At food-tech company NotCo, AI wears the chef’s hat, creating plant-based recipes for some very familiar products with some less familiar ingredients. Take chicken. The main ingredient in NotCo’s NotChicken line? Strawberries. And inside its NotMilk line, you’re going to find cabbage, pineapples, and coconut — a combo more suited for smoothies but together creates a creamy, foamy milk option ideal for iced cappuccinos.
That’s the data-driven approach from Giuseppe (yes, they named the AI) who works back and forth with some very human chefs to develop products adopted by well-known brands from Shake Shack to Burger King. At the core of NotCo’s success, says Fernando Machado, NotCo’s chief marketing officer, is asking “Why Not?” And then letting AI start inventing.
“We use AI to create a version of plant-based that could foam infinity-times better than any non-dairy, and actually, I would argue that foams even better than 2%,” Fernando Machado, NotCo’s chief marketing officer said on the Signal Stage in July. “This is all partnering with AI to make this stuff happen.”
You can hear more from Machado and NotCo’s unique approach to upending the food world today from his talk at Signal 2023 in the video below or read our lightly edited transcript.
TRANSCRIPT:
John Battelle
The beginning of Signal, the first few years, was to bring the outside in. We were focused on the digital acumen pillar of getting better and learning how to go to market and internally digest the digital revolution. Most of the entrepreneurs that came to this stage back in those early days were from the Silicon Valley. They had a similar profile as you might expect. What I’m very proud of the team here at P&G has done is brought people from all over the world who are changing the world. We’ve already seen some of them, but I’m about to introduce you to another. His name is Fernando Machado. He is the chief marketing officer of a company called NotCo, and they are reinventing the food industry. So please join me in welcoming him to the Signal stage.
Fernando Machado
Thank you, thanks for having me. I confess, it feels a little bit weird to be here. I never thought I would be at Signal after 18 years of Unilever. But it’s a real pleasure to be here. I couldn’t say no, to the invitation. When I got the invitation, I was like, ‘Why not?’ Right? In fact, #why not is where the entire principle of NotCo is predicated on. When the company started around seven, eight years ago, our founder and CEO started the business because he felt that the food industry was rather inefficient, a lot of trial and error, long lead times to develop things. He thought that the food industry was also inefficient when it comes to sustainability. I think we all know, like what I’m talking about here. So he started with the company in his house in the garage, I think, actually, it was in his sister’s room. I’m not joking. In his started with mayo, the company started in Chile. Mayo, in Chile, is one of the top countries in terms of per capita consumption, and he created a plant-based meal in the company.
If you go to the website, this is what you read. ‘We are a data company developing artificial intelligence to disrupt the food industry.’ I find that a little bit boring. So I’ll try to explain with my own words what we do. I cannot tell that to my mom, she’ll be like, ‘What the hell are you talking about?’ So we have an AI chef. The name of the AI chef is Giuseppe. It’s based on the Italian painter Giuseppe Arcimboldo from the mid-1500s. You probably have seen some portraits with vegetables and fruits, it’s a classic type of portrait. It’s basically like creating art with plants. That’s why the chef is called Giuseppe. What we do so far, and I’ll go back to the ‘so far’ in a second. What we do is we take animal-derived products, we do an analysis on the product like chromatography, evaluate the physical properties and all that good stuff and we do all that, we put on the AI and the AI spits five recipes. Then our chef team and R&D, so it’s always AI plus humans. We take that and prepare the recipes and go back to Giuseppe saying, ‘Hey, number three was the best.’ Then Giuseppe spits another five and so on and so forth. We converge into the target really quick.
