While it made its name with branded consumer products such as NotMilk, Chilean foodtech firm NotCo expects that within three years it will be generating the bulk of its revenues from using its AI tools to help industry partners develop better products.
“I envision that in the next three years, the main source of our revenue is going to come from our technology, from our software, from the b2b services, and then from the foods part of the business,” NotCo cofounder and CTO Dr. Karim Pichara told AgFunderNews.
“This is the future of the company. Kraft Heinz was our first b2b partnership, but we also have a bunch of new partnerships with big CPGs looking to create entirely new products from concept to final formulation or to renovate products to reduce costs or make them compliant with regulations that are changing all the time.”
He added: “In many cases, they want to modify formulations but keep the same sensory profile, and we’ve been working with them on a lot of those projects. It’s not just about plant-based, it’s about coming up with elevated products in every category. We are now moving towards a software as a service (SaaS) business, where we are offering our technology directly to other companies.”
What next for NotMilk, NotBurger in the US?
NotCo declined to comment on the performance of NotCo branded products in the US, which it entered in 2020 with NotMilk in the chilled aisle before pivoting to shelf-stable products early last year. It did not provide any update on NotBurger and NotChicken, which were launched in the US in 2022/2023, while the NotCo US website now redirects to its b2b technology site, inviting firms to partner on AI-powered formula and concept generation, and cost reduction projects.
Former Chobani exec Meredith Madden, who was brought in to run NotCo US in February 2023 and promoted to general manager of North America later in the year, left the post in June 2014 to take over as CEO of NotCo’s joint venture with Kraft Heinz and has not been replaced.
Announced in early 2022, the joint venture has since developed a range of co-branded products including sliced cheeses, mac & cheese, and Oscar Mayer hot dogs and sausages. It is not clear whether NotBurger and NotChicken—which are listed on the KraftHeinz/NotCo JV website without any accompanying retail stockists—are to be phased out or if they will evolve into co-branded products with KraftHeinz/NotCo branding.
One industry source told AgFunderNews: “It would make sense to me to stop brand building and fighting to gain distribution [for NotCo’s branded consumer products] when KraftHeinz will do it for you.”
‘From design, we are an AI company’
Speaking to us after announcing the launch of NotCo’s Generative Ai Fragrance Formulator, a tool that generates fully-formed fragrance formulations “in seconds” based on natural language prompts, CTO Karim Pichara explained: “When we had the first version of [NotCo’s proprietary AI platform] Giuseppe, we realized we had a company, and we started from there. From design, we are an AI company.”
Thanks to Giuseppe, some of NotCo’s products such as NotMilk feature unexpected ingredients such as pineapple and cabbage, which Giuseppe identified as a combination that would generate lactones, aroma compounds found in dairy products.
However, its focus is often less about coming up with novel ingredient combinations than optimizing formulations and tweaking processing steps, often using familiar ingredients to meet a specific price point, for example, said Pichara.
“So recently, for example, we partnered up with a big CPG company in the US to develop two flavors for them using our generative AI technology where a big part of it was to get to a specific target cost.”
While the ingredients list for NotCheese, for example, might look similar to other plant-based ‘processed’ type cheese on the market, “Sometimes minor changes in processing or just in the ratio of one ingredient vs another can create a big effect on the final product,” he noted.
When humans approach formulation problems, they fix one variable at a time, and then test again, he said, “But that is clearly not the most efficient way to get to a final solution. The AI can deal with all of the parameters at the same time and optimize everything.”
He added: “Some companies are using AI very specifically to identify functional ingredients or finding or creating the best protein for specific functionality. But our focus has always been on end-to-end product creation, not just in plant-based meat and dairy, and we’ve been adding new capabilities as we go, such as our work on generating fragrances and flavors.”
For some CPG partners, he said, “We’re giving them the idea, the packaging, the nutritional profile, and at the same time, bringing a path to the solution.”
“A completely different set of economics open up when you use GenAI to create highly transactable intellectual property (in this case, formulations) at scale. With <$30k infrastructure cost and bespoke generative AI models, we can do in seconds what it takes experts to do in weeks or months and generate novel IP within a $40B+ perfume market or a $600B+ consumer goods market..” Aadit Patel, SVP, AI product and engineering, NotCo
Fragrance formulations ‘in seconds’
As part of a partnership with Chilean flavor and fragrance co Cramer, NotCo has been training Giuseppe on Cramer’s database of tens of thousands of fragrance formulations to develop a tool enabling it to develop formulations in seconds based on natural language prompts such as ‘an ocean scent on a breezy summer day on a tropical island.’
Historically, this could take days, weeks, or even months depending on the complexity or novelty of the brief, claimed Pichara. “We’ve developed scents that evoke memories and emotions – such as the nostalgic aroma of Christmas, the freshness of a spring morning, the tranquility of a beach at sunset, and the lively atmosphere of summer parties.”
According to Pichara: “We had to connect natural language with the jargon used by fragrance companies and then integrate that with the knowledge of the molecules and all the volatile compounds to understand the effect of flavor ingredients and combinations, and ratios [how much strawberry vs vanilla, say] on the final fragrance.
“We have automated and streamlined the entire process.”
While AI won’t replace trained flavor and fragrance scientists, he predicted, it will significantly speed up the development process enabling staff to become much more efficient: “Maybe one senior fragrance expert can now deal with 10 or 20 projects at the same time instead of three or four.”
“We began testing Giuseppe’s ability to evoke emotions through scents and were amazed by the early results. Our partnership with Cramer, combined with their expertise in formulation and raw materials, has allowed our AI to learn at an exponential pace.” Aadit Patel, SVP, AI product and engineering, NotCo
Partnerships with F&F companies
Asked about how partnerships might work, he said: “In Chile, we’re working with Cramer when it comes to the co-creation of new fragrances, and there’s no one better than them in that region to work with.
“But we are open to similar collaborations with other flavor or fragrance houses in other regions. We are the pioneers when it comes to generative AI for fragrance creation. So there’s huge potential to deploy our technology.”
He added: “As we are also a CPG company, we already have relationships with many flavor & fragrance houses in different countries, because they’ve been helping us with the sourcing of flavors for our products.”
Sponsored
Sponsored post: The innovator’s dilemma: why agbioscience innovation must focus on the farmer first