The Future of Food Service: Robotics, AI and Sustainability – Nipun Sharma, SJW Robotics

This is Techstrong tv. Hello and welcome to Techstrong tv. I'm Bonnie Schneider.

Climate change brings more extreme weather. This not only affects individual daily life, but it also impacts food prices and even the availability of what we want to order for dinner at restaurant. Well, innovative solutions are more critical than ever for the food service industry.

Today we'll be discussing the role of AI, robotics and automation to improve the efficiency and sustainability of food service. Joining me now is Nip and Sharma, c e o of s j w Robotics, who says Robotic Restaurant Automation is a game changer for the food service sector. Appreciate you having here, nip on A pleasure to be here.

Great. Well, can you begin by sharing a little bit more about your background and how you came to lead SS j w Robotics? Sure.

So I've spent the last 17, 18 years in the restaurant industry, and Bonnie, I've done everything you can think of. I've run steakhouses chains, I've run vegan restaurant chains, Chinese, middle Eastern Indian. Uh, I've been an entrepreneur.

I've expanded restaurant chains all over the world. I've been a public company, c o o, so really immersed into the entire business of, uh, of, of the food industry. And, you know, one of the things that really started disturbing us to me personally about 10 years ago, was a shortage of labor and, uh, in different pockets of different countries.

We started noticing while demand keeps going up, nobody wants to cook at home, uh, the supply side is kind of broken and the labor's becoming an acute problem, uh, uh, in addition to sustainability, of course. Uh, so for us, uh, you know, we looked at automation to see how we could bridge that gap, uh, in the lack of labor. And I started calling up on, uh, food robotic companies, uh, to see if I could work with them.

The solution could work for us. Came back a bit disappointed and realize we're not quite there yet. So at this point, I coincidentally met my co-founder, an expert in robotics and automation, and we decided to go from first principles.

How would we make the restaurant of the future? How would we check all the boxes of labor, sustainability, quality, accessibility, and start a restaurant from scratch. If we take away the human limitation of a restaurant, we can really stretch the boundaries on machine learning, on robotics, on ai.

And that's how we sort of got into it. We started doing some initial tests. We did a sustainability study, a scoping study, can this actually work?

Uh, at that point we raised a bit of capital, did a prototype. People got to come and see the machine, they got to kick the tires, press a button, taste of food. And then after that we really started rolling and raising additional funds, uh, partnership with Compass Group in Canada.

And Donatos Pizza is really getting our business model in shape now. That's, uh, fascinating. It really is what you imagine the future to be if you had to close your eyes and envision restaurants of the future.

Um, our audience at Techstrong is very interested in AI and robotics as well. Can you explain more how, how you're using these technologies and what you're doing? Sure.

You know, we, in the restaurant industry, especially managing large chains, we use a lot of data, right, to predict our sales. We know there's a ice storm in Montreal, maybe people won't come to your restaurant, so don't cook too much and staff accordingly. There's a storm in Florida, tomato prices are 400%.

Well, let's make recipes that don't include tomatoes. We try our best. We're not with a few, uh, data points that we have that could be weeks or months old.

What if we could do that real time? And so that's managing the supply side. Also, the inventory, right?

Uh, food can expire, food is mishandled, human leaves chicken outside for too long. There's all these factors that come, uh, that make our inventory management systems less efficient in a non-automated world. So we take away all these inefficiencies, look at data points that are real time.

So for us, AI and machine learning goes into two categories. One is predictive maintenance of the machine. We know certain parts are being used quite a lot.

It's gonna break down. That's make sure replacement parts are on hand person who can fix. It's also ready and available during a downtime.

And managing the food inventory effectively. We know chicken's expiring in 12 hours. That's gen generate, uh, a promotion, 50% of your chicken dish.

So managing all this stuff in a robotic environment, we can really push the boundaries. If you look at human kitchens, it designed for the average human right, like five foot, three inches average strength. Let's make a restaurant kitchen so they don't hurt themselves, but they can make food really, really quickly.

