It can whip out a crispy dosa or make you a fiery chicken masala. And it will never ask you for a raise
(Illustration: Saurabh Singh)
WHEN THE TRADITIONAL restaurant business got upended by the pandemic with more and more individuals choosing to have meals delivered to their doorstep, Karan Tanna, the founder and CEO of Ghost Kitchens, faced a dilemma peculiar to those who run cloud kitchens. How do you expand your kitchens rapidly to cater to the growing demand from the food delivery model and yet be able to train and retain the requisite staff while also ensuring that the taste of your meals remains consistent through a large geographical stretch?
The cloud kitchen model which began a few years ago with the advent of the online food aggregators—where meals are prepared only for deliveries, thus saving on the cost of operating a dining establishment, and often with multiple food brands operating from a single kitchen—got a fillip during the pandemic. Today, it is among the hottest businesses to be in, with investors pumping in big money and cloud kitchens mushrooming across cities.
Tanna’s over 100 kitchens that operate out of neighbourhoods in multiple cities capture the palette of India’s burgeoning middle class—from biryanis, pizzas, Chinese dishes to momos and more. Faced with the opportunity and challenge of scaling themselves up quickly, Tanna turned to an option many large cloud kitchens in India are taking: automation.
Many of his outlets already employ machines such as fryers that whip out items like French fries or pans that cook parathas, flatbreads and patties with little or no human intervention. All that is required of his staff here is to load the necessary ingredients and press the digits required to select the dish. But now, after trying them out in a couple of his kitchens in Mumbai, he is rolling out more complex machines such as those that dish out the perfect crisp dosa or automatic woks that can churn out anything from a Chinese fried rice and pasta to a spicy chicken masala, all on its own, across his kitchens. Next on the line in a couple of months, Tanna says, is a pizza machine that can make pizzas on its own.
“It’s amazing. You select the knob for [French] fries, and that automatic basket carries the fries into the oil only once it is heated to a specific temperature and removes it after a precise amount of time has passed,” says Tanna. “Let’s face it. There are several processes in a restaurant business—toasting buns, heating parathas, frying something—where humans don’t add much value at all.”
After more misses than hits, robots have gradually begun to take small but steady steps into our kitchens in the last few years. There is a salad-making robot, Sally by California-based Chowbotics. There is another California-based company, Miso Robotic’s Flippy, which, it is said, can flip 150 burgers in an hour. Another of its robots, the tortilla chips-making Chippy, is currently being tested by the American fast food brand Chipotle. There are pizza-making robots, such as the one developed by a small French pizza chain Pazzi. And then there is Moley Robotic Kitchen (by London-based Moley Robotics), whose two robotic arms hanging over an oven, with the click of some buttons on a touchscreen, can pick up and put down saucepans and spatulas, stir, whisk and flip, and cook some 5,000 different dishes. There is even a robotic bartender, by the Italian company Makr Shakr, whose arms can measure, mix, shake, pour, and even garnish cocktails.
Most of these machines have just been introduced into the market in a small scale or are prototypes close to deployment, and there is still probably room for improvement, but the once remote likelihood of robots dishing out meals in our kitchens is gradually becoming more of a reality. We may have assumed that robots wouldn’t make their way into the food and restaurant business since this industry is seen as one that requires a human touch in a way that, say, manufacturing does not. But robotic companies and many in the hotel industry are betting the trade-off, lesser human touch for greater efficiency, will be worth it. Besides, most robots currently are taking over work that is dull and sometimes dangerous. So, while the robots sweat it out in the kitchen, human employees are free to focus on other tasks.
In India, automation in commercial kitchens is mostly coming up in the cloud kitchen space. “The cloud kitchen is a very brick and mortar business. It needs a lot of capital and extensive operations to scale up. If you can bring down human intervention, naturally scaling becomes more efficient,” Tanna explains. “If you teach a human to cook a certain thing, if he quits, you need to train someone else. Even if he does not quit, there is no assurance he will cook the same exact thing all the time. If you automate the process, it becomes more efficient. And it helps you grow and scale faster.”
