be massive, and so that’s what’s driving that cost of goods up. The guys that I’m seeing that are doing a good job a-and the one that I’m working with, they are taking a lot of that stuff that’s very simple to answer, make it a meal, make it a large, [00:07:00] make it a combo, whatever.

Those simple repetitive things, they’re grabbing that, and they’re processing that inside their own infrastructure and sending an answer back as opposed to having to call a commercial, service where you’re burning tokens, right? And so they’ve been able to bring the cost of goods down to something that’s reasonable that they could then charge the restaurant and still turn a profit and, you’ll be a financially viable company.

So that’s thing number one. Thing number two that’s kinda difficult is, you’re there and, the semantics around it. So I’m gonna pick a… Not that I’ve done business with them or anything, but I’m gonna give you a great example. You’re familiar with Whataburger, down there in Florida.

They’re all over the place. They’re in Atlanta now. everything on their menu, this is just an example, everything on the menu starts with what, right? And so when you’ve got an, an AI that’s listening to that and you say, “I want a Whataburger,” the first response of a typical AI is to say, give you the definition of Whataburger is, right?

It’s thinking you’re asking it what is a burger, right? [00:08:00] And so if you think about other menus, Zaxby’s is another example. Everything starts with a Z, right? a lot of these menus don’t have common English words in them, and so that’s another challenge that these, kind of providers have to overcome, is simple stuff like that you and I take for granted.

Tal Clark: Yep. Yep.

Marcus Wasdin: And so one of the things that, that I think a lot of people ask and are worried about, solutions like that is is it gonna eliminate jobs? Is it going to, take someone’s job at the restaurant? And, my assertion is it’s not gonna take someone’s job. What it’s going to do is it’s going to allow you to take these tasks that are super repetitive and super burdensome and quite frankly, things that really drain the energy of a lot of employees, and allow those team members to now focus on if there’s guests inside the restaurant, let me make sure that they’re having a great experience.

Let me make sure they’re having a great experience when they pull up to the window, when I give them their food or if I’m walking it out. I think it’s really gonna be a force multiplier i- for the people that are in the restaurant so they can actually deliver a better guest experience as opposed to [00:09:00] eliminating, particular roles. Highly repetitive things I think are ripe for AI and take them off the shoulders of those folks.

Tal Clark: That’s interesting because if what you just said ends up being the case, first of all, it sounds like we’re still very early. I mean, we’ve been talking about AI for two or three years, and it sounds like we’re probably still a good bit away from mass adoption or mass use of this. But what you just said, I believe, is that, you’re not gonna replace people, which we– none of us want to see happen, right?

And, you’re not gonna replace people, but from the operator perspective, that means you’re adding cost, right? ‘Cause if you’re paying for AI and you’re keeping the same number of people, you’re adding cost. So the only way that works then is if you’re using what you just said and your guest experience improves to the point or your food improves to the point that your revenue is actually going up. 

Marcus Wasdin: Yep.

Tal Clark: Cause you’re not working on the cost side, you gotta work on the revenue side. So, so it sounds like thought process on this is the way that it’s [00:10:00] successful is that AI ultimately drives better business results for these operators, not at the expense side, but at the revenue side.

Marcus Wasdin: Correct. And so I think there is probably some expense side savings to a certain extent, but I think that the revenue side, especially for the solution I just mentioned, which is that kind of drive-thru piece, one of the things that, that is also interesting if you think about it, is, a lot of restaurants, especially QSR, they have local, funds that they can use for local marketing funds, right?

So, for the folks that are in the restaurant space will know exactly what I’m talking about. Typically, a restaurant brand will charge you some sort of percentage in addition to your royalty to then do national advertising and things like that. But a piece of that comes back to you as the local franchisee or operator to deploy locally around your restaurant.

So they typically sponsor sports teams, or they put up billboards near their place, or they put road signs up, on the highway, like they sponsor that. They do those types of things. But a lot of [00:11:00] times they don’t deploy all that money, and that money’s just sitting there, right? And so one of the ways to potentially offset some of the cost of this thing coming in is could you open up the funnel for those local marketing funds to be used to offset the cost of these types of products that are coming into the restaurant, right?

Because, for example, let’s say you’re a partner with Coca-Cola, right? Coke always wants you to try to upsell a Coke or make it bigger or whatever. They want you to sell more Coke, right? And so, and they, and as they deploy a ton of money to restaurants to do that, right? That’s why the Coke cup is in the picture on the menu and all that kind of fun stuff.

And so you can say, “I want every employee to do that,” but they don’t, dude. Like they… And it’s not because they don’t want to, it’s because they get busy, and they just, they forget to try to sell you a Coke or upsell a Coke. Well, guess what AI doesn’t do? AI never forgets.

Tal Clark: There you go.

Marcus Wasdin: Will always do that, right?

So now there’s a channel that is more consistent, if you will. And if you can make that bridge and potentially get [00:12:00] some of these op- or some of these brands to open up local store marketing dollars that the restaurant’s already paying, because the revenue, the percentage is going out, if they could use that money to offset the cost of a drive-thru QSR, voice thing, I think that’s a big unlock too.