Uh, so we take that limitation away and say, Hey, what if my average human being was 10 feet tall and 18 arms and foreheads and Arnold Schwarzenegger strength, my kitchen would look different, right? So when we looked at our kitchen, we decided to go on first principles. Here's the raw materials, here's what the finished product needs to look like.

Let's create a system using advanced, uh, robotics and automation to make that food without the limitation of a human hand movement. Hmm. So when it comes to automation of, of food, I was reading that a different types of cuisine, uh, place different challenges.

Now that's interesting probably for anybody that cooks, you know, people have their preferences of what's easier to prepare, but I guess robots do as well. So can you talk about that, maybe using the example of Asian food versus Italian food and what's easier and what's not as easy and why? So, You know, when we first started a company, it was not meant to be a company.

It was solving our, my personal problems in our restaurant kitchen. And because I had done pretty much every cuisine time, coincidentally for us, the hardest to automation was Asian food because of the variability of raw materials. We think about it, there's sloppy noodles, there's chunks of broccoli, there's shredded carrots, there's flat snow peas.

How do you measure all these? We had to invent our own sensor. We're doing all this stuff.

So we figured if we could do Asian, we could do anything. If we can't do Asian, let's just leave the next generation to come up with a solution. Uh, and if you look at the quick service, uh, restaurant industry, we've become pretty good at managing operation that scale.

And again, remember, we don't have the best employees applying. So the system has to be pretty strong. So in most concepts, there's a prep area where the food is, you know, chopped your specification and placed in coolers.

Then there's a, how do you heat the food? Is it a wok, is it an oven, is it a grill? And how do you garnish it?

So it's a fairly linear motion if you think about how food is put together at a Chipotle station or a subway. So we've taken those basic principles of the quick service restaurant industry. So every cuisine will probably have a different way of heating the core food, but everything else kind of falls in a very similar linear pattern.

So we've started, of course, with I think, the most complex Asian food, uh, second cuisine, which happens to be, I think the easiest happens to be pizza. Uh, it's a simpler process to cook. You know, conveyor ovens replace a complicated walk system.

And then between these two ranges, I think we'll be able to do Mexican, Italian, Indian, everything else, because there's a similarity in all these things. There's some kind of a starch, there's some kind of a protein, there's some kind of a sauce, there's some kind of a cooking mechanism. So if you look at it from that perspective, uh, the range of, uh, food cuisines, uh, and again, we really think people buy food, uh, by brands.

You don't buy a butter chicken from a Chinese restaurant. So we really wanna partner with established brands. So the customers have an expectation how you'll be able to manage like Donna's Pizza, you know, you know what the pizza tastes like, what the quality standards are.

So when an automated machine is cooking it, we need to make sure that we can meet those customer expectations or exceed them. So we become really, uh, the conduit for brands, uh, to go autonomous. Have you had, um, I'm sure you have with your testing, but customers not be able to tell the difference between robot made and human made.

So, so, so, uh, if you, uh, maybe you've seen a video, uh, and the most proud, every time somebody's come to look at a machine and taste the food, they always say, oh my God, this is the best pat I've ever had. And that's what I'm most proud about, because nobody cares about your machine. People come here for food.

If your food is not star, you're not gonna win. Uh, but, you know, no, I love to see our machine is very futuristic, but really, if you think about it, it's not a very, uh, uh, mind boggling, uh, format of what we're doing in most restaurants. You know, the, the genius comes from the chef's recipes.

The, the robot is not really making the recipe, the sauces come pre-made. What happens in a controlled environment like a robotic restaurant is we don't make mistakes. We don't overheat, we don't over sauce.

Uh, so that means the recipes that are originally crafted by a great chef, uh, they remain intact. So we'll be as good, if not slightly better than the average restaurant using the same recipes. And I read that some of your new self-contained units Inc include refrigerated storage for up to 350 meals, including all proteins, vegetables, sauces, and starches that can make up to 60 meals per hour.

Wow. Um, can you walk me through how that Works? Yeah.

So a machine, uh, you know, the first section is a refrigerated part. There's 16 silos that contain raw materials and vegetables and proteins. There's 10 silos that contain liquids, a separate one for oil and liquid eggs.