More automation will come in commercial Indian kitchens, those in the cloud kitchens business say, because efficiency and the standardisation of dishes will become even more crucial as the food delivery model aims to further shorten the time it takes you to order a meal on an app and have it delivered at your doorstep. “People on Twitter don’t understand,” says Eshwar Vikas, cofounder and CEO of the food robotics company Mukunda Foods, referring to the criticism on social media that Zomato’s plans of ensuring food deliveries in less than 10 minutes will mean that food will not be freshly cooked and delivery partners would be exploited. “Today, using machines, you can cook a plate of fried rice in less than two minutes. You can grill sandwiches and fry kachori in just a few minutes. It’s a fact.”
Kitchen robots and more automation, according to Vikas, will be the key in ensuring food delivery platforms like Zomato succeed in breaching the 10-minute food delivery barrier. Mukunda Foods, one of the largest food robotics companies currently in India, will probably play an important role in that. Zomato, which plans to use its robots in its “finishing stations” to make super fast deliveries, last month acquired a 16.66 per cent stake in the company. But, this hasn’t stopped its rivals Swiggy and Ola (through its quick grocery delivery arm Ola Dash) from purchasing Mukunda Foods’ products to kickstart their own 10-minute food delivery plans, according to reports.
Mukunda Foods did not start as a robotics company, but as a restaurant chain. “I was looking to establish a McDonalds of Indian food like dosas, idlis and vadas, like a chain of 1,000 outlets,” Vikas says. The dream soured by the time the third outlet came up in Bengaluru. “We couldn’t maintain the consistency and quality of food,” he says. So Vikas and cofounder Sudeep Sabat, both electrical engineers, created a machine that makes dosas by itself.
This machine, called DosaMatic, was meant only for their outlets. But the resultant press coverage and requests by other hotels to make such a product for them too made the two to pivot their business to creating robots for commercial kitchens instead.
Today, its multiple machines have automated, according to Vikas, about 90 per cent of the cuisine we order through delivery apps. The company is also working on an automatic cooking-cum-vending machine, which could be placed in areas like large offices, where with the press of a few buttons, individuals would be served freshly cooked meals like biryanis, pastas, noodles, upma and poha.
“There are a lot of challenges in restaurants—everything from that of manpower, consistency in taste to the high attrition rate. We want to help restaurants move from a manpower-oriented business to a process-oriented business. You want your business to be limited by demand, not by the supply (of manpower), especially not the quality of your supply,” he says.
But robotics isn’t just promising to change our commercial kitchens. Just as one set of machines enables food businesses to cater to this ever-growing culture of ordering-in, to an extent even manipulating people to do so, there is another set, much closer to our lives and our kitchens than we think they are, telling us we will never need to order another meal because we are too bored to cook in our homes. Using the latest in robotics and artificial intelligence (AI), and fitted with spatulas, cameras and sensors, they promise to be the most revolutionary thing that hit our kitchens since the invention of the refrigerator. That just by putting in the ingredients, pressing some buttons, out will come a hot meal, big enough to feed a family, in less time than traditional cooking methods take.
Among them is the machine that Bengaluru-based robotics startup Nymble has developed. All a human needs to do here is to load the ingredients, then using an app on the phone and the display screen on the machine, select the dish they want the machine to prepare. “There are primarily two aspects of cooking—the mechanical and the intelligence aspects,” says Rohin Malhotra, Nymble’s co-founder. The first aspect involves our hands and fingers performing tasks such as grasping, picking, dispersing and stirring. The second aspect involves intelligence gathering and processing: the picking of visual cues like the onions turning brown or the pasta sauce thickening, to decide what step is to follow next. Nymble’s machine, which has tentatively been named Julia, Malhotra claims, can do all these tasks. It uses cameras (both thermal and a regular one), multiple sensors to constantly generate and once this information is processed using AI, the machine will decide, much in the way a human cook does, when the flame is to be lowered or certain ingredients added.