Tal Clark: Well, that’s cool. And that’s, sort of a, a refreshing perspective, I think, because I think the general public hears about AI, they hear about, maybe hear about AI in restaurants, and that’s, the fear of AI, right? It’s gonna replace my job, or it’s gonna replace somebody else’s job.

But, so I think that is, I appreciate your perspective on that, and it’s a refreshing one in that it has the opportunity to improve the business results for the business and the, results for the consumer as well that’s going to that restaurant without the cost of jobs, which is, which is good.

Marcus Wasdin: And look, in a simple way, if I try to distill it, like AI is gonna help with the repetitive tasks that are draining to most people, right? And, what’s gonna become more and more [00:13:00] important with the team members in the restaurant is judgment and how to interpret what’s happening with, if it’s a system giving you data, do you have the scars on your back to say, “That looks right,” or, “That doesn’t look right.”

And, hey, the AI may say something wonky to a customer as they drive around, because that’s gonna happen, and you recover and those types of things. So that guest experience and that judgment and that care with guests is gonna become more and more important as other parts of the business start to get a little more automated.

Tal Clark: Okay. Well, that’s good. Let’s sort of put, a bow on the AI thing. We’ve talked about it quite a bit, but if you were… And let’s do it by if you were sitting down with a restaurant CEO today who was considering beginning their AI journey, what sort of five-minute conversation would you have with them?

What would be your advice for them?

Marcus Wasdin: My advice would be don’t go wide open initially on an AI strategy. Yeah, sure, let’s talk about a global AI strategy for the [00:14:00] business and what that might look like, but don’t put that into action immediately. What I would say, as part of that strategy, grab a specific challenge or problem. Let’s identify that inside the restaurant and, see if the– if we were to resolve this problem, what type of benefit would that mean to your restaurant?

And is this problem also something that can be potentially suited for AI to solve? So go after a specific thing that delivers an outcome. Once you see the outcome be delivered, then you turn– you open the floodgates on de-deploying that particular solution, right? Get that moving through, and then go to the next thing.

What’s the next biggest thing that you can go after? And as you start to get some of those incremental wins, then it gives you the authority and the credibility with your franchisees, ’cause, a lot of times, as a brand CEO, you may have a handful of company-owned restaurants, but what you’re really trying to do is build out systems and processes that franchisees are willing to pay, build more stores for, pay those royalties for, fly your flag for, right?

And so [00:15:00] start to build up the credibility with the smaller wins on AI deployments. Get two or three of those under your belt, then you earn the right to stand in front of the franchisee, advisory council and say, “We’ve been delivering these wins. Now we want to open the strategy up a little bit more, and let’s go.” And so if you can’t deliver those small incremental wins, you don’t necessarily have the right to do that. Now, a good CEO, and there’s a lot of them out there, could convince the franchisee advisory board to do it because there’s so much hype about AI to go wide open on the strategy. But if they go wide open on the strategy and it doesn’t work and they’ve burned a lot of money doing that, then that’s tough to come back from.

Like you lose credibility on that, and they probably won’t ever listen to you again on the matters of AI. So, my counsel to CEOs is find a specific problem that AI can address, resolve that problem, maybe do it again, and once you get that credibility, then you can stand in front of your operators and say, “This is what we’ve done. This is what it’s delivered. Everyone agrees. Let’s now open up the floodgates a little bit more.” [00:16:00] So, baby steps basically is what I would say.

Tal Clark: Okay. Well, that’s good. And then, let’s, let’s– I know one of the, other things too, Marcus, that you do is you’re advising private equity firms, evaluating restaurant acquisitions today. What are investors looking for today that they weren’t looking for five years ago? I mean, I could, we could go back into some of the things we’ve talked about already.

But just briefly, what are they out there looking for? We see a lot of churn in the restaurant industry right now in regards to, I mean, there’s bankruptcies occurring that we’re all aware of out there right?

Marcus Wasdin: Closures…

Tal Clark: there are, restaurants looking for private equity, right? Opportunities. So what are the private equity firms looking for? What do you see out there?

Marcus Wasdin: I think it’s a great question. I, let me back up just a little bit and say, I think there’s a lot of different lanes that these private equity firms are in, and they’re all looking for something a little bit different. what I would say is, a couple [00:17:00] of the items that I like to see is, most private equity firms aren’t necessarily in it to make big investments and turn a, struggling brand around.

So if it’s a, if it’s a transaction where there’s a brand that has massive brand recognition, but maybe they’ve just fallen on hard times or they’ve been stumbling a little bit here lately, the, private equity firm that’s looking at that, they’re looking for is there brand value where I don’t have to establish this new brand?

But if I can fix the food quality and make sure that it’s a great food, kind of experience, and I can fix the overall guest experience piece through… A lot of that is digital today, right? so if I can fix some of those pieces and also deliver consistently on my operations, then I can take what is something that I don’t have to build, which is the brand equity in this particular brand, and I can go and fix the operational components of this and try to revive this brand to get it moving forward, right?