Uh, so when an order is placed, uh, you know, we pick up the raw materials. In an Asian concept, it typically goes into a blanching station to heat up the food before it hits to the walk. In our case, it goes to a steam tunnel, and that's triggered only when an order is place.

And then the food gets, uh, cooked and plated and bold comes with the lockers on the, the side. You tap your phone and, uh, your meal is ready. And we also modeled this to be a restaurant.

This is, should have the same capacity of a typical restaurant should generate the same sales as a typical restaurant. So 350 meals, uh, is what the sweet spot is. Uh, uh, we could be slightly lower than that.

Uh, we can decide be higher than that. We're only doing three 50 meals as a restaurant, quick service restaurant, you're on the higher end of what the capacity needs to be. Of course, that needs to be managed by the recipes.

If I have 50 items, three 50 meals, not a whole lot, I can cook. If a one item that's 350 meals, that's very predictable. Also, write balance and the number of menu items.

And that's where, by the way, AI comes in to really support us and help us make sure our menus are meeting the expectations, uh, both in terms of shelf life and customer, uh, desires. Well, food waste is one of the biggest in the world in terms of, uh, carbon emissions, emitter of carbon emissions. So in what ways does SS j w robotics technology reduce food waste and contribute to environmental sustainability?

So in the restaurant business, you know, a lot of the food is, uh, wasted because of shelf life. Uh, shelf life because your inventory wasn't effectively managed or shelf life when the refrigerator breaks down in the middle of the night and you don't know until you come in the next morning. So when you're monitoring something 24 7, uh, we make sure the inventory moves along properly, both in terms of stocking it so we can predict our sales smartly, avoiding food wastage, generating promotions if food is about to expire.

So we can take all these, uh, precautions or we could become better every day as we learn about the demand factors for our inventory. So that's one way we really, uh, put a big, uh, dent on reducing food wastage. The other thing we do for c o two emissions is, uh, uh, we don't use gas.

So it's all induction cooking, and we did our initial study, we have 70% less carbon emissions compared to a comparable restaurant in the quick service space. And that's interesting things. Uh, restaurants, you know, the inefficiency comes from the equipment.

Uh, your air conditioning, a restaurant is generating a lot of heat in the, in the kitchen. Uh, and of course, uh, the restaurants are designed for speed and quality and not for sustainability. That means in the agent concept, you have a pot of boiling water all day to blanch your food.

In our case, we generate steam only when the machine, uh, gets in order, it remains idle otherwise. So taking away all these inefficiencies, not having interior lighting and all that stuff, we're able to make a pretty big dent on, uh, sustainability of, uh, carbon emissions. And that's, uh, you know, something you wouldn't have, uh, already thought of.

But one of the things everyone loves about restaurants is that they deliver and we can get whatever we want whenever we want it. So how do you, how do you implement technology like this with home food delivery? So, you know, that's the, uh, the greatest and the greatest thing and the greatest tragedy for the restaurant industry.

I think in the last decade, the biggest growth has come from delivery and covid accelerated the inevitability. Uh, people increasingly love getting delivered. All kinds of food, not just the old pizzas and stuff that we had before.

Uh, you know, these delivery companies came into our lives. Uh, what problematic is, uh, delivery companies charge a pretty big amount for restaurants. It is not really profitable for most restaurants to engage in delivery, but you do it because you have no choice.

And the reason is, you know, food is only about 25% of a total cost. 30% is about labor, about 10% is occupancy cost. So in our model, we don't have any onsite labor, and we have a very small footprint, so practically almost no rent.

So when we've taken away about 35% of, uh, the cost of a typical restaurant, we become actually profitable for delivery. So food in our restaurants, by the way, can be ordered onsite on the screen from your phone, from the Uber delivery app. If we're agnostic, the difference is either you come pick up the food or the driver comes pick, uh, picks up the food.

So we've actually made, uh, delivery profitable. And, and again, if by taking away the inefficiencies and the massive costs a typical restaurant has, what if you could eliminate those costs? Again, food is only about 25, 30% of your costs.

Let's make sure that our numbers, uh, aligned with those costs. That's interesting. And, and this was, I'm just curious, how do you, uh, manage security though, if you're, you know, if everyone's a robot, you know, to keep everything safe?