One of the beauties of our model is that we are not anchored in one single ingredient. If you are like me, I drink a lot of non-dairy not because I’m lactose intolerant just like the taste. But they come with certain limitations in terms of the physical properties. I will talk about that in a second. In our case NotMilk, which is one of the products we have. By the way all the products are NotMilk, Not Chicken, NotBurger. It all starts with ‘Not.’ That’s why the name of the company is the Not Company. We say NotCo, it’s kind of easier. But it’s not just one ingredient. NotMilk has cabbage, pineapple, coconut, and so forth. If you look at the environmental footprint, for most plant-based products, they will smash the animal proxy, in terms of environmental footprint. I just pick water usage and C)2. You can see there the difference. I could have picked, NotChicken, one of the key ingredients of NotChicken, just to make the point is a strawberry, you would never guess. I think that if you ask someone from R&D to formulate a plant-based chicken, they would probably not start with strawberries. That’s also the beauty of the AI. It really pushes us out of our comfort zone, in terms of which ingredients to explore. Usually, we come up with some like really cool combination of ingredients, not chicken, this is the water usage and CO2.
I know you’re probably asking, how does it do taste-wise, because taste is the main driver of the category. In Flexitarians, which tend to be the target audience of plant-base, they don’t compromise. It’s not like vegan. Vegan, like they are willing to compromise on taste. Flexitarian they won’t. So it really needs to match the animal target and hopefully also be much better than the other plant-based products that are out there. So that’s what we target with every product. I’m not going to bore you, explain the graph and all that but it’s basically statistically significantly parity to an animal-target and superior to the main competitors in the respective countries where we have the products.
The other cool thing about our process is the time it takes to develop. I think you probably empathize that working at Unilever I was running quick stability tests at higher temperatures because I was rushing to get the stuff out so not lose the Walmart window to ship my product. And NotMayo, which was one of the first product we created, took us 18 months. As you can see, we are kind of getting better. We are not, the AI is. Actually the AI is getting better at matching the target. Right? I mean, we created a version of Cadbury’s, like NotCadbury in two months, and there is no one there in the team who is an expert in chocolate. The guy Karim was one of the founders, who was our AI boss. He’s a PhD in astrophysics using AI. He probably doesn’t know how to bake a cake. He knows AI. Our VP in the AI field, if you look at his CV is NASA. Seriously I work with someone who works for NASA? Yes, we do. So it’s a very high caliber of people developing the model partnering with R&D, and partnering with a very talented team of chefs.
NotCustard which is a product that you can buy. All the plant-based custard and plant-based shakes from Shake Shack, they are made by NotCo. If you go to Shake Shack or if you go on Uber Eats you will see the description of the product. It’s a partnership that we have with them and is a pretty awesome product. Three weeks to develop. The Shake Shack guys couldn’t believe when we came with the prototype and they try and it’s like, ‘Yeah, you nailed.’ Don’t take my word for it. Go buy it and try for yourself. It’s no wonder we have so many partnerships. We have a very strong partnership with Starbucks in Latin America with Alsea who operates Starbucks in most of the countries in America. We have a partnership with Dunkin in Latin America, we have partnership with Domino’s, we have the Shake Shack partnership. We supply either NotChicken or NotBurger to 8/9 countries for Burger King in Latin America. So we have like lots of QSR partners which is a cool way to also get the brand known because it’s a real partnership. You’re not just like the supplier. It goes on the merchandising, we do a joint advertising plan together and all that good stuff. We have a joint venture with Kraft Heinz. So we are doing every single one of the key products from Kraft Heinz to lounge in the US. This is a is a GAV for the U.S. market. So we collaborate with them. I dare you to do a NotMayo blind test at your house and someone tell me the product is plant-based. it tastes exactly the same. We just launched it NotCheese which is the Kraft Singles and there is more coming on that space. Kraft Singles NotCheese, 30% market share in the places where we have the product, 80% incremental for the retailer with a higher value. So it’s a really good story to introduce the product in retail. And Kraft Heinz is number two most innovative company, thanks to that partnership. I’m not just saying that if you go to the Forbes article and read you see what I’m saying.