Unlike many other under-development robots, Nymble’s Julia is a sleek and compact device that can easily fit on to a kitchen counter. “The kitchen is the most prized real estate in a house. Have you seen the cabinets, the boxes and dabbas stacked atop one another? It’s got to be compact to make it inside a kitchen,” Malhotra says.
Malhotra felt the need for such a machine, he says, several years ago when he stepped into college. Then, back home, his mother had a new task for him and his two sisters. The three siblings would from now on have to cook dinner for the family. Throughout their childhood, Malhotra says, the three had seen their mother struggle through her days and nights juggling her career as a pharmaceutical scientist while whipping out meals for the family. But thrust with this new responsibility into the kitchen, he realised, he says, just how much of a drudgery cooking everyday was. “We looked for all sorts of solutions, like options for healthy meals being delivered, but nothing worked,” Malhotra recalls. “As a mechanical engineer, I couldn’t help but think, ‘Surely, there is some way some of this process of cooking could be mechanised’.”
EVEN FOR THE MOST enthusiastic of cooks, one of the most dull aspects of life is the need to feed ourselves about thrice a day for the rest of our lives. If you’re a woman, you’re probably going to be unfairly saddled for not just yours, but your family’s meals too. Like the many appliances in our kitchens, these machines also attempt to retrofit traditional women’s work into the lives of modern women. But the promise they make here is that cooking will now be transferred from the woman in the house to this new machine.
“Imagine freeing up 2 to 3 hours a day for half a billion women—imagine the immense economic impact you’re creating by giving 20 per cent of their waking hours back to them,” says Cohan Sujay Carlos, the Bengaluru-based founder of Mechanical Chef that is developing a prototype of a kitchen robot. “The problem is we have made these great technological advancements in so many aspects of our lives. But tech has not walked across the corner and taken a look at the kitchen really. We are in the modern age. But why is a woman stirring a pot for 30 minutes? Once she has worked out a recipe, why does she have to do it all over again and again?”
Carlos, an AI researcher who worked at IBM in the US before relocating to Bengaluru where he established a consulting firm, has been developing a robot that cooks meals for several years. Currently in its seventh iteration, he has been refining the product, eliminating the many problems such as the mess early versions of the robot created, the bulk and size of the product and the ability for users to take apart and clean the many containers where ingredients are loaded (in the past, when kitchen robots were first introduced, many overlooked this aspect, leading to food particles sticking to these containers and causing food poisoning). The device cannot chop vegetables currently—this is too complicated, Carlos explains, and can be handled for later iterations—and while one cannot also deep-fry items, bake, or make items that require some amount of shaping—it can do everything else. “That’s the great thing about Indian cooking—most operate on a simple premise of heating and stirring,” he says. About 100 recipes have been programmed into the machine, but users can also tweak these recipes, and over time, more recipes will be added. “From the technology standpoint, I would say it is ready,” he says. “Now, it is a matter of raising money and getting it into manufacturing.”
Malhotra believes his robots should hit the markets by the middle of next year. They initially plan to launch in the US, before introducing the product in India. According to him, because they will constantly be adding newer capabilities and things like recipes to the product, it will be sold with a recurring amount payable every month. “Think of it like Netflix. There is no fixed catalogue. It keeps giving you new movies and shows. Similarly, we keep giving new features, new capabilities, and recipes,” he says.
At the time the machine will be sold, Malhotra expects, that around 500 recipes will be programmed into the machine. While users will not be able to add new recipes or tweak them, there will be a constant system of feedback, and over a few weeks, the robot will be able to customise a meal according to the user’s preferences. “Think of it like a new human cook who joins your house. She won’t know immediately how you like your dish prepared. But over a week or so, she will have learnt how much spice to put, how much oil,” he says. “Only more efficient.”
But isn’t cooking as much, if not more, art than science? About human estimation and feeling than scientific precision? A throaty laughter fills the phone. “That’s a very romanticised version of cooking,” Carlos says. “If anything, the machine shows, it can be all science.”