Uh, so, so our machine, by the way, had to go through an N Ss F certification process. Little details like, you know, we have the locker systems where, you know, BOL is ready, uh, but there's a machine inside. Uh, people could put their hands inside and, you know, have serious problems when the machine hits you.

Mm-hmm. Uh, so the design of the machine has been, uh, uh, by, by default make sure that it's, uh, it's very safe for consumers. There's a wall between, uh, the robot and, and, and the, and, and the food locker system.

So you cannot really put your hand through it. Uh, and most parts of the machine are inaccessible, uh, by anybody except the engineers that come to service it. Uh, so for example, our partners encompass, they put in the raw materials like, uh, broccoli and chicken, the machine, when the machine prompts you to do so, they change the air, uh, HVAC filters when it's time to replace them, and general maintenance.

Uh, and these, uh, initial machines are all going into, you know, compass facilities like universities, hospitals, and office towers. They already have a lot of people working in these buildings. And, uh, we've estimated they'll spend a few hours, uh, every day to make sure that machine, uh, there's nobody that's drops something.

Uh, of course we have cameras and everything else to make sure that, uh, the machine is also safe from customers and, and vice versa. So the box be self-contained. The only point of access is the food locker for customers, and that's safely, uh, secured to make sure you can't really put your hands all the way through.

You know, it, it's funny, I can envision this in, in airports and, you know, places where you get a lot of people, like you said, a hospital or an office building. Um, kind of a two part question. Is there any place where people, where you, you envision, um, your, your this, this idea, this concept being where people might not think it's the most obvious place?

And also do you, what do you see in the future? I mean, do you see, we'll actually have robots taking orders? I mean, where do you see this going in the future?

Well, I, I think, uh, uh, the obvious place for a lot of these robots is where you can find labor. So a lot of 24 hours facilities, like hospitals, um, and airports. So you mentioned airports, which is amazing because, uh, uh, the labor costs in airports are significantly higher because the employees need to be vetted and you have to park your car, et cetera, et cetera.

The smoke control environment, the demand in airports is obscene. Uh, and we've seen after covid, not all airports have recovered. Many restaurants are operating at half capacity or simply closed down first.

The obvious basis is where the need is, uh, where the labor shortage is very acute. And then it's about convenience, right? How close can you be to your customers because of a small, uh, footprint?

And because we're real estate agnostic, we can have restaurants where traditionally you couldn't have restaurants. So for example, in an airport, I don't have to be in a food court. I can just be closer to the terminal.

How can we be more closer to everybody else? So I think, uh, the, uh, evolution of this thing, I can see every brand having, uh, a bunch of robotic restaurants to compliment the existing restaurants. Uh, I can also see, uh, as we get better, uh, as we get more technology on, on communicating with the customers, and we already have, uh, AppSec like Uber and loyalty AppSec from brands, uh, that we can communicate with our smartphones.

Increasingly, voice technology is getting integrated on acceleration will actually, you'll be able to, uh, ask the machine to make certain food for you, ask about your post on allergens. A lot of that technology, and we're partnering with companies like Nvidia that gives us access to, uh, voice generation and ai, uh, to make sure the communication between people is great. And again, order ahead.

So you know that you're gonna be in that area around two 30. The machine will start cooking when it senses your phone is com, uh, some somewhere close by. So the convenience factor is gonna be great.

Um, and accessibility is gonna be great. So if we're an environment with senior citizens, uh, homes where they're not looking at a small phone, they'll be able to talk to the machine and get their orders ready. So I can see that advancement in accessibility becoming more interesting.

It Really is. It's just a, it's a, it's a fascinating to imagine and, and I think it'll be great to, uh, get these meals to more people that maybe couldn't get them and also solve the, the labor shortage that you mentioned and reduce emissions. So it sounds like a win-win.

Um, Nipun Sharma, c e o of s j w Robotics, thank you so much for joining us on Techstrong tv. Thank you for having me, Bonnie. Really nice to be here.

Absolutely. All right, well stay with us on Techstrong tv. We're gonna have a lot more coming up.