But this is not the part that excites me the most. Because I think that if everyone does really well, the game of matching to the animal target, it will probably become commodity. I mean, if everyone is doing that, well, it becomes a price game. What really excites me is when I go to San Francisco, where is our R&D in AI lab, and the guards are playing with the AI in ways that I never thought would be possible. This will sound like I’m making it up. But when I was a kid, and I was a bit weird, but I used to think of a color that doesn’t exist, and for obvious reasons, that’s not possible to do. But the AI can help us do that type of stuff. I will never forget, when I when I went to San Francisco for the first time in they were like trying stuff in the kitchen. And they said, ‘Hey, Fer, people call me For, like, Hey, try this.’ And it was like chocolate, like caramel. I don’t really know what it is. So we fed what chocolate was to us, based on the chromatography and all the analysis that we do, and we asked Giuseppe to create a flavor that’s not chocolate, but that humans will like, based on the fact that they like chocolate. And they was like, ‘Holy shit.’ Look that was not a very technical term, but that was my real reaction. We showed the product to people who are chocolate experts and they could not pinpoint what the product was. They would say, ‘Oh, it reminds me of this, reminds me of that.’ But they wouldn’t say, ‘Oh, this is it.’ Because there was no it for what we created. That’s just one example. Every time I go there, they’re like, ‘Try this coffee that there is no coffee. Or hey, ‘Try this mole sauce that there is no chocolate and there is no animal product.’ It is just amazing what we can do with the technology.
These are just some examples of things that we have already launched or are about to launch. But honestly, like the limits of the technology are way ahead of some of the stuff that I’m showing here. This is a solving a real problem. I don’t know if you guys drink plant-based milk. If you do you probably know that it’s really hard to form. That’s a problem for Starbucks. Starbucks today, I sells around 70% of the drinks, they are cold drinks, which makes it even harder to form because the machinery was not necessarily developed for cold drinks. When you buy a cold drink, and it’s plant-based and it doesn’t foam, when you get to your table or when you get to your car or if you order delivery, the foam goes down. Right. So it feels like Starbucks is bullshitting you because like it’s like a headspace that’s the size. Then people go back there and ask for more milk or Starbucks, this is a hypothetical example, Starbucks has to fill more and then you know what happens with the margin? Right? So we use AI to create a version of plant-based that could foam infinity times better than any non-dairy actually. I would argue that foams even better than 2%, and it’s done certified vegan, halal, kosher, allergen-free. So this is all partnering with AI to make this stuff happen.
This is an actual example that we did in Brazil. In Brazil, there is a large segment of protein shakes. The problem with a protein shakes that are animal-derived that has animal milk is that it’s very hard to find the right balance between protein and carbs, it tends to have more carbs. So for a 15 gram protein milk-derived shake, you have like 16 grams of carbs, which is weird, because maybe they should call it a carb shake. So we created a product that has 15 grams of protein, 3.5 grams of carbs. It was very tempting, we had to project that onto a building which owns 80% of the market share in Brazil. This is just to show that we are having fun because every time I go there, they show me something that I find mind-blowing. We are launching at the end of this year, a platform here in the U.S. which we will call NotCo drops. This is not public yet. Because like I don’t know if there will ever be a market for plant-based mole in the US. But I think is a really cool story to tell. To explain how the hell we use AI and precision fermentation to create a mole. And I want you to tell those stories. So we created NotCo drops, where we’re going to do like very limited batch in the lab and sell. I’m going to lose money, of course, doing that. It’s more of a marketing thing than a source of revenue. The first one will be a plant-based fried chicken skin, because why not? Right.
Just to close, because I don’t want to run over time. I think what I was saying next goes really well with the previous conversations. This is the cover of The Economist, I think it’s from April of this year. I love this cover. Because many of the questions that I asked about AI they relate to, ‘Oh, Fernando, do you think AI will kill all of us or steal the jobs of everyone?’ I don’t know, maybe because I’m Brazilian and I’m hopelessly optimistic about things. I think that AI to me is something that will be used to augment our ability as humans, our creativity as humans, to find solutions to problems that we could not solve by ourselves. In our case, it will be to invent the food industry for good because why not? Thank you very much, guys. Thanks for having me.