Now, a lot of times, and I experienced this with my, some of my tenure ear- earlier in my career. sometimes these brands get ignored to the point where you’ve just got restaurants that are in trade areas that [00:18:00] just will not support them anymore, right? No matter how much operations you try to fix, no much- no matter how good your food is, there’s just trade areas around that just do not support the brand anymore, and it’s moved on.

And unfortunately, you’ve got to be, pretty, diligent and, facing the truth and getting rid of those so you can focus your resources on the other parts that you’re moving forward. So I think they’re looking for that, right? They’re looking for opportunity to, is there a trade area that, th- is there a basis where I can grow from, right?

I know I’m gonna have some, folks out there that are underperforming, but if I trim those, focus on the others, get the food right, get the ops right, the brand recognition’s already there. Can I get the experience right? Let’s drive that forward. So I think that’s what a lot of guys are looking for today.

And I think what you’re seeing is there’s transactions where, a brand might have been part of a broader portfolio company, right? And so maybe the portfolio company, has, certain resources, and they want to focus on other brands that are doing a little bit better, and [00:19:00] this brand could be doing better, but maybe they just don’t have the time, energy or, dollars to, to spread across all their brands.

And again, I think that’s another thing that, private equity firms are looking for, which is, “Hey, is there a brand out there that may be just getting starved a little bit in its current environment? And if we gave it a little bit of love, right, could we… could, it blossom and move forward again?”

So,

Tal Clark: There you go.

Marcus Wasdin: A couple things there that, I think are interesting. But it- you’re right. There’s a lot of closures out there, right? There’s a lot of, it’s tough, and it goes back to what I said before, man. Looks like a simple business, but it is not. And yeah, it’s so tough, and I wish many times that I, didn’t have the bug for restaurants.

There’s so many other industries that are easier. But, I do have that bug, and I’ll be in it probably un- until I die, and I, I just love the space and trying to figure those problems out.

Tal Clark: Yeah. Well, it’s, and it’s great people in the space, and you’re providing great advice to people. So, look, really appreciate you being on. Before we drop off, I wanna get a couple of rapid-fire questions for you, if [00:20:00] that’s okay. What is one leadership principle you live by?

Marcus Wasdin: Oh gosh, I learned it a long time ago in my consulting life when I first came out. It’s, the two, two-part answer. First is hire for aptitude, train for skill, and reward for performance.

Tal Clark: There you go.

Marcus Wasdin: The three things, and so that kind of goes back to hiring people that are smarter than you.

I want people around me that are smarter than me, and as long as they’re showing that aptitude and that fire in their belly to go learn stuff, you can teach them what you want them to know. And then, when they deliver, reward them, right? That’s, one thing. And then the, second part of that is, trust.

Like building trust to me is a big thing. and trust leads to culture. Like you can’t create culture. Like culture’s kind of like a mist or a fog. Like you have to create the conditions for it to appear. You can’t manipulate it directly and things like that. So trust, like basically outline your plan.

So say what you’re gonna do, and then do what you say and back people up. And, if you’ve got those people that you’ve hired right, and you build that trust and that culture starts to emerge, sky’s [00:21:00] the limit. And I’ve seen that happen a couple times in my career.

Tal Clark: That sounds great. Okay. What is a restaurant brand that you admire right now and why?

Marcus Wasdin: Buddy, it’s one that’s here, near and dear to my heart here in the, the Atlanta area, and I go to it way, way, too much, and that’s, Chick-fil-A, man. They’ve been the perennial… Just the level of service that they provide, right? And the level of consistency. Typically, when you go to a Chick-fil-A, you know exactly what you’re gonna get, and it’s gonna be the same every single time.

You know that, I’m gonna pull into a parking lot and see that there’s a line, around the building, but I know I’m still gonna get through that line in five or six minutes, right? So I have this expectation that they’ve set over years. And they just do it well, man. They just, they just really do, and their AUVs speak for it.

And it’s, I’ve always admired them and I’m a big fan.

Tal Clark: That’s great. All right. That’s good. I think I agree with you on that. And lastly, what professionally speaking are you most excited about for [00:22:00] 2026?

Marcus Wasdin: But yeah, I think it’s, we’ve danced around the topic a little bit, the whole discussion, which is, man, I can’t wait to see some AI get deployed out there that’s actually delivering some returns and some business value. I think it can happen in 2026. we still got half the year left. I can’t believe we’re halfway through the year, dude.

Tal Clark: Crazy, I know it. Yep.

Marcus Wasdin: But I, I think you’re about to start to see some of these things start to hit that, that are gonna make a difference from an AI perspective and, I’m excited about that. And then I think 2027’s gonna, continue to move that as well. But, I know that’s probably a cop-outy sounding answer, but, the twist I’ve got on it is I wanna see real outcomes, man.

I wanna see real value delivered, and I think we can do that before the end of 2026.

Tal Clark: Okay. Marcus, thank you so much for joining us today. Your leadership in the industry has made a huge impact on restaurant operations and the guest experience. For our audience, you can find Marcus on LinkedIn. We’ll link to his page in the show notes. You can follow [00:23:00] along with new episodes at instant.co/podcast or wherever you get your podcasts.

Please share the episode and leave a review. Thanks for tuning in to the Instant Payments Podcast.