Miso Robotics and AWS collaborate to scale restaurant robotics

The How We Built This blog series features conversations with C-suite executives that are building innovative restaurant technologies on AWS

The recipe for restaurant recovery begins with efficient operations and satisfied customers. But inflation, rising food costs, and supply chain disruptions present daunting headwinds. According to the Bureau of Labor Statistics, foodservice employee turnover rates rose from 4.8 percent to 6.9 percent in 2021, the largest jump of all industries. Now, restaurants are turning to leading-edge technologies to build resiliency.

I had the pleasure of speaking with Chris Kruger, chief technology officer at Miso Robotics, about how Amazon Web Services (AWS) and Miso Robotics have collaborated to help restaurants improve operational efficiency and enhance customer satisfaction. Their products, Flippy 2, Flippy Lite, CookRight, and the upcoming Sippy, are shaping the kitchen of the future.

Restaurants are eager to optimize kitchen efficiency. How did the idea for Flippy originate?

Miso has been dedicated to food technology and robotics since our inception. Before Flippy, there were some previous incarnations, like Flippy flipping burgers at Dodger Stadium. I joined Miso as they figured out that Flippy at the fryer was perfect—it is a job that is repetitive and tedious, which is a great fit for a robot. This is not about replacing human work—it is about having technology do tedious jobs that humans do not enjoy as much.

We often hear from our customers that worker strain is, quite literally, a top pain point. How are restaurant robotics alleviating this?  

We have heard from restaurants that there is a lot of employee dissatisfaction around the manual labor aspects of restaurant work, whether at the fryer or flipping burgers. Even without a tight labor market, these positions have traditionally been challenging for employee retention. Our technology helps restaurants to stay open longer, and more often, with less manual labor required. Our customers can promise their employees more rewarding work with less strain.

Tell us about Miso Robotics’ cloud migration journey. 

While all our products work on premises, we stream data to the AWS Cloud to analyze performance, scale quickly, and proactively predict restaurant needs. The cloud has been a major part of our success.

You’ve used the AWS Migration Acceleration Program (AWS MAP), a comprehensive and proven cloud migration program. How did that accelerate your cloud migration?

AWS MAP has been instrumental for us. When it was time to migrate, we reached out to AWS, and they immediately connected us with resources that directly supported our migration activity and connected us with key AWS Partners. The AWS team and its Partners helped us fill internal staffing gaps. AWS MAP helps us focus on our products and customers, not our infrastructure. Now, with AWS MAP, we have accelerated our ability to optimize our team to experiment, scale, and innovate on behalf of our customers.

This sounds oddly familiar in that Miso is helping restaurants optimize workforce efficiency—we hear from restaurants consistently that the labor shortage has made this critical.

For our customers, a lot of the labor that would have gone directly to fry cooks usually gets redistributed. Our restaurants are seeing considerable improvement in speed of service. Labor optimization powers the restaurant’s success, and from an employee perspective, they are no longer frying the food—they are learning how to interact with a robotics system. I could see a day when restaurant staff members are trained up on AWS services and could have their own view into the robot to see how it is doing in the restaurant.

Speaking of robots, Miso has now announced a formal collaboration with AWS RoboMaker. How has this been instrumental?

Robotics is multidisciplinary, and as a software developer, it can be daunting to make changes to the stack. We are fortunate to have a physical testing lab here with Flippys in it, but even so, there is a delay to test software on a Flippy unit. Without a simulation environment, we would not get visibility into the software development that powers our innovation. The simulation environment empowers our developers to innovate in the cloud and test their changes prior to submitting them. This provides quicker and more cost-effective feedback, which helps us build more features and scale more quickly.

We hear a lot about apps and web development pipelines. What does a robotics software delivery pipeline look like? 

Excellent question. Robotics software is ROS based so we do use the robot operating system but even still, there are probably over 100 different open source packages that come together in our stack. That is a big stack to maintain, and we are constantly upgrading components, building new features, and tuning performance. We are articulating and pulling apart aspects of the system so that it is easier and less daunting to develop in a particular part of the system. Then we have our overall integration that we run. We are testing the longevity of the software and cooking food here in our lab on an almost-daily basis. It is exciting to track the product’s performance to see how the changes we make impact our customers.

Does optimizing software delivery pipeline also accelerate go-to-market efforts for new products, like Flippy 2 and Flippy Lite?

Yes, and an example is when Chipotle came to us and asked, “Can you help us with a robot that cooks chips? And we want it to do a few other things too.” We took their requirements, designed a system, and built the whole thing in simulation prior to giving them an answer. We wanted to see that the robotic arm and sensors that we were proposing would work properly together and provide the performance level that they desired. We simulated the entire product prior to starting the engagement. They benefited from this level of rigor, and the simulation allowed us to develop Chippy for Chipotle in less than 1 year. If we did not have the foundation from Flippy and the simulated continuous integration pipelines, that would not have been possible.

How else has AWS powered operational efficiency?

We build dashboards on Amazon QuickSight, which allows everyone in an organization to understand the data, that measure the cloud-based testing that we are doing and give us metrics to monitor our progress on stabilizing new code. One idea for a new system that we are developing originated within the engagement we have on AWS RoboMaker. We are also learning how AWS IoT Greengrass, an open-source edge runtime and cloud service, might help us replace existing data pipelines. AWS RoboMaker has opened a number of new areas for us; it’s hard to list all the ways that we are using the plethora of services available from AWS.

What ways does AWS help the sheer scale of Miso’s deployment to fast casual restaurants (FCRs) like Chipotle and Panera?

What AWS and AWS Partners have helped us with is how to think at that scale. AWS helped us on data privacy, data retention, and how we think about segmenting of data across regions. AWS has been great in helping us understand different ways to architect new solutions into our cloud services. That’s the difference for me. Robots are going to be everywhere, and AWS helps us to scale and achieve a global footprint.

How do AWS and Miso Robotics improve customer satisfaction?

We monitor our robots’ performance daily, which helps us to proactively reach out to restaurants so that things run smoothly. Our vision is for a restaurant to have multiple Miso products working in unison, integrating with the back of the house to offer a better experience for our customers. We hear from our customers that, after the robot is in place, they do not want it to leave. It is a valued team member that shows up every day and gets the job done. As White Castle put it, “Please do not take my Flippy away!”

Do you see Miso Robotics expanding into hotels or even airports?

We see a lot of possibility there, especially with CookRight’s coffee monitoring that we brought to Panera. Hotel kitchens are a natural fit. Sippy is our automated drink program that will be available next year, and we are excited to deploy a number of products that work together in restaurants and across the travel and hospitality industry for years to come. We are just getting started.

Conclusion

Though robotics automation once seemed a futuristic vision, Miso Robotics’ expanding roster of restaurants and rapid product development alongside AWS are clear proof that robotics are reshaping kitchens today. As restaurants seek new ways to overcome the labor shortage, improve operational efficiency, and enhance the customer experience, the cloud, automation, and robotics will all play a role. To learn more, view Chris Kruger’s session at re:Mars or watch Industry Innovators: Travel and Hospitality on demand to hear how Dine Brands Global and White Castle are using robotics